Professional Documents
Culture Documents
VOLUME 1
Thom Brooks
LEIDEN BOSTON
LEIDEN BOSTON
2011
ISSN 2211-2014
ISBN 978 90 04 203426
Copyright 2011 by Koninklijke Brill NV, Leiden, The Netherlands.
Koninklijke Brill NV incorporates the imprints Brill, Global Oriental, Hotei Publishing,
IDC Publishers, Martinus Nijhoff Publishers and VSP.
All rights reserved. No part of this publication may be reproduced, translated, stored in
a retrieval system, or transmitted in any form or by any means, electronic, mechanical,
photocopying, recording or otherwise, without prior written permission from the publisher.
Authorization to photocopy items for internal or personal use is granted by KoninklijkeBrillNV
provided that the appropriate fees are paid directly to The Copyright Clearance Center,
222 Rosewood Drive, Suite 910, Danvers, MA 01923, USA.
Fees are subject to change.
CONTENTS
Notes on Previous Publicationvii
List of Contributors....................................................................................ix
Introduction 1
part I
Practical Reason
Standards, Advice, and Practical Reason11
Chrisoula Andreou
Does Rationality Consist in Responding Correctly to
Reasons?25
John Broome
Practical Reason, Value and Action57
Alison Hills
Normativity and Practical Judgement77
Onora ONeill
part II
Particularism
Ethics Without Reasons?95
Roger Crisp
Defending the Right107
Jonathan Dancy
part III
Moral Realism
Moral and Theological Realism: The Explanatory Argument125
Russ Shafer-Landau
Anti-Reductionism and Supervenience147
Michael Ridge
vi
contents
part IV
Virtue Ethics
Han Feizis Criticism of Confucianism and Its Implications
for Virtue Ethics173
Eric L. Hutton
Aristotelian Virtue and the Interpersonal Aspect of Ethical
Character207
Maria W. Merritt
Good but Not Required?Assessing the Demands
of Kantian Ethics237
Jens Timmermann
Virtue, Character and Situation259
Jonathan Webber
part V
viii
List of Contributors
Chrisoula Andreou is Associate Professor in Philosophy at the
University of Utah.
Thom Brooks is Reader in Political and Legal Philosophy at Newcastle
University. He is editor and founder of the Journal of Moral Philosophy.
John Broome is Whites Professor of Moral Philosophy and Fellow of
Corpus Christi College at the University of Oxford.
Roger Crisp is Professor of Philosophy and Fellow of St Annes College
at the University of Oxford.
Jonathan Dancy is Professor of Philosophy at both the University of
Reading and the University of Texas-Austin.
Timothy Hall is Associate Professor in Philosophy at Oberlin College.
Alison Hills is Lecturer in Philosophy and Fellow of St Johns College at
the University of Oxford.
Eric Hutton is Associate Professor in Philosophy at the University of
Utah.
S. Matthew Liao is Associate Professor in the Center for Bioethics at
New York University.
Maria W. Merritt is Assistant Professor at the Johns Hopkins Berman
Institute of Bioethics.
Onora ONeill is Professor Emeritus in Philosophy and former Principle
of Newham College at the University of Cambridge. Baroness ONeill
of Bengarve is a Crossbench Peer in the House of Lords.
Martin Peterson is Associate Professor in Philosophy at Eindhoven
University of Technology.
list of contributors
Introduction
Thom Brooks
Ethics and Moral Philosophy
Ethics and moral philosophy is an area of particular interest today. This
book brings together some of the most important essays in this area.
The essays have all appeared recently in the Journal of Moral Philosophy,
an internationally recognized leading philosophy journal. This book is
divided into five parts: practical reason, particularism, moral realism,
virtue ethics, and ethics and moral philosophy more generally. My
introduction will give some background to these areas and introduce
the essays in this book.
Part I: Practical Reason
The first four essays focus on the topic of practical reason. Chrisoula
Andreou (2006) asks the question: is there a mode of sincere advice in
which the standards of the adviser are put aside in favor of the standards of the advisee? Andreou argues that the advisers standards should
not be set aside when offering sincere advice to others. Her claim is that
it is a mistake to hold that practical reason judgements evaluate actions
from within the agents system of standards as opposed to the system of
the judgers.
The second essay is by John Broome (2007).1 He asks whether rationality consists in responding correctly to reasons. Many philosophers
answer this question in the affirmative: rationality does consist in
responding correctly to either reasons or beliefs about reasons. Broome
argues that this widespread understanding is wrong. Rationality is not
about correctly responding to reasons or beliefs about reasons. Instead,
I am rational where I respond correctly to my beliefs about reasons.
1
This essay was originally published in a special issue on Metaethics arising from a
special conference at the University of Cambridge organized by Fabian Freyenhagen
and Hallvard Lillehammer. Essays included in this book include Broome (2007), Hills
(2007), ONeill (2007), Ridge (2007), and Shafer-Landau (2007).
thom brooks
This issue was Journal of Moral Philosophy volume 4, number 1 published in 2007.
introduction3
project that move beyond his earlier work into further interesting
domains.
Part III: Moral Realism
Part three concerns moral realism and an indirect debate between Russ
Shafer-Landau and Michael Ridge. Both essays were originally presented together in a panel at Journal of Moral Philosophy co-sponsored
conference on Metaethics.3 Shafer-Landau addresses sceptical arguments against theological realism and moral realism. He turns his focus
to an explanatory argument: moral facts lack causal-explanatory power
and, thus, we have most reason to deny the existence of moral facts.
Shafer-Landau argues that there is no parity between these realisms.
The vulnerabilities of one are rather different from the other and the
theological realist has the greater vulnerability. Moral realism may be
in a more safe position than we initially thought.
Michael Ridge does not directly address this particular line of argument, but he does focus on Shafer-Landaus then current work on
moral realism. Ridge focusses on Shafer-Landaus attempts to explain
supervenience with reference to his mora realism. Ridge argues that
these attempts are unsuccessful. Moreover, there are several implausible conclusions that follow from it. Ridge argues that we must have
greater caution in applying explanations for supervenience in the area
of the philosophy of mind in other contexts, such as metaethics.
Part IV: Virtue Ethics
The fourth part includes three chapters on virtue ethics. The first is
very different from the other two. This part begins with a fascinating
and highly illuminating essay by Eric Hutton. Hutton (2008) explores
the implications for contemporary (and largely Western) work in the
area of virtue ethics arising from Han Feizis critique of Confucianism.
This chapter will be of great interest to those interested in virtue ethics,
comparative philosophy, Chinese philosophy, and the non-Western
thought more generally.4 Confucius has been seen to offer a theory of
See footnote 1.
It has long been a goal of the Journal of Moral Philosophy to offer a mainstream
venue for work in non-Western philosophy. We continue to work arduously to attract
3
4
thom brooks
virtue ethics. Hutton considers the work of Han Feizi, an early critic.
Huttons project is to investigate whether these criticisms of Confucius
and his theory of virtue ethics offer revealing criticisms of virtue ethics
as understood today. Hutton answers in the affirmative and he argues
that this work helps highlight a new problem overlooked by current
virtue ethicists.
The next two chapters focus on character and virtue ethics. Maria
Merritt (2009) examines whether the Aristotelian idea of character as
firm and unchanging relates well to the empirical reality. She argues
that we should focus our attention primarily on substantive ethical
considerations and not characterological assessments. JonathanWebber
(2006) also considers our character in light of new developments in
experimental psychology.5 He argues that the traditional account has
much to recommend it. Experimental psychologists assume a behavioralist construal of traits in terms of stimulus and response rather than
inclinations. The traditional account has greater predictive power and
rests on a more defensible foundation.
Part V: Ethics and Moral Philosophy
This final fifth part brings together seven chapters that stretch
across ethics and moral philosophy more broadly. Timothy Hall (2008)
examines the oft-claimed moral distinction made by many non-
consequentialists between doing and allowing harm. Hall introduces a
third distinction, namely, denying resources. He argues that this third
distinction is separate from the first two and an important feature that
non-consequentialists should take into account.
The next two chapters are by S. Matthew Liao. First, Liao (2997)
considers the concept of time-relative interest originally introduced by
Jeff McMahan. This concept is meant to help us solve certain puzzles
about the badness of death, but it has also been used to explain how
abortion is permissible. Liao challenges this finding on a number of
grounds and argues that the justification of abortion must be found on
different grounds. In his second chapter included here, Liao (2010)
turns our focus to the question of what it is about our moral status as
new submissions in this area and it is an area of great personal interest as well (see
Brooks 2002a).
5
Webbers essay was the winner of our only prize for best article published in the
Journal of Moral Philosophy for 2006.
introduction5
human beings that gives this status such importance. Is this merely
speciesism? Liao argues in favour of what he calls the genetic basis for
moral agency as an account of rightholding. This account has the
attraction of accounting for all human beings without advocating some
form of speciesism.
Martin Peterson (2009) examines the number problem (see Brooks
2002b). The number problem is how to best decide whether to save
group x or y where these groups lack no morally relevant differences
and you may only choose one or the other. Different options are considered: we should save the greatest number, we should toss a fair coin
or we should set up a weighted lottery. These are all jettisoned in favour
of a new mixed solution where both fairness and the number of persons to be potentially saved count.
The following chapter is about our duties to those in severe poverty.
William Sin (2010) considers what follows if persons in affluent countries could easily save those in severe poverty in other countries. Sin
argues that accepting a principle of easy rescue whereby it is wrong for
these people not to provide easy rescue may lead to a problem of
demandingness. In such cases, what may first appear to be trivial sacrifices may result in great demands.
Alison Stone (2004) examines the essentialism debates of feminists.
Are there properties that are essential to women which all women
share? Stone argues against essentialism and instead favours the view
that women share a genealogy: women acquire feminity by appropriating and reworking existing cultural interpretations of feminity. This
anti-essentialist position is not only more sound than essentialism, but
it may also be more easily reconciled with feminist politics.
The final essay is on the demands of Kantian ethics. Jens Timmermann
(2005) assesses the alleged demandingness of Kantian ethics concerning our duties in a highly illuminating essay that defends Kantian ethics from several criticisms made of it.
Conclusion
A final comment should be made about how these essays were originally selected for publication. All chapters originally appeared in recent
issues of the Journal of Moral Philosophy. I founded the journal with
Fabian Freyenhagen in 2003 and it launched the following year. The
JMP has been published quarterly since 2009. All submissions were
thom brooks
subjected to rigorous anonymous peer review by our international editorial board and referees. We publish a list of our referees in the final
issue of each annual volume. Our standards are high and our acceptance rate is less than 7%. We receive submissions from all over the
world with the great majority (about two-thirds) coming from the
United States or Great Britain. Furthermore, we have endeavoured to
ensure high quality review standards with swift turnarounds and we
are normally able to vet about 90% of all submissions in three months
with about 85% reviewed in two months or less.
The essays in this volume have satisfied our high standards and all
have appeared in recent issues of the Journal of Moral Philosophy ensuring their quality and their timeliness. Each was selected for this book
because of its real contribution to the general topic. Rather than offer a
single narrative, the essays instead may be read in any order and present a number of important perspectives and insights on the topic that
should help provide further clarifying illumination on central debates
and ideas in the field. Readers interested in learning more about what
other essays have been published in the JMP should consult this books
bibliography towards the end where the full publication details of all
articles is listed. Anyone interested in submitting new work for future
issues should submit through our online submission system found on
our website.6
This book is amongst the first in our new Studies in Moral Philosophy
book series to be published by Brill, the publishers of the JMP. The
book series aspires to fulfil the same high standards and quality of its
sister journal. We will aim to publish leading work in the areas of moral,
political, and legal philosophy. This book helps launch this new series.
I hope you will enjoy reading the essays in this book as much as
Ihave. Many thanks are due to Suzanne Mekking at Brill for her strong
support of this book and new book series. I am also most grateful to
Fabian Freyenhagen for his assistance over the years, as well as to our
advisory committee for this book series. Finally, my most sincere
thanks to the authors for choosing the Journal of Moral Philosophy for
their work. The journals success is primarily not through any particular effort of its editors, but rather a reflection of the high quality of our
authors contributions. We look forward to producing further volumes
on other topics.
6
The Journal of Moral Philosophy website may be found here: http://www.brill.nl/
jmp.
introduction7
Bibliography
Andreou, Chrisoula (2006). Standards, Advice, and Practical Reason, Journal of Moral
Philosophy 3(1): 5767.
Broome, John (2007). Does Rationality Consist in Responding Correctly to Reasons?
Journal of Moral Philosophy 4(3): 34974.
Brooks, Thom (2002a). In Search of Shiva: Mahadeviyakkas Virashaivism, Asian
Philosophy 12(1): 2134.
Brooks, Thom (2002b). Saving the Greatest Number, Logique et Analyse 45(177/178):
5559.
Crisp, Roger (2007). Ethics Without Reasons? Journal of Moral Philosophy 4(1):
4049.
Dancy, Jonathan (2007). Defending the Right, Journal of Moral Philosophy 4(1): 8598.
Hall, Timothy (2008). Doing Harm, Allowing Harm, and Denying Resources, Journal
of Moral Philosophy 5(1): 5076.
Hills, Alison (2007). Practical Reason, Value and Action, Journal of Moral Philosophy
4(3): 37592.
Hutton, Eric (2008). Han Feizis Criticism of Confucianism and Its Implications for
Virtue Ethics, Journal of Moral Philosophy 5(3): 42353.
Liao, S. Matthew (2007). Time-Relative Interests and Abortion, Journal of Moral
Philosophy 4(2): 24256.
Liao, S. Matthew (2010). The Basis of Human Moral Status, Journal of Moral Philosophy
7(2) (2010): 15979.
Merritt, Maria W. (2009). Aristotelian Virtue and the Interpersonal Aspect of Ethical
Character, Journal of Moral Philosophy 6(1): 2349.
ONeill, Onora (2007). Normativity and Practical Judgement, Journal of Moral
Philosophy 4(3): 393405.
Peterson, Martin (2009). The Mixed Solution to the Number Problem, Journal of
Moral Philosophy 6(2): 16677.
Ridge, Michael (2007). Anti-Reductionism and Supervenience, Journal of Moral
Philosophy 4(3): 33048.
Shafer-Landau, Russ (2007). Moral and Theological Realism: The Explanatory
Argument, Journal of Moral Philosophy 4(3): 31129.
Sin, William (2010). Trivial Sacrifices, Great Demands, Journal of Moral Philosophy
7(1): 315.
Stone, Alison (2004). Essentialism and Anti-Essentialism in Feminist Philosophy,
Journal of Moral Philosophy 1(2): 13553.
Timmermann, Jens (2005). Good but Not Required?Assessing the Demands of
Kantian Ethics, Journal of Moral Philosophy 2(1): 927.
Webber, Jonathan (2006). Virtue, Character and Situation, Journal of Moral Philosophy
3(2): 193213.
part I
Practical Reason
* My thanks to Noell Birondo, Sarah Buss, Earl Conee, John Dreher, Stephen Finlay,
Bruce Landesman, Elijah Millgram, Russ Shafer-Landau, Cynthia Stark, Mariam
Thalos, Mike White, and two anonymous referees for valuable comments on earlier
versions of this paper. Some of the ideas in this paper were presented at the Eastern
Division APA meeting in December 2002 and at the Pacific Division APA meeting in
March 2004. I appreciate the helpful feedback I received during these meetings.
1
Plato, Republic (trans. G.M.A. Grube; Indianapolis: Hackett Publishing Co., 1986),
589e.
12
chrisoula andreou
This thesiswhich is taken for granted by many, including, for example, Bernard Williams, Gilbert Harman, and Donald Hubin3might
seem quite plausible. Indeed, it might seem as though the following
even stronger thesis is warranted:
Thesis 2: All genuine and sincere advice is such that, in cases of conflict,
the standards of the adviser are put aside and actions are evaluated from
within the advisees system of standards.
Though there is, I grant, some plausibility to the idea that giving
genuine and sincere advice involves seeing things from the advisees
perspective rather than ones own, thesis 2 is false. Indeed, even thesis
1which is significantly weaker than thesis 2, and which is taken for
granted by Williams, Harman, and Hubinis, I will argue, untenable.
I will first directly address two sorts of cases in which it might appear
as though an adviser is evaluating things from within the advisees system of standards even though this system conflicts with her own; such
cases are, I think, best interpreted in ways that dissolve this appearance.
I will then argue that the nature of sincere advice precludes an advisers
putting aside her own system of standards in favor of a competing system of standards.
After defending my position, I will consider its implications for a
popular Humean view of practical reason. According to this Humean
view, practical reason judgmentswhich are concerned with what an
agent has reason to do or not to do and which can figure as advice
evaluate actions from within the agents (as opposed to the judgers)
3
See Bernard Williams, Internal Reasons and the Obscurity of Blame, in
B. Williams, Making Sense of Humanity (Cambridge: Cambridge University Press,
1995), pp. 3545; Harman, Moral Relativism; and Donald Hubin, The Groundless
Normativity of Instrumental Rationality, Journal of Philosophy 98.9 (2001),
pp. 44568.
14
chrisoula andreou
Moes intention is akratic or not), Jen can affirm that if Moe intends to
make his spouse jealous, he should flirt with the video store manager.
Her judgment is thus based on Moes intention, rather than on either of
their standards, which are, at least for the moment, being put aside.
Intention-based judgmentslike Jens judgment that Moe should
flirt with the video store managerfigure prominently in our interactions with others. When I tell the operator that I am trying to call
Greece and ask her what I should do, she does not, before answering,
inquire into the reason for my call so as to determine whether or how
well my intention fits with my standards. She simply provides me with
the information I need to carry out my intention. Similarly, when I tell
my financial adviser that I want to liquidate some funds in order to buy
a Corvette, but also want, as much as possible, to avoid paying penalties, she does not, before advising me, inquire into the reason for my
proposed auto purchase so as to determine whether or how well my
intention fits with my standards. Again, she just proceeds to provide
me with the information I need to carry out my intention. (She also
provides me with information concerning how carrying out this intention will affect my prospects of carrying out other intentions I have
expressed to her; but this is still a far cry from making an evaluative
judgment from within my standards.) We have, so far, no support for
the thesis that there is a mode of sincere advice in which the evaluation
is made from within the advisees system of standards.
Advice where a Restricted Set of Options is under Consideration
Suppose a friend who is against abortion but is also not ready to become
a mother finds out she is pregnant and, fearful and distressed, comes to
you for advice. After reflecting on your friends casetaking into
account, among other things, how the various available alternatives
would affect those involvedyou come to the conclusion that your
friend should have an abortion. You know, however, that your friend
would never have an abortion. For her, the only conceivable options are
to have the baby and raise it herself or to have the baby and give it up
for adoption. So, when your friend asks you what she should do, you
tell her that she should give the baby up for adoption. Though sincere,
this advice seems to conflict with your standards, and so it might be
assumed that the advice is based on your friends standards. But there
is really no conflict between your advice and your standards. To the
contrary, your advice reflects your standards. More specifically, it
16
chrisoula andreou
reflects the fact that your standards favor the following ranking:
(1) your friend has an abortion; (2) your friend has the baby and then
gives it up for adoption; (3) your friend has the baby and then raises it
herself. Your advice implicitly involves a contrast between options (2)
and (3), which are the restricted set of options under consideration.
Given that your friend would never consider having an abortion but
has not yet formed any intention about what to do with the baby once
she has it, your advice focuses on damage control or on making the
best of things. Such advice is still derived from your standards and so
we do not have a case in which the standards of the adviser are put
aside in favor of the standards of the advisee.
My approach thus far has been defensive. In the next section, I will
offer a positive argument in favor of my contention that there is no
mode of sincere advice in which the adviser puts aside her own system
of standards in favor of a competing system of standards.
3. Why Advisers Must Favor Their Own Standards
It might be suggested that even if I could conclusively show that the
evaluative judgments we, as individuals, offer as advice are never based
on the standards of the advisee, this may simply reflect a defect on our
part, rather than something about the nature of advice. Maybe we are
substandard advisers, self-centered and resistant to taking the perspective of others even when this is called for. In this section, I will argue
that the nature of sincere advice precludes an advisers putting aside her
own system of standards in favor of a competing system of standards.
Suppose a potential adviser is being asked for some advice. Suppose
further that the request is not for intention-based advice, and that there
is a question as to whether the adviser is to use her own system of
standards or the advisees competing system of standards to arrive at an
all-things-considered evaluative judgment that can serve as sincere
advice. If, on the one hand, the adviser is a cognitivist about her standards, then she accepts her standards as true. It thus makes sense for the
adviser to rely on her own standards in making an evaluative judgment. This is because, in cases of conflict, a sincere adviser relies on her
own beliefs rather than on the beliefs of the advisee, as is illustrated by
the following example (in which the conflict is between what are
uncontrovertibly beliefs): Suppose that, according to both Nico and
Sarah (who may have shared standards for evaluating actions), Nico
should really relax and enjoy himself for a few hours. Suppose further
that, while Nico believes that going to a bar for a couple of drinks would
be both relaxing and enjoyable for him, Sarah thinks Nico is mistaken
about this. Her view is that if Nico goes to a bar for a couple of drinks,
he will actually end up stressed and depressed. Clearly, Sarah will, if she
is being sincere, advise Nico against going to a bar for a couple of
drinks. Though she can agree that, relative to Nicos beliefs, Nico should
go to a bar for a couple of drinks, she cannot sincerely provide Nico
with advice that conflicts with her own beliefs. (Of course, if she is
unable to convince Nico that he is mistaken, she can sincerely tell him
to do what he thinks he should do, on the basis of her view that Nico
should not just blindly follow her advice.5)
If, on the other hand, the adviser is a non-cognitivist about her
standards and thinks of judgments derived from an individuals standards as essentially expressive, then again the adviser will have to rely on
her own standards in arriving at a sincere all-things-considered evaluative judgment. For a sincere expressive judgment will aim to convey
the judgers own attitudes, and so will be derived from the judgers
standards, not from the standards of the advisee. So whether the adviser
is a cognitivist or a non-cognitivist about her standards, it makes sense
that her all-things-considered evaluative judgment will be derived
from her own standards.
Of course, an adviser can still make judgments like According to As
standards, A should (all-things-considered) X, which she can affirm
even if her evaluative judgment is that A should (all-things-considered)
not X. But, unlike the judgment A should (all-things-considered) X,
the judgment According to As standards, A should (all-thingsconsidered) X is about a system of standards, rather than being an
evaluation derived from within a system of standards. My conclusion is
that the nature of sincere advice precludes an advisers putting aside her
5
Do what you think you should do can be interpreted as derived from within
Sarahs system of standards for evaluating agents (rather than from within Nicos potentially competing system of standards for evaluating agents, or from within Sarahs system of standards for evaluating actions). My paper is concerned with judgments
derived from standards for evaluating actions. In any case, in saying Do what you
think you should do, Sarah is not putting aside her own system of standards for evaluating agents in favor of Nicos potentially competing system of standards for evaluating
agents. (Nicos system of standards for evaluating agents might require that, in cases of
conflict, an agent should faithfully follow the dictates of a certain authority rather than
do what he thinks he should do.)
18
chrisoula andreou
own standards and evaluating things from within the advisees system
of standards.
Clearly then, when it comes to seeking advice other than intentionbased advice, there is something to be said for seeking out someone not
only sincere, knowledgeable, and clear-headed, but also someone who
largely shares ones own standards, or at least the standards that are
likely to be relevant given the issue at hand. Such a search resembles the
search for a potential true friend, given Aristotles conception of friendship. For, the true friend is, according to Aristotle, another self .6 More
specifically, a true friend is not only someone whose intentions one
trusts, but also someone whose judgment one trusts, and so someone
who at least generally approve[s] of the same things.7 With such an
individual, one is, according to Aristotle, more able both to think and
to act, since one is better guarded against temptations and errors.8
Atrue friend is prone to encourage one to reason well and to be steadfast when one is weak-willed, since she will not want her friend to go
wrong; she might also, in some cases, be better positioned to reflect on
a particular situation.
One might, of course, be interested in having ones standards challenged. In such cases, it might also make sense to seek out someone
who, though she often approves of the same things, does not share all
of ones standards, and, in particular, does not share the standards that
are relevant given the issue at hand. Considering a defense of standards
that one does not share, especially if the defense appeals at least in part
to standards that one does share, is sometimes a sensible strategy for
improving ones ability to think and act. This strategy is, however, dangerous in cases where one is tempted to put aside ones standards
because they are proving to be quite demanding; for, in such cases, one
might be prone to act on rationalizations one is not really convinced by.
There is also the possibility that part of the reason one is asking a
particular person for advice is because one is interested in getting
aclearer picture of this persons standards, perhaps in order to get a
clearer picture of how this person would be likely to react to an action
one is contemplating. Though the point of this might be to open ones
standards up to challenges, it need not be. For the person one is asking
6
Aristotle, Nicomachean Ethics [cited hereafter as NE], trans. David Ross, rev. J.L.
Ackrill and J.O. Urmson (Oxford: Oxford University Press, 1998), IX.4, 1166a3132.
7
Aristotle, NE, IX.3, 1165b2729.
8
Aristotle, NE, VIII.1, 1155a1516.
for advice need not be someone whose judgment one trusts; she might,
for example, simply be someone with power over ones prospects.
If the nature of sincere advice precludes an advisers putting aside
her own standards in favor of the standards of the advisee, then in what
sense does being a good adviser require seeing things from the perspective of the advisee? My reasoning suggests that one respect in
which the good adviser will see things from the perspective of the advisee is that the good adviser will not ignore the decisions that the advisee has already made. For, a good adviser will, at least typically, focus
on providing advice that is not bound to be ignored. She will thus need
to know what options the advisee is willing to consider, and the decisions that an advisee has already made are relevant for determining
this. Also, an advisees decisions determine what intention-based
advice is applicable. There is arguably a further respect in which the
good adviser will see things from the perspective of the advisee. For
there is a great deal of plausibility to the idea that the good adviser will,
in some cases, base her advice on the agents likes and dislikes. But all
this is consistent with the advisers favoring her own standards.
Note that I am not committed to the view that a good adviser is willing to impose her standards on others. To the contrary, my view is consistent with the idea that a good adviser will never attempt to pressure,
let alone compel, others to act in accordance with her standards, even
if she is convinced that her standards are correct. For my view is consistent with the idea that a good adviser will recognize her fallibility
and respect the autonomy of others. As John Rawls stresses, even if
there is an objectively or intersubjectively valid set of standards of evaluation and one sees ones standards as capturing ones views concerning these valid standards, one can still, in light of the burdens of
judgment, acknowledge a reasonable pluralism of standards based on
reasonable disagreement.9
4. Standards, Advice, and Practical Reason Judgments
If my reasoning in this paper is correct, then there is no room for
thepopular Humean view of practical reason that is accepted by, for
example, Bernard Williams, Gilbert Harman, and Donald Hubin.10
9
John Rawls, Justice as Fairness: A Restatement (Cambridge, MA: Belknap Press,
2001), pp. 3537.
10
See Williams, Internal Reasons and the Obscurity of Blame; Harman, Moral
Relativism; and Hubin, The Groundless Normativity of Instrumental Rationality.
20
chrisoula andreou
For those who are not opposed to this idea, the popular Humean view
might seem plausible for several reasons. First, the view allows that, as
is commonly taken for granted, the aptness of a practical reason judgment does not vary depending on who is making the judgment. This
follows from the fact that there is, according to the Humean view, a
single system of standards that is relevant for arriving at practical reason judgments, namely the agents system of standards. Second, the
In the words of proponents of the popular Humean view, practical reason judgments take the agents perspective on a situation (Williams, Internal Reasons and the
Obscurity of Blame, p. 36), illustratea relativity to the agents values (Harman,
Moral Relativism, p. 62), evaluate actions from within the agents system of norms
(Hubin, The Groundless Normativity of Instrumental Rationality, pp. 46566).
Note that the popular Humean view differs from simple instrumentalism, according
to which practical reasoning and practical reason judgments are not in order until the
agent has settled on some end. For a defense of this latter view, which is less popular
among Humeans, see Candace A. Voglers Reasonably Vicious (Cambridge, MA:
Harvard University Press, 2002), especially chapter 7. For a critique of the view, see
Christine Korsgaards The Normativity of Instrumental Reason, in Garrett Cullity and
Berys Gaut (eds.), Ethics and Practical Reason (Oxford: Clarendon Press, 1997),
pp. 21554.
Note also that it is common ground that the popular Humean view is not Humes
view. There is, nonetheless, a great deal of controversy concerning how Humes position, particularly in the Treatise, is best interpreted. For two very interesting discussions concerning this issue see Elijah Millgrams Was Hume a Humean?, Hume Studies
21.1 (1995), pp. 7593, and Nicholas Sturgeons Hume on Reason and Passion, in
Donald Ainslie (ed.), Humes Treatise: A Critical Guide (Cambridge: Cambridge
University Press, forthcoming).
11
popular Humean view can account for the connection between accepting a practical reason judgment and being at least somewhat motivated
to act accordingly. Since the standards that are relevant for arriving at
practical reason judgments are the agents, and since the agents standards are essentially expressive, it makes sense that, barring akrasia,
agents are motivated to act in accordance with practical reason judgments that they accept. Finally, unlike the view that an agents reasons
are determined by the agents felt inclinations, which is explicitly
rejected by prominent Humeans as overly-simple, the popular Humean
view recognizes that, when there is a conflict between our standardsand our felt inclinations, it is our standards, rather than our felt
inclinations, that we are bound to see as necessarily relevant in evaluating our own actions.
Despite these potential virtues, the popular Humean view is untenable because it conflicts with the fact that the nature of sincere advice
precludes an advisers putting aside her own system of standards in
favor of a competing system of standards. As evaluative judgments that
can figure as sincere advice (and, more specifically, that do figure as
sincere advice when sincerely expressed in the form You should X),
practical reason judgments cannot be derived from a system of standards that conflicts with the judgers system of standards. Given that,
when agent and judger are not one and the same, the agents system of
standards can conflict with the judgers system of standards, it follows
that the idea that practical reason judgments evaluate actions from
within the agents system of standards is unacceptable.
It might seem as though a small alteration can save the popular
Humean view. The Humean can abandon the idea that practical reason
judgments are evaluations derived from within a set of standards, and
maintain instead that practical reason judgments either amount to or
derive from judgments about the agents standards. The Humean can,
more specifically, maintain that the practical reason judgment A should
(all-things-considered) X amounts to or derives from the judgment
According to As standards, A should (all-things-considered) X. If
practical reason judgments amount to or derive from judgments
about the agents standards, then a judger could affirm the practical
reason judgment A should (all-things-considered) X even if her
competing standards support the evaluative judgment A should (allthings-considered) not X. A judger could, for example, affirm the practical reason judgment Al should (all-things-considered) eat the veal
that has been placed in front of him by his host even if the judgers
22
chrisoula andreou
competing standards support the evaluative judgment Al should (allthings considered) not eat the veal that has been placed in front of him
by his host.
The view that practical reason judgments amount to or derive from
judgments about the agents standards is, however, problematic. This
becomes apparent via consideration of first-person practical deliberation. When, in practical deliberation, one concludes that one should
(all-things-considered) X, this practical reason judgment figures as an
evaluative judgment derived from within ones system of standards
rather than as a judgment about ones system of standards, or as a judgment that derives from a judgment about ones system of standards.
Consider, for example, the following straightforward bit of practical
deliberation:
Premise: Animals have the right to a decent quality of life.
Conclusion:
I should boycott factory-farm-produced meat.
The conclusion certainly seems to be derived from within a system of
standards. The premise, which figures as providing direct support for
the conclusion, could not figure in this way if the argument could be
rewritten in either of the following ways:
Premise: Animals have the right to a decent quality of life.
Conclusion: According to my standards, I should boycott
factory-farm-produced meat.
Premise: Animals have the right to a decent quality of life.
Conclusion 1: According to my standards, I should boycott
factory-farm-produced meat.
Conclusion 2:
I should boycott factory-farm-produced meat.
Unlike the judgment According to my standards, I should boycott factory-farm-produced meat, the judgment Animals have the right to a
decent quality of life is not a judgment about the agents standards. As
such, the latter judgment cannot provide direct support for the
former.
It might be suggested that the premise in the original bit of practical
deliberation must not be taken at face value, but must instead be interpreted as elliptical for the judgment According to my standards, animals have the right to a decent quality of life. But why should we accept
this suggestion? It seems dogmatic to insist that the agent cannot
takeher conclusion to be directly supported by a judgment about the
rights of animals, but must instead take it to be directly supported by a
judgment about herself. If, however, we allow that the agent can take
her conclusion to be directly supported by a judgment about the rights
of animals, then we must also allow that the agents practical reason
judgment is an evaluation derived from within her system of standards,
rather than a judgment that amounts to or derives from a judgment
about her system of standards. But then we cannot accept the suggestion that practical reason judgments amount to or derive from judgments about the agents standards. So the slightly altered Humean view
fares no better than the popular Humean view it was meant to replace.
5.Conclusion
In the bulk of this paper, I argued against the thesis that there is a mode
of sincere advice in which the standards of the adviser are put aside in
favor of the standards of the advisee. Though plausible-sounding, this
view is untenable. There are, I grant, cases in which it might seem as
though an adviser is evaluating things from within the advisees system
of standards even though this system conflicts with her own; but such
cases can, and indeed must, be interpreted in ways that dissolve this
appearance. For the nature of sincere advice precludes an advisers putting aside her own system of standards in favor of a competing system
of standards. It follows that a popular Humean view of practical reason
is untenable.
* Thanks to Jacob Busch, Olav Gjelsvik, Christine Korsgaard, Sven Nyholm, Derek
Parfit, Wlodek Rabinowicz and Andrew Reisner for very useful comments. Thanks
also to audiences at Cambridge, Harvard, Oslo, Oxford and St Andrews, who all contributed a great deal. My work on this paper was supported by a Major Research
Fellowship from the Lever-hulme Trust, to whom I am very grateful.
26
john broome
Another purpose of the paper is simply to explore a part of the relation between rationality and reasons. I shall look for what truth there is
in equivalence and its relatives. I shall give most attention to this one
side of equivalence, because it is nearest to the truth.
Entailment. Necessarily, if you are rational you respond correctly to
reasons.
28
john broome
30
john broome
3
Thomas Nagel pointed out to me that believing p at one time might constitute a
reason to believe p at a later time. The fact that you believe p one time constitutes, at a
later time, indirect evidence of p: you would probably not have believed p if you had
not had evidence for it. But in the claim I am objecting to, the beliefs are supposed to
be contemporaneous.
32
john broome
C3. Necessarily, if you are rational then, if you believe p and you believe
that if p then q, you believe q.
C4. Necessarily, if you are rational then, if you intend to F and you believe
you cannot F unless you G, you intend to G.
4
A point made by Gilbert Harman in Change in View (Cambridge, MA: MIT Press,
1986).
34
john broome
be the case that the conditions of rationality hold only because of their
corresponding attitudinal reasons. There might simply be a class of
attitudinal reasons like the ones I described, and rationality might
consist in responding correctly to them. The reasons might explain
the conditions, even though we identify the reasons through the
conditions.
So attitudinal reasons provide a potential defence of equivalence and
the reductionist idea that rationality consists in responding correctly
to reasons. However, I shall now argue that this defence fails. Even
restricted to attitudinal reasons, equivalence is false.
Your attitudinal reasons may conflict with each other; you may have
an attitudinal reason to F and an attitudinal reason not to F. I shall give
an example soon. When your reasons conflict, you respond correctly
to them so long as you do not go against their balance. Inevitably, even
if you respond correctly, you will not satisfy one or other of your conflicting reasons. But I explained that these reasons impose strict liability. That is to say, in failing to satisfy one of them you are inevitably
irrational. So, although you respond correctly to reasons, you are irrational. Therefore rationality is not equivalent to responding correctly
to attitudinal reasons.
Here is the argument more briefly. To be rational, you have to satisfy
some particular conditions. But attitudinal reasons can be overridden,
so responding correctly to them cannot guarantee that you satisfy any
particular condition. Therefore rationality is not equivalent to responding correctly to attitudinal reasons.
Next an example. Suppose you intend to G, and you believe you cannot G unless you H. According to R4 among the examples above, this
intention and belief together constitute an attitudinal reason for you to
intend to H. But suppose also that you intend not to H. According to
R2, this intention constitutes an attitudinal reason for you not to intend
to H. So you have an attitudinal reason to intend to H and another not
to intend to H.
To respond correctly to your reasons in this case, you must go with
the balance of your reasons. Let us say that the balance is in favour of
your not intending to H. Then you respond correctly to your reasons if
you do not intend to H. But you intend to G and you believe you cannot
G unless you H. Given that, the condition of rationality C4 set out
above tells us you are not rational. So you respond correctly to your
reasons but you are not rational. Therefore, rationality is not equivalent
to responding correctly to reasons.
36
john broome
But again, actually you may be responding correctly. This reason not
to intend to G may be opposed by another attitudinal reason. Perhaps
it is opposed by a belief that you ought to G. So far as I can see, this
could go on for ever.
I conclude that attitudinal reasons do not provide a successful
defence of equivalence. Entailment, restricted to attitudinal reasons,
may yet be true:
Necessarily, if you are rational, you respond correctly to attitudinal
reasons.
38
john broome
This cannot be generally true, for the following reason. Suppose you
believe your reasons require you to F, and you intend to F, but you are
prevented from Fing by something outside your mind. You may be
rational. For instance, suppose you believe your reasons require you to
vote for Gore, and you intend to do so, and you do all the appropriate
things, but surprisingly the machinery works in such a way that you
vote for Bush. You fail to do what you believe your reasons require you
to do, but you might nevertheless be rational.
This is another application of the principle that rationality supervenes on the mind, which I introduced in section 2. If the machinery
had worked differently but your mind had been just the same, you
would have voted for Gore. In that case, we may presume you would
have been rational, since voting for Gore is what you believe your reasons require you to do. Given the way the machinery actually works,
you vote for Bush. But the operation of the machinery cannot affect
your rationality. Since your mind is the same in either case, you are
rational in this case too.
The consequence is that the direct enkratic condition cannot be true
unless whether or not you F is entirely determined by the properties of
your mind. The condition cannot be true if Fing is a bodily act, or a
state such as living in Australia.
It might nevertheless be true in important cases. I shall come back to
those in section 5. But for cases where your Fing is not determined by
the properties of your mind, we need to modify the core condition
some more. We can make it:
Enkratic condition. Necessarily, if you are rational, then, if you believe
your reasons require you to F, you intend to F.
the scientist prevents you from being entirely rational. So this is not an
objection to the enkratic condition.
The enkratic condition says that akrasia is irrational. This is a common opinion, and one I shall accept without more ado. I treat this as a
concession, since I am about to show that it gives the best support I can
find to modified entailment. I do not think the condition as I have
stated it is exactly true. A more accurate version is:
Necessarily, if you are rational, then, if you believe your reasons require
you to F, and you believe you will F if and only if you intend to F, you
intend to F.
But for the purposes of this paper, I shall stick to the simplified version
I started with.
As a further concession, I am happy to strengthen the enkratic condition by adding an explanatory or counterfactual clause. We get (also
adding a universal quantifier):
Strengthened enkratic condition. Necessarily, if you are rational, then,
whenever you believe your reasons require you to F, you intend to F and
an appropriate explanatory or counterfactual connection holds between
your belief that your reasons require you to F and your intending to F.
40
john broome
5. Meeting Your Own Standards
7
See Graham Priest and Koji Tanaka, Paraconsistent Logic, in Stanford Encyclopedia
of Philosophy, http://plato.stanford.edu/.
8
This is the view of rationality presented by T.M. Scanlon in What We Owe to Each
Other (Cambridge, MA: Harvard University Press, 1998), pp. 1832.
42
john broome
These two formulae are equivalent if we interpret you respond correctly to beliefs about reasons to mean that, whenever you believe your
reasons require you to F, you F and an appropriate explanatory or
counterfactual connection holds between your belief that your reasons
require you to F and your Fing. So the strengthened direct enkratic
condition, like the strengthened enkratic condition, constitutes an
interpretation of modified entailment. It says that rationality entails
responding correctly to beliefs about reasons.
We can go further. We are pursuing the thought that the direct enkratic condition replaces all the separate conditions of rationality. If it
does, that one condition encompasses the whole of rationality.
Consequently, we could convert modified entailment from a one-way
conditional statement to a biconditional. We would arrive at modified
equivalence. We would have shown that rationality is equivalent to
responding correctly to beliefs about reasons. This sections particular
interpretation of this idea is that rationality is equivalent to meeting
your own standards.
I have two objections to this conclusion. The first is to the last step of
the argument: modified entailment cannot be extended to modified
equivalence as I have just suggested it can. The second is an objection
to the direct enkratic condition, which means it is an objection to modified entailment itself, interpreted as the strengthened direct enkratic
condition.
Objection to Modified Equivalence
For the moment, let us suppose the strengthened direct enkratic condition is true. I shall argue it cannot possibly be the only condition of
rationality. So modified equivalence does not follow.
The direct enkratic condition is supposed to replace all the various
more particular conditions of rationality that I mentioned. For instance,
instead of the condition that you do not have contradictory beliefs, we
have a particular instance of the direct enkratic condition: that you do
not have contradictory beliefs if you believe your reasons require you
not to have contradictory beliefs.
The particular conditions of rationality assert strict liability: you are
not rational if you do not conform to them, whatever you believe about
44
john broome
Suppose you do not believe God exists, but you believe your reasons
require you to believe he does exist. According to the direct enkratic
condition, restricted to states of mind, you are irrational. But that is not
necessarily so.
Suppose you believe there is no evidence of Gods existence, and that
is why you do not believe he exists. Indeed you are unable to believe he
exists. Nothing you could dogoing regularly to church, or taking a
course of religious instruction, or anything elsewould bring you to
believe he exists. However, you believe your reasons require you to
believe God exists, on grounds of personal safety. You believe it is a
good idea to believe he exists, just in case he does. You may be perfectly
rational in having these beliefs.
To be sure, your belief that you ought to believe God exists may be
false. At least two arguments can be raised against it. One comes from
evidentialism. Evidentialists think personal safety is no reason to
believe God exists; only evidence can constitute a reason. So they think
you are wrong to believe your reasons require you to believe God exists.
Still, even if evidentialism is true, you need not be irrational in having
a different opinion. Your belief that your reasons require you to believe
God exists will be false, but it need not be irrational.
A second argument comes from the principle that ought implies can.
Since you cannot believe God exists, perhaps it follows that it is not the
case that your reasons require you to believe God exists. Still, even if it
is true that ought implies can in this context, you need not be irrational
in believing the opposite. Once again, even if your belief is false, it need
not be irrational.
That shows the restricted version of the directed enkratic principle is
false. On the other hand, what is its attraction? Why should we think
you are necessarily irrational if you are not in a mental state that you
believe your reasons require you to be in? You would indeed be irrational if your mental states were controlled by your intentionsif, necessarily, whenever you intended to be in a mental state, you were in it.
The enkratic condition tells us you would necessarily be irrational if
you did not intend to be in a mental state you believed your reasons
required you to be in. Consequently, you would necessarily be irrational if you were not in it. But actually many of your mental states,
including many of your intentional attitudes, are not controlled by
your intentions.
We might be inclined to believe the direct enkratic condition even
so. We might think a rational person simply has a psychological
disposition to be in a mental state whenever she believes her reasons
46
john broome
This is a variant on the direct enkratic condition. But the formula has
to be interpreted carefully if it is to catch Scanlons meaning correctly.
We must give a particular interpretation to the clause if you believe
your object-given reasons require you to F.
One way to satisfy the condition expressed in this clause would be to
have a belief that you could express using a sentence of the form My
object-given reasons require me to F. But hardly anyone ever has that
sort of a belief, because hardly anyone has the concept of an objectgiven reason. Another way to satisfy the condition is by having a belief
that you could express using a sentence of the form X requires me to F,
where X (whether you know it or not) constitutes your object-given
reasons. For instance, you satisfy the condition if you have a belief that
you could express by the sentence The evidence requires me to believe
the climate is changing. A technical way of putting this is that believe
in the conditional clause sets up a referentially transparent context.
We can make the transparency explicit when Fing is believing or
intending. For believing and intending, Scanlons condition comes
down to:
Necessarily, if you are rational, then, if you believe your evidence requires
you to believe p, you believe p.
Necessarily, if you are rational, then, if you believe your reasons require
you to F, you intend to F.
The first of these formulae is plausible; I do not deny it. The second is
simply the enkratic condition, which I accept.
So Scanlons condition may well be true of beliefs and intentions.
It makes better psychological sense than the direct enkratic condition.
We can expect automatic, subpersonal processes to follow your beliefs
about object-given reasons rather than your beliefs about state-given
reasons. In my example, you believe it would be safest to believe God
48
john broome
I think Parfit is saying that rationality is equivalent to responding correctly to what I have called non-normative beliefs about reasons. This
section considers whether that is so.
Suppose you believe there is a tiger-snake on the path ahead, and
you respond by stopping. Do you respond correctly to your belief?
Probably not. If there is a tiger-snake on the path ahead, that is one
reason for you to stop. But probably you have other reasons either to
stop or not to stop. Perhaps you need to get back home, or perhaps you
are enjoying the exercise. All these reasons together determine whether
13
or not your reasons require you stop. There is therefore no such thing
as responding correctly to just one of them. You can respond correctly
only to all of them together; you do so by doing what all of them
together require. Consequently, we could not count you as responding
correctly to a non-normative belief in just one of the reasons.
Could you ever respond correctly to a non-normative belief about a
reason? The tiger-snake does not provide an example, but are there any
examples? Suppose you have a non-normative belief that a reason
exists, and suppose that reason, if it did exist, would be enough on its
own to make it the case that your reasons require you to F. I call a belief
of this sort a perfect reason. Then, if you F (and if an appropriate
explanatory or counterfactual connection holds between your belief
and your Fing), we could count you as responding correctly to your
belief.
An Example
Here is a plausible example of that sort. Suppose, gazing at the horizon
at night, you believe you see a red light. What you believe is that you see
a red light. Plausibly, if true, this would be a perfect reason to believe
you see a coloured light. Therefore, if you respond to your belief
thatyou see a red light by also believing you see a coloured light, plausibly you are responding correctly.
Moreover, it is also plausible that, unless you respond correctly in
this way, you are not rational. Necessarily, if you are rational you
respond correctly to this reason. Rationality entails responding correctly to it. I do not assert this is true; only that it is plausible. I cannot
think of any more convincing example. So for the sake of pursuing the
argument further, I shall assume it is true.
If it is, it illustrates another sort of strict liability. You are irrational if
you do not respond correctly to your non-normative belief about a reason, and this is so whatever your normative beliefs may be. Section 3
examined strict liability in responding to a reason. Here we have strict
liability in responding to a belief about a reason.
This raises a question. The claim is that you are strictly liable for
responding correctly to a non-normative belief of yours, whatever your
normative beliefs may be. But might you not have a normative belief
that imposes a conflicting liability? Is that a problem?
In the example, you have the non-normative belief that you see a red
light. I said that, plausibly, you are not rational unless you believe you
see a coloured light. But suppose you also have the normative belief
50
john broome
that your reasons require you not to believe you see a coloured light.
Perhaps you know you will be severely punished if you believe you see
a coloured light. According to the direct enkratic condition, you are
not rational if you believe you see a coloured light. So it seems you are
not rational whether or not you believe you see a coloured light. Is that
a problem?
It is not, because I have already rejected the direct enkratic condition
in section 5. This may be a case like the one where you believe your
reasons require you to believe in God. Suppose, given that you believe
you see a red light, nothing you can do would stop you from believing
you see a coloured light. Even if you believe your reasons require you
not to believe that, there need be nothing irrational about you if you do
believe you see a coloured light.
An alternative reaction to the difficulty would be say you must be
irrational anyway, because your normative belief is irrational. If you
can see a red light, you must be irrational in believing your reasons
require you not to believe you see a coloured light. But that is not credible. To be sure, according to evidentialism your normative belief is
false. According to evidentialism, your evidence provides all the reasons you can have for believing or not believing anything. In this case
you have conclusive evidence that you see a coloured light, so you have
a conclusive reason to believe you see a coloured light. The fact that you
will be punished for doing so provides no reason against it. But whether
or not evidentialism is true, it is debatable. We must not impugn a persons rationality just because she has a belief that conflicts with
evidentialism.
In any case, in one way or another, we can accept strict liability
for this example. Plausibly you can respond correctly to your nonnormative belief about a reason, and you are not rational unless you do.
Practical Reasons
The example of seeing a red light is contrived, and it is an example of a
reason to believe. Can we find an example among practical reasons
reasons to do something?
Before I try out an example, I need to make an adjustment to the
notion of responding correctly to a non-normative belief. Suppose you
have a non-normative belief whose content would, if true, be a reason
to do some act. Suppose the act is not a mental one. Then it could not
be the case that, necessarily, if you are rational you do this act. That
52
john broome
and necessarily, if you are rational and you have this belief, you intend
to do this thing. That is to say, rationality entails that you respond correctly to your belief.
This is another assertion of strict liability. You are not rational if you
fail to respond to your non-normative belief, whatever your own beliefs
may be about the normative situation. This time I am not convinced.
Suppose you have a normative belief that conflicts with your nonnormative one. Suppose you believe that your reasons require you not
to change your appointment. That might be because you have some
false, bizarre normative theory according to which you should avoid
pain on Wednesdays much more than on Tuesdays. Then the enkratic
condition entails that you are irrational unless you intend not to change
your appointment. But Parfit thinks you are irrational unless you
intend to change your appointment. We may take it for granted that
you are irrational if you have both intentions; since they contradict
each other, they violate C2 among the conditions of rationality listed in
section 3. So, whatever intentions you have, you are irrational.
What should we conclude? A similar problem came up with the red
light example. But there the conflict between your normative and your
nonnormative belief was generated by the direct enkratic condition.
Isolved it by rejecting that condition. In our present practical example,
the conflict is generated by the enkratic condition. The enkratic condition is on stronger ground, and I accept it.
We might instead claim you are irrational because of your bizarre
norma-tive theory. We might say it cannot be rational to hold such a
bizarre theory. But I see no grounds for that claim. There is no logical
inconsistency among your beliefs; your belief that your reasons require
you not to change your appointment is not logically inconsistent with
your non-normative belief in the big conjunction. Moreover, your
belief in a bizarre normative theory may be supported by all the evidence you have. Perhaps you acquired it during your upbringing by
testimony from people you had good reason to believe were reliable.
We are not entitled to impugn your rationality just because you hold a
false and bizarre normative theory.
I conclude that you may rationally intend not to change your
appointment, despite your non-normative belief whose content would,
if true, be a perfect reason to change it. The content would, if true,
make it the case that your reasons require you to change your appointment. But that is only because of the normative facts, and you do not
believe those normative facts. Instead, you have a normative belief that
implies your reasons require you not to change your appointment. You
may not be irrational in having this belief. Consequently, you may
intend not to change your appointment, and nevertheless be rational.
A parallel argument could be brought to bear on any non-normative
belief in a practical reason. You may be rational even if you fail to
respond correctly to it.
That need not be the end of the story. Derek Parfit has pointed out to
me a way of going further. The problem revealed in the example is that
your non-normative belief might conflict with a normative belief of
yours. If you respond correctly to your normative belief, you will not
respond correctly to your non-normative belief. That is why you may
be rational even if you do not respond correctly to your non-normative
belief. But suppose you have no contrary normative belief. Then you
have no grounds for failing to respond correctly to your nonnormative belief. In that case, we might conclude you are necessarily
irrational if you do not respond correctly to it.
That may be correct for Parfits particular example of pain on
Tuesday. In this example, it is very obvious that what you believe, if
true, would be a perfect reason for you to change your appointment.
Given that, perhaps it takes a normative belief (which would have to be
bizarre) to oppose its obviousness, and make it rational for you not to
change it. I have no firm view about this example.
However, the rule does not generalize. The question is this. Suppose
you have a non-normative belief in a reason. Suppose that what you
believe, if true, would be a perfect reason for you to F. Suppose you
have no opposing normative belief. Are you necessarily irrational if
you do not respond correctly to your belief? That cannot be so in general. It need not be obvious that what you believe constitutes a perfect
reason for you to F. It might be a perfect reason only because of some
fact that you do not believe and to which you have no easy access. In
that case, you may fail to F and yet be rational.15
For example, suppose you believe this liquid is orange juice. Suppose
that, if it is indeed orange juice, that is a perfect reason for you not to
drink it, because you have overnight become so allergic to orange juice
that drinking it will kill you. But suppose nothing has given you any
inkling of this allergys sudden onset. Then you might drink the liquid
and not be irrational in doing so.
15
This point is made by Andrew Reisner in Conflicts of Normativity (unpublished
DPhil thesis, Oxford University, 2004), ch. 2.
54
john broome
Conclusion
This section has investigated the idea that rationality is equivalent to
responding correctly to non-normative beliefs about reasons. These are
the conclusions that have emerged.
First, to many of your non-normative beliefs about reasons, there is
no such thing as responding correctly. You can respond correctly only
if what you believe would, if true, constitute a perfect reason. But many
of your nonnormative beliefs about reasons are about reasons that conflict with others. Your belief about a tiger-snake is an example.
Second, plausibly there are some non-normative beliefs about reasons such that, necessarily, if you are rational you respond correctly to
them. Your belief about seeing a red object is an example. This means
that, at a pinch, we may once again endorse modified entailment.
Ireproduce it here:
Modified entailment. Necessarily, if you are rational, you respond correctly to beliefs about reasons.
58
alison hills
2
J.D. Velleman, The Aim of Belief , in his The Possibility of Practical Reason (Oxford:
Oxford University Press, 2000).
In other words, practical reasons are reasons for action, but more precisely, they are reasons to perform actions that meet the standards of
success intrinsic to action. It is perhaps not obvious what it means to
have reasons to meet these standards, but I hope that this will become
clear through a number of examples. On the assumption that this intuitive connection holds, I will set out three different conceptions of
action and corresponding success conditions, and explain how each is
linked to a particular conception of practical reason and, in two cases,
to a theory of value too. My goal is to describe these different accounts,
rather than to defend any in particular, though I will suggest that some
are more satisfactory than the others.
3
See C.M. Korsgaard, The Locke Lectures (forthcoming, available online at:
http://www.-people.fas.harvard.edu/~korsgaar/) and J.D. Velleman, Practical Reflection
(Princeton: Princeton University Press, 1989), and his The Possibility of Practical
Reason. Korsgaard thinks that acknowledging these standards of success will leave us
in a position to answer the sceptic who challenges the authority of practical rationality.
Velleman hopes that this will turn out to solve the problem of how an agents practical
reasons are related to her desires.
4
In this paper, I discuss only practical reasons, rather than practical reasoning or
practical rationality. I do not expect the link between practical reasons and correct
practical reasoning or practical rationality to be straightforward, so the argument here
should not be taken to endorse any particular view about practical reasoning or practical rationality.
5
I am therefore assuming for the purposes of this paper that practical reasons are
essentially reasons to act, rather than reasons for emotions, desires or anything else.
We do often talk of reasons for wanting X or for feeling sad or angry. But here I will
assume that, insofar as these are practical reasons, they are reasons to take steps in
order to feel the relevant emotion or to acquire the pertinent desire: in other words,
they are reasons for action after all. If there are reasons simply to have a certain
emotion, for example, rather than to bring it about, they are of a very different kind
from the practical reasons described in this paper, and so, I will assume, deserve to be
treated differently and labelled differently too.
60
alison hills
2.Execution
There is something quite attractive about the idea that the basic evaluation of an action as a success or failure is in terms of whether the
agent achieved his aim or not. And there is also something appealing
about tying reasons for action to the success conditions of action. But
unfortunately the associated theory of practical reason does not fit at
all well with our intuitive grasp of what practical reasons we have. It is
simply not plausible that agents have practical reasons to satisfy their
desires, no matter how ill considered, nor is it plausible that they have
no reason to do anything that they do not already want to do.
Consider a famous example from Williams.8 Suppose that I think
that a bottle of clear colourless liquid contains gin, whereas in fact it is
full of petrol. If I pour myself a glass and drink it, is my action s uccessful?
I intended to get myself a drink from that bottle and I did. According
to the simple theory, my action is successful because I got what I aimed
at, and I had reason to do what I did. It is not very plausible, however,
that any consideration favoured my drinking petrol, nor that I had any
reason to do so.
There is a worse problem for the simple account, however. One of
my aims in acting was to get a gin and tonic, and I did not get one, so
I failed to achieve my aim. It appears to follow from the simple theory
that my action is a failure and also a success, that I had no reason to do
it and also that I had some reason to do it.
The simple theory appears, therefore, to be incoherent. But it
seems that whilst keeping to the spirit of that theory, we can modify it
to avoid this problem. The success of my action should not be judged
on the basis of an aim of mine (e.g. to drink from that bottle), which
is based on a false belief. The satisfaction of my other goal (to get a
gin and tonic), by contrast, is relevant to an evaluation of my actions
success. This accords with my own judgment of my drinking the petrol,
namely that my action was disastrous, and that I had no reason to
do it.9
8
B.A.O. Williams, Internal and External Reasons, in his Moral Luck (Cambridge:
Cambridge University Press, 1981).
9
There are other ways of restoring the coherence of the simple account. We could
claim that I performed more than one action, one successful (drinking from the bottle), one not (drinking a gin and tonic). This seems to multiply the number of actions
I performed more than is necessary or plausible, however. Alternatively, we could
claim that my action was successful in one respect and not in another. But this conflicts
with my own, very plausible judgment, that my drinking petrol can in no way be
described as a success.
62
alison hills
1.Reasons for action are essentially connected to the success conditions of action:
all reasons for action are reasons to perform successful actions.
2.An action is successful if it is the successful execution of the agents aims (broadly
conceived to include aims that are not based on false belief or ignorance).
64
alison hills
The problem is that we began with a quite plausible claim, that success in action was related to achieving your aim. But this linked practical reason to your current desires or goals: a successful action, and
what you had reason to do, was to get what you wanted. It was reasonable to ignore those of your actual desires based on false belief or ignorance: for it made sense to describe your action as failed when you
drank petrol instead of a gin and tonic. You, here and now, would probably class those actions as failed. But it does not make sense to call your
action a failure simply because you have not satisfied some desire that
you do not yet have. This simply does not fit with the conceptionof
action as execution. The only reason to make this modification is
because without it the associated conception of practical reason is
unacceptable. The modification is ad hoc, and on that basis should be
rejected. But without it the theory of practical reason is not adequate.
In addition, if we do say instead that practical reasons are reasons to
satisfy the desires that you would have if you had no false beliefs, and
you had all relevant information, we run into a well-known problem.12
On this account, all reasons for action are instrumental reasons to satisfy your aims. But if you really have a reason to take the means to your
aim, then it seems that you must have a reason to pursue that aim in the
first place (or at the very least, no reason not to). Otherwise, you could
acquire reasons to do stupid, pointless and immoral actions just by
forming the appropriate desires. If you find yourself with a desire for
which you have no reason, for example, a desire to kill yourself and
others, then you have no reason to take the means to satisfy it. Rather,
you have reason to get rid of the desire, if you can. The claim that there
are only instrumental reasons for action seems to be false.13
Discussed, amongst many others, by C.M. Korsgaard, The Normativity of
Instrumental Reason, in G. Cullity and B. Gaut (eds.), Ethics and Practical Reason
(Oxford: Oxford University Press, 1997), J. Broome, Normative Requirements, Ratio
12 (1999), pp. 398419; J. Broome, Normative Practical Reasoning, Proceedings of the
Aristotelian Society, Supplementary Volume 75 (2001), pp. 17593; and D. Enoch,
Agency, Schmagency: Why Normativity Wont Come from What is Constitutive of
Agency, Philosophical Review 115 (2006), pp. 199241.
13
The best response to this objection may be to insist that the problem itself is incoherent. Practical reasons just are reasons related to success in action, where this is the
successful execution of the agents aims. Asking questions about whether or not we
have practical reason to have the aims that we in fact have is simply a mistake. Compare
the parallel argument about belief. Epistemic reasons relate to the correctness conditions of belief; in the crudest terms, they are reasons to have true beliefs rather than
false. Someone might well ask you whether you have reason to form propositional
attitudes that aim at the truth: what reason do you have to form beliefs? There may be
12
66
alison hills
The appeal of action as production is perhaps the reason why utilitarianism (or more generally consequentialism) is perennially popular.15 If action aims at producing good results, then as an agent, the
fundamental way in which you should respond to value is by p
roducing
more of it (thereby acting successfully). So some maximizing version of
consequentialism seems to be inevitable, for example, utilitarianism
(everyone has reason to maximize happiness) or egoism (everyone has
reason to promote her own interests).16
There are some benefits to thinking about successful action as the
production of some good, and practical reasons as reasons to promote
the good. For example, it is not true that all practical reasons are instrumental reasons, so the problems that afflict theories of practical reason
that are purely instrumental do not apply. The question why be an
agent? can be answered. According to action as production, there are
certain standards that you must meet if your action is to be successful,
bad. Moreover, an agent can successfully produce good results without aiming at anything in particular. So the aim of producing good results (or even of producing
any results) cannot be attributed to the agent herself. This is a problem for the theory, since it is not obvious how else to understand the idea that actions aim at
production.
15
This might be a friendly amendment to Foots claim that the intuition behind
consequentialism is that it can never be right to prefer a worse state of affairs to a better. But whereas most people do not think at all in terms of states of affairs, it is more
plausible that they do think of actions as aiming to produce good results (P. Foot,
Utilitarianism and the Virtues, in her Moral Dilemmas [Oxford: Oxford University
Press, 2002]).
16
The picture is so seductive that it influences even Velleman, who initially seems to
be trying to give a rather different account, whereby the constitutive aim of action
is selfknowledge (Velleman, The Possibility of Practical Reason, pp. 2024; Practical
Reflection). Action, or perhaps autonomous action, is distinguished from mere behaviour, according to Velleman, in that an agent is in conscious control of his action
whereas one is not similarly self-aware in other forms of behaviour (e.g. in Freudian
slips). This sounds quite different from the idea of action as production. But in order to
derive a theory of practical reason from this constitutive aim, Velleman expands his
story so that you have most reason to do what it would make most sense for you to do.
For example, it might make most sense for you to cook dinner, since you are hungry
and reasonably enjoy cooking, so that is what you have most reason to do (Practical
Reflection, pp. 198ff.). In order to derive a full theory of practical reason from the minimal idea that action involves self-knowledge, he resorts to action as the production of
self-knowledge. And it is here that he goes wrong, because as Setiya points out, even if
self-knowledge is in some sense characteristic of autonomous action (or acting for reasons) it is not in the sense that action can produce more or less of it, and we have reason
to produce more. See K. Setiya, Explaining Action, Philosophical Review 112 (2003),
pp. 33993, especially pp. 37188.
namely your action must promote the good, and since you have reason
to meet those standards, because the good is to be promoted, you also
have reason to be an agent.
Making this response to the problem obviously means leaving aside
the possibility of an explanation of reasons for action in terms only
of the standards of action, without reference to value. Here instead the
reason to be an agent, that is, to perform actions that have certain
standards of success, is explained in terms of the reason to produce
certain values. It may seem now that characterizing action in terms of
production is making no contribution at all to the theory of practical reason. But this is not so. Claims about the nature of action are
no longer playing the kind of ambitious role that Korsgaard and
Velleman, for example, hoped that they would. They are not explaining
the whole of practical reason. But they still have a purpose. Instead of
deriving all practical reasons from the nature of action, we have an
interconnection between value, practical reason and the nature of
action. Values are the kind of thing that should be brought about as
much as possible. Action is essentially production; successful action
is the production of good states of affairs. So we have reason to act
successfully and to produce good states of affairs and these practical
reasons can be seen as connected both to the nature of value and to the
nature of action.
The difference between thinking of action as production and as execution is substantial and they tend to support different theories of practical reason. This is particularly noticeable if the agent does not intend
or aim at producing what is in fact a good state of affairs.
Suppose that it is good to maximize happiness but the agent instead
intends to make herself rich. She may successfully execute this intention without maximizing happiness. According to action as execution,
her action is successful. According to action as production, her action
is a failure, even though she did what she set out to do.
Now suppose that she does not execute this intention very well, and
she ends up poor, but at the same time making others happy. According
to the view that action is essentially production, her action is successful: she has produced a good state of affairs. According to the view that
action is essentially execution, her action has failed; she has not
achieved her aim. In these circumstances, it may be a matter of luck
that she produced a good outcome. But equally it may be predictable
that by forming apparently unrelated aims and intentions, she may
68
alison hills
achieve a good state of affairs (either by successfully or, as in this example unsuccessfully, pursuing the alternative ends).
It is for this reason that consequentialist theories of practical reason
may be self-effacing, that is, they may require an agent not to follow
that theory of practical reason. For example, a good utilitarian may aim
to look after herself and her friends and carry out that aim, rather than
aiming to maximize happiness, precisely because she has reason to
maximize happiness. This may be the best way for her to do so. This
aspect of consequentialist theories has been thought to be a problem by
some and defended by others.
But the real problem for this theory is much worse: there is something very peculiar about classing an action as a success when the agent
not merely has completely failed to do what she aimed to do, but may
not even have been carrying out an aim at all. She may have brought
about an event with good consequences, but this is not the same as acting successfully. In fact, she may be able to produce the good consequences without acting at all, that is, without carrying out any particular
goal or intention.
For example, suppose that tripping over the kerb I fall againstanother
pedestrian who was about to step into the road, not having seen the
lorry speeding towards him. My tripping has saved his life, bringing
about good consequences, and therefore is classified as successful
action. But obviously it is not: it is not an action at all.
The theory of action as production, as it stands, may classify as
examples of successful action kinds of behaviour or events that are not
actions.17 The combination of claims that successful action just is the
production of good results, and that practical reasons are reasons to
perform successful actions, is unacceptable, and we will have to reject
one of them.18 If we are to maintain the intuitive connection between
practical reason and successful action, we will have to find a different
way of relating action, value and practical reason.
17
The theory that successful action is the production of good results and the claim
that actions aim at producing therefore come into conflict, at least when the latter is
understood in terms of the agent herself aiming at bringing about some results. A similar criticism is made of these kinds of theory by T. Schapiro, Three Conceptions of
Action in Moral Theory, Nous 35 (2001), pp. 93117.
18
This is not an objection to maximizing consequentialist theories of practical reason, for it is obviously open to someone who wanted to maintain that practical reasons
are reasons to produce good outcomes to drop the claim that practical reasons are
reasons to perform successful actions. But this claim is intuitively plausible, and a theory that links practical reason, value and action is intuitively appealing.
A common thought is that action can express value. This was taken
extremely literally by Wollaston, whose views are described by Schapiro
as follows:
The thoughts which Wollaston takes to be expressed in action are essentially propositional in nature. For example, by firing on an approaching army, a body of soldiers offers the salute of an enemy, a salute
whichdeclares that the two armies are indeed foes. If the actual relation
which holds between them is one of friendship, then the proposition
embodied in this act is false. Similarly, by using and disposing of an
object in his possession, a man declares it to be his own. If he does not
in fact stand to the object as an owner to his property, then the declaration is false Action, on this account, is a form of assertion, a mode of
expression in which we purport to represent things as they really are.
And this, Wollaston claims, is the sense in which actions can be or fail to
be in conformity with the nature of things, independently of their
consequences.19
19
20
70
alison hills
something, you commit to it in the following sense: you aim at it, try to
achieve it, choose what to do on the basis of their contribution to it,
and monitor your success in achieving it. When you intend to X for its
own sake, you aim at it and try to achieve it for its own sake.21
Intentions can express value in the following sense. If something
is valuable for its own sake, then it is worth pursuing for its own sake.
If something is worth pursuing for its own sake, then it is appropriate
to aim at it, to try to achieve it, to choose your actions on the basis of
their contribution to it, and to monitor your success at achieving it for
its own sake. So when you intend X as an end, you treat X in ways that
would be appropriate if X really were worth pursuing for its own sake.
Similarly, if you intend Y as a means to some other end Z, you treat Y
in ways that would be appropriate if it really were valuable as a means
to Z.
For example, if you intend to kill someone, your intention is inappropriate, because killing is not worth pursuing for its own sake. If you
intend to help Sarah because she is in need, your intention is appropriate. Intentions can also express non-moral value. It is appropriate for
you to commit yourself to having a hearty breakfast, but it is not appropriate to intend to eat a large lunch, because the former is good for you,
the latter is not.
Note that this is different from the traditional claim that action is
under the guise of the good, according to which, when an agent desires,
aims or intends something, she sees it as good in some respect. An
agent can intend to eat a large lunch without regarding eating a large
lunch as good or worth-while in any respect. But she will thereby treat
eating a large lunch in ways that would be appropriate only if it had
qualities that she herself does not think that it has.22
The simplest way of connecting action and commitment is to say
that actions aim at having appropriate commitments, and that an
21
This conception of intention as commitment is very similar to that defended by
M. Bratman, Intentions, Plans and Practical Reasoning (CSLI, 1999), except that he
does not discuss the connection between intention and value.
22
As a consequence, there is a problem in understanding in what sense actions aim
at appropriate commitment, since this is not something at which the agent herself
aims. The best way of understanding this claim is, I think, the following. Anyone who
can be considered an agent has intentions, and in appropriate circumstances, those
intentions are commitments that can be assessed as appropriate or inappropriate. Since
commitments can be evaluated in this way, and actions essentially involve commitments, it is reasonable to assess actions by this standard, even though agents themselves need not aim at having appropriate commitments.
action is successful if the intention on which the agent acts is appropriate. An intention is appropriate if the agent intends to X for its own
sake, only if it is valuable for its own sake; he treats Y as if it were valuable as a means only if it is valuable as a means.23 The action you have
most reason to do is the action whose commitment is most appropriate. We can call this action as commitment.
The relevant conception of practical reason, given this idea of
action as commitment, is that an agent has reason to perform some
action if the intention on which the agent acts is appropriate (that is,
appropriately responsive to value). As in the case of action as production, the goal here is not to explain fully practical reason in terms of the
nature of action, but rather to give an account of action, practical reason and value that fits together. The view of value that fits with action
as commitment is quite different from that underlying action as production, where value was thought of as that which should be brought
about. Instead, value is conceived of as that which should be responded
to in the right way, for example, by aiming at it, trying to bring it about
and so on. According to this view, what is important in terms of action
is not production of value but orientation with regard to value.24
Action as commitment might well be suited to a kind of nonconsequentialist moral theory, for what is important about action is
commitment (e.g. trying not to kill people) rather than production
(e.g. making sure as few people are killed as possible). For example, it
may support a version of the doctrine of double effect, for the distinction between what is intended (what one is committed to) and what is
not intended (whether it is foreseen or not) is very important on this
view of action.
The idea of action as commitment, and related theories of practical
reason and morality, are interestingly close to Kants conception of the
good will. According to Kant, a good will wills the right action for
the right reason (the relationship in Kant between practical reason
and value is too complicated and controversial to discuss here). But a
23
Intending to perform Y for the sake of X is an appropriate commitment only if X
is valuable and Y is a means to X, however. If X is not an appropriate end, the agents
overall commitment cannot be appropriate even if Y is a suitable means to the end she
in fact chose.
24
Correct orientation with regard to value requires at least that you intend to X for
its own sake only if it is valuable for its own sake. But it may also be necessary that you
formed appropriate intentions for the right reasons, that is, that you intend to X for its
own sake because it is valuable for its own sake.
72
alison hills
problem often raised for Kants good will is relevant here: surely it is
important not just what someone wills, but also what they actually do.
In our terms, we may have given an account of successful intention,
but we have missed out an aspect of action that is surely crucially
important: carrying out that intention. No one could call an action
successful however worthy the intention if the agent utterly failed to
execute it. Suppose that you form an appropriate intention: to work all
morning. But then you find yourself engrossed in a variety of entertaining but trivial distractions. This is not a paradigm of successful
action. Nor could it be said that you did what you had reason to do. You
had reason to work all morning, not just to form the intention to do so.
This suggests that commitment alone cannot be what is essential
to action. It must at least be combined with another feature. And the
obvious candidates are the two that have already been discussed: execution and production. This leaves us two possibilities: first, that successful action is an appropriate intention plus successful execution of
that intention;25 second, that successful action is the production of
good consequences together with an appropriate intention.
There is a problem with the latter theory as it stands. An agent may
have an appropriate intention and produce good consequences, but not
through successfully executing her intention. For example, you may
intend to help others and you may actually improve their welfare, but
precisely by failing to act on your intention to help and thus forcing
those others to fend for themselves and become self-reliant. This surely
should not count as successful action, however. So we need to amend
the second theory to: successful action is the production of good consequences through an appropriate intention. This rules out accidentally
contributing to good consequences whilst at the same time having a
good intention.
This brings the two accounts much closer together, for in both cases
successful execution of an appropriate intention is essential to successful action. But there may still be a difference between the two. The successful execution of an appropriate intention may not produce the best
consequences, for the long-term consequences may be much worse
25
As before, execution here must not just be that the intention is in fact satisfied, or
even that the intention is satisfied though some behaviour of the agent. The behaviour
must be caused in the right way by the agents aim in order for the agent to
execute her intention: the agent must control her behaviour in order to satisfy her
intention. It is very difficult to say what this control consists in, and unfortunately this
is a question I cannot pursue here.
74
alison hills
If you are inclined to say no, you will be drawn to theories of practical
reason that emphasize reasons to produce good consequences; if you
would say yes, you will be supporting a theory of practical reason that
emphasizes reasons to respond appropriately to value.
Your view on value obviously may feed in and influence this decision
too. These two conceptions of action both support conceptions of value
as something to be produced and as something to be responded to in
other ways. How important is either of these two components? If you
think that it is fundamental to value that it is to be brought about as
much as possible, you should favour the conception of action that
emphasizes production. If, by contrast, you think that responding to
value in ways other than production is important, you should favour
action as commitment plus execution.
I cannot settle this question here, so I will leave these two accounts
of action still standing. Practical reason for each of these is a complex
matter, including reasons to respond to value appropriately, instrumental reasons (to carry out intentions successfully) and perhaps reasons to produce good consequences too. But we have been led to this
more complicated account of practical reason by an attempt to give an
adequate account of action (and of value) and simpler theories proved
to miss out one or more of the essential features of successful action.
5. Basic Action and Mature Action
There is still a serious question, however, over whether our final
accounts of action are too complicated. Can we consider small children
and animals as forming appropriate intentions and carrying them out?
If not, are we to conclude that they do not act after all?
This raises a significant question for the whole approach of this
paper. Can we in fact give one account of action, identifying a single
standard of success? Perhaps different kinds of action have different
standards. For example, perhaps animals and small children perform
basic actions which have as their standards of success: the agents
successfully executing her aim or intention. Perhaps we have been setting out a different kind of action, which we might call mature action,
whose standard involves an assessment of what the agent is doing
(through an assessment of her intention, and perhaps of what she produces) as well as whether she successfully achieved her aim.
I think that there is something to this complaint, but that there are
important connections between the different conceptions of action.
76
alison hills
78
onora oneill
1
Phillipa Foot, Morality as a System of Hypothetical Imperatives, in her Virtues
and Vices (Oxford: Basil Blackwell, 1978), pp. 15773.
the meeting with friends at the pub that I shall arrange. But this is not
to say anything about the particular acts that I will do, beyond stating
what types they will be with some generality. The definite descriptions
do no more than specify the types of act that are to be done. Although
they individuate the agent by whom and the occasion on which they
are to be done, they do not distinguish between the innumerable ways
in which the specification of action might be met. It may seem, then,
that the gap between norms (for which reasons may be given) and
enactments of those norms (for which no complete reasons can be
given) marks a limit of practical reasoning.
Taking a Relaxed View of Norms
So the optimistic thought that norms are apt for practical reason
because they combine propositional structure and content (making
them apt for reasoning) with a capacity to guide action (making them
practical) is too quick. This might suggest some problem with reliance
on the idea of a norm. The term has a chequered history, and may seem
in some ways to be too narrow and in others too broad to provide a
useful focus for practical reasoning. However, I do not think that the
difficulties lie specifically in the use of the term norm, rather than of
other terms for practical propositions that combine pro-positional
structure and content with claims to be action guiding (such as principle, rule, standard, law). We should not and we need not be the p
risoners
of past uses of the term norm (particularly now that the term normative
is used in a much broader way). As I see it, there are reasons for being
rather relaxed in our understanding of both terms, and rather self conscious about their relations to other terms that serve cognate purposes.
Indeed the only feature of norms that it seems to me essential to retain
is the thought that their use is to guide the way action changes the
world rather than to describe the way the world is. A focus on direction
of fit suggests four issues that might arise if we relied on traditional,
more restrictive understandings of norms. I shall discuss these in turn
and suggest that there is nothing particularly problematic about centring discussion of practical reasoning on norms.
1. Norms and Other Directives
One set of considerations that might lead us to wonder whether norms
offer too narrow a focus for practical reasoning is that there are other
80
onora oneill
82
onora oneill
lot of history here. Some strands in that history urge us to think that
norms provide a uniquely suitable focus for practical reasoning precisely because they are socially embedded, because they are the norms
of this or that group at this or that time, and that this is what makes
them important, practical and motivating. Others have seen the understanding of norms as socially embedded as a great deficiency of a sociological understanding of norm-based practical reasoning. If norms
are seen as practical propositions that are entrenched in the lives of
some group, the objection runs, they offer a highly problematic focus
for practical reasoning. Some embedded practical propositions may
provide poor reasons for actionand the fact that they are embedded
adds only the questionable weight of an argument from authority.
(We need only think of appeals to entrenched norms of revenge, honour killing or female genital mutilation.) By contrast, abstract principles of action that at a given time are not well embedded may sometimes
offer significant reasons for action.
However, letting go of the traditional, narrower sociological understanding of norms as embedded is not problematic. In fact, I suspect
that current usage of the term has long since distanced norms and normativity in discussions of ethics and practical reason from the older
sociological conception of norms as socially embedded, although there
are pockets of lingering confusion.
4. Norms as Motivating
And there is a fourth point that follows from accepting a relaxed view
of norms. An explicit rejection of the older sociological conception of
norms as socially embedded undermines the psychological corollaries
of such views, and requires us to set aside the assumption that norms
are invariably motivating. There is no doubt a complex story to be told
about the ways in which norms come to be accepted, understood,
rejected and modified, and about the emergence of particular norms in
the lives of individuals and societies. There is also no doubt a story to
be told about what Sabina Lovibond has called the animal precursors
of practical reasoning.3 However, once we accept a broader understanding of norms as practical propositions that may or may not be
embedded in the life of this or that group, we will also have to accept
3
Sabina Lovibond, Practical Reason and its Animal Precursors, European Journal
of Philosophy 14.2 (2006), pp. 26273.
84
onora oneill
reasoning may look algorithmic. The instructions for utilitarian calculation tell us to list all options; to reckon the probable outcomes of each;
to calculate and sum the expected utilities of these outcomes for all
parties; and finally to maximize. In practice, as we all know, utilitarian
calculation cannot even approximate the underlying algorithm: we can
specify only selected options; we are often unsure about their probable
outcomes; our calculation of expected utility for anyone (let alone
everyone) is pretty gestural. Only the maximizing looks even close to
algorithmichowever, on reflection we see that it too is not wholly tied
down and could be done in various ways.
On second and more cheerful thoughts, it perhaps is not important
if norms are not algorithmic. For they can at least formulate constraintson or advice for action, and perhaps all that matters is that we
choose some act that meets the constraint or recommendation set by a
norm that we are seeking to respect. Yet how are we meant to do this?
A standard answer is that it is a matter of judgement. But this is not
reassuring. Invoking judgement without explaining how it is to work
seems to leave us no clearer about what we need to add to norms if they
are to offer practical guidance. Yet perhaps practical reasoning can take
us no further.
Writing of theoretical rather than practical judgement, Kant commented that reason could not take us all the waya point that has been
endorsed by many later writers:
[I]f it [the understanding] wanted to show generally how one ought to
subsume under these rules, i.e., distinguish whether something stands
under them or not, this could not happen except once again through a
rule. But just because this is a rule, it would demand another instruction
for the power of judgement, and so it becomes clear that although the
understanding is certainly capable of being instructed and equipped
through rules, the power of judgement is a special talent that cannot be
taught but only practiced.6
6
Immanuel Kant, Critique of Pure Reason (trans. and ed. Paul Guyer and Allen W.
Wood; Cambridge: Cambridge University Press, 1999), A133/B172. Here Kant is
writing specifically about subsumptive or determinant judgement, the most standard type of theoretical judgement. See below for some comments on reflective
judgement.
86
onora oneill
Picking and Choosing
88
onora oneill
90
onora oneill
92
onora oneill
13
See Barbara Herman, Performance and Obligation, in her The Practice of Moral
Judgment (Cambridge, MA: Harvard University Press, 1993), pp. 15983; and Onora
ONeill, Instituting Principles: Between Duty and Action, in Mark Timmons (ed.),
Kants Meta-physics of Morals: Interpretative Essays (Oxford: Oxford University Press,
2002), pp. 33147.
part II
Particularism
96
roger crisp
In his earlier work, Moral Reasons, it seems that Dancy may have
been aiming to defend SH. At one point, for example, he considers the
common charge against his particularistic view, grounded on holism
in the theory of reasons, that it is always bad knowingly to cause pain.2
Rather than allowing that a particularist could accept that indeed it is,
Dancy seeks to show that it is not.
In opposition to both versions of holism is atomism:
2
98
roger crisp
This is consistent with SH, since it may be the case sometimes that
the property some action has of being a causing of suffering to a nonrational sentient being does not count against it, or even counts in its
favour.
Inconsistent with SH would be:
T*: In every possible case, the fact that some action is one of causing suffering to a non-rational sentient being counts against it.
Note that these principles are quite similar. Note further the implication of the distinction between the criterion of rightness and decisionprocedures which allowed us to focus on Dancys real concern.
Someone who accepts T may think that, in practice, it would be wiser
to teach children T*, on the ground that one maximizes the chances of
action in conformity with the real reasons.
But even if we assume that this is not the case and that we should
teach children T rather than T*, it seems because the principles are so
similar that the difference between particularism and generalism may
not be especially great. To be sure, an ideal moral agent, according to
the particularist, is likely to need pretty good judgement to tell when a
case is an exception to a principle. But a generalist like Ross is also
going to require judgement about the weight of different principles in
particular cases.
The second problem is Dancys unreflective reliance on commonsense views about which reasons we in fact have. One obviously plausible strategy for understanding how reasons work is first to offer a
(defeasible) account of certain reasons, and then assess how they
behave in different circumstances. But such an account should not be
taken directly from common sense, as it so often is in writing about
100
roger crisp
not begin enquiring into our reasons, or their nature, on the assumption that they must, to be reasons, function universally in the way that
atomism says. But reflection upon the nature of explanation suggests
analogies with the nature of justification which throw doubt upon WH.
Important here is the distinction Dancy draws between favouring
and enabling (pp. 3843). Here is one of Dancys own examples:
1. I promised to do it.
2.My promise was not given under duress.
3. I am able to do it.
4. There is no greater reason not to do it.
5.So: I do it.
Premise 1, Dancy suggests, is a favourer: the fact that I have promised counts in favour of my doing what I have promised. 3, however,
merely enables 1 to favour 5. It does not itself add to the reasons in
favour of acting. Dancy believes that 2 is also an enabler rather than a
favourer.
In an earlier article on particularism, I sought to draw an analogy
between explanation and justification, claiming that just as what
I called a full explanation of some event, such as an action, will include
everything, including background or enabling conditions, so the full
story about justification should not leave anything out.4 Dancy says
that, because I insist on bringing enabling conditions into the justification, in that earlier argument I do not respect the distinction between
right-making features and enabling conditions (p. 47). I did not, however, wish to claim that all enabling conditions must feature in a justification (though they should feature in a causal explanation). For
example, if I now have a strong justification for -ing, the fact that God
has not just struck me dead is one of the conditions that enables me
now to act. But it should not be part of my justification. Rather, what
should be taken up into the justification are all those allegedly merely
enabling conditions that are in fact justificatory. This is why I am what
Dancy calls an expander (p. 96), and why I seek what he calls guarantees (pp. 9597). As Dancy puts it: Every ultimate reason is a feature
whose presence guarantees that there is a reason (i.e. that there is
Roger Crisp, Particularizing Particularism, in B. Hooker and M. Little (eds.),
Moral Particularism (Oxford: Clarendon Press, 2000), pp. 2347, esp. at pp. 3436,
4244.
4
102
roger crisp
something to be said for doing the action), and any reason we can rest
satisfied with is an ultimate reason (p. 96; see also pp. 12527).
Dancys first objection to this idea is that a guarantee that there is a
reason to do the action need not itself be a reason in favour of doing it,
since the only way I could reach a guarantee would be by adding enablers and the absence of disablers to an incomplete reason. That would
mean that anything that could be a reason for doing the action has got
submerged in a host of considerations that are doing other jobs. The
crucial point here is that a combination of a favourer and an enabler
need not itself be a favourer at all (p. 96).
On that final point, as I have said, I think Dancy is right. But I think
also that the method of expansion he describes is better understood in
terms of finessing a favourer. To use the example of promising again,
what happens is that, once we learn that a promise has to be freely
given, we amend the claim that having promised to counts in favour
of -ing to the claim that having freely promised to counts in favour
of -ing. So I have little doubt that premise 2 should be incorporated
into the justification of my action.
In support of his claim that what is doing the normative work in this
case is not the combination of the facts that I have promised and that
I was not under duress when I promised, but rather just the fact that
I have promised, Dancy suggests that even those whose promise has
been deceitfully extracted often feel some compunction in not fulfilling
their promise, even though they realize that their promise does not
play its usual reason-giving role. I confess that this is a phenomenon
I fail to recognize. But even if it exists, it is best explained as a remnant
of our moral upbringing, in which we are taught the simple principle
that it is wrong to break promises. A second example of Dancys concerns the reason you have to help someone who asks you the time. If
their purpose were to distract you so that their accomplice could steal
your bag, that reason would not exist. But Dancy is unwilling to say
that the real reason must be construed as their asking you the time for
some appropriate purpose. Again, this seems to me a case where the
alleged enabler should be incorporated into the favourer, enabling us
better to understand the justification in question.
Here we should also note Dancys readiness to allow the possibility
of a principled but holistic ethic, which might include a principle
suchas:
104
roger crisp
better, given their fundamental role in ethics, to the law of the conservation of energy. Something like this view that we should have an open
mind on how reasons behave is what lies behind WH.
But the particularist analogy fails. Justificatory reasons must function universally, like standard causes. The WH picture seems to be the
following. At the quantum level, we might have some set of initial conditions s which issues in outcome o. But we cannot strictly claim that
s causes o, because we know that some identical set of initial conditions
s* might issue in an out-come p which is non-identical with o. What we
have here is a case of genuine indeterminism, which is to that extent
inexplicable: it just happens that way. Similarly, it may be that some
feature of an action f functions as an ultimate justificatory reason for
performing some action, but in a similar situation fails so to function
or even has the opposite polarity.
Ultimate reasons for action cannot be like that, however, since they
help us to understand why someone has a justification for acting. If
I claim to be justified in -ing because of some feature of -ing, but
then refuse to allow that this same feature justifies me in -ing, then
I have further work to do in specifying exactly what my reasons are.
A staunch holist might refuse, but this is to follow the white rabbit.
What is more likelyas usually happens in Dancys caseis that the
justification will in fact be further specified, but holism retained
through the use of a distinction between enablers and favourers that,
respectable enough though that distinction is, is not operating here.
WH allows that, as things turn out, some reasons may in fact always
have the same normative force across different cases. So let me end by
saying a little about Dancys views on invariance:
It may be true that every reason may alter or lose its polarity from case to
case, even though there are some reasons that do not do this. If they dont
do it, this will be because of the particular reasons they are. Invariant
reasons, should there be any, will be invariant not because they are reasons but because of their specific content It is like the claims that a man
can run a mile in four minutes, that Sam Smith is a man, and that Sam
Smith cannot run a mile in four minutes. These claims are compatible,
and so are the claims that reasons are variable qua reasons though some
reasons are (necessarily, given their content) invariant. The invariance,
where it occurs, derives not from the fact that we are dealing here with a
reason, but from the particular content of that reason. (p. 77)
The thought here is that, just as some men turn out to be able to run
a mile in four minutes and some do not, so some reasons may turn out
to be invariant and others not. But Dancy would be unhappy with this
view of reasons:
It is, I suppose, conceivable that though the vast bulk of reasons function
according to a holistic logic, there are a few whose logic is atomistic. But
if this were true we would have a hybrid conception of rationality. There
would just be two sorts of reasons, each with their own logic, and moral
thought would be the uncomfortable attempt to rub such reasons
together. It is much more attractive, if at all possible, to think of our reasons as sharing a basic logic, so that all are atomistic, or all are holistic.
(p. 77)
This passage seems to be in effect a restatement of the holistic conception of a reason, according to which how a reason functions in one
case is clearly going to be independent of how it functions in another,
because of the possibility of variance. But we have seen that the very
concept of a reason is atomistic and generalist. So in that sense whether
some feature is a reasonthat is, an ultimate reason (the kind that
really matters)or not will not be independent of how it functionselsewhere. If it functions differentlyif its polarity switches, for example
then that will be a clue that what we are dealing with is not a
counter-example to generalism but a non-ultimate reason.
In conclusion, let me restate my main claims. The central topic of
debate here is whether holism or atomism captures the truth about
normative reasons. Holism can be understood as a weak view about the
106
roger crisp
Strong holism says, effectively, that if this is true there can be no general principles stating the reason-giving status of any such feature.
Weak holism allows that there will be such a principle for any invariably reason-giving featureand that there may be such features.
Crisp starts by considering the stronger form of holism. He makes
various points about the sort of argument that I have produced in its
favour. One way of arguing for holism is by appeal to examples. Crisp
repeats his complaint (in the Hooker/Little volume1) that the examples
I offer fail to distinguish between ultimate and non-ultimate reasons.
In his earlier version, he characterized an ultimate reason as a consideration that we can rest satisfied with as a reason. This made it easy for
me to say that I was perfectly satisfied with lots of reasons that Crisp
wanted to call non-ultimate. Now he has a different account of the ultimate: These reasons are non-ultimate or derivative, and depend for
their normative force on certain higher-order reasons which can be
seen to subsume them into a single generalist form.2
What is going on here? There are, I think, various possibilities. One
is that what this means is that the considerations we ordinarily think of
as reasons are not really reasons at all. Properly speaking, they are parts
of reasons. A part of a reason is not subsumed into something higherorder; it is an element in a complex which is itself a reason. No part of
B. Hooker and M. Little (eds.), Moral Particularism (Oxford: Oxford University
Press, 2000).
2
Roger Crisp, Ethics Without Reasons?, p. 44, this volume.
1
108
jonathan dancy
that complex is more a reason than any other part. If we look inside
that reason, dismantle it, as it were, what we find are things which do
not favour anything (at least not here). The only real reasons are the
ultimate ones.
A different interpretation allows that non-ultimate reasons are reasons all right, but claims that these reasons get their normative force
from being sub-sumed under certain higher-order reasons, which are
invariant. (I think this claim of invariance is what Crisp means by his
talk of their having a generalist form.) This interpretation would fit
something Crisp said in his earlier attempt; he held there that the virtues were ultimate reasons, suggesting that the ultimate reason for
doing many things was that doing them would make ones life go better.
It seems to me, however, that though one can generalize in this way, in
doing so one goes a step too far. The reason why things make ones life
go better is that they are fun, rewarding, engrossing and so on, worth
doing in their own right. It is not the other way around. Perhaps better:
that something is fun, or in some other way worth doing, is not a reason to do it because one has a general reason to make ones life go better, and having fun is one way of doing that (sometimes). That this
would be fun is its own reason to do it, and ones life goes better to the
extent that one is able to do a wide variety of the things that one has
good reason to do. The so-called ultimate reason is the derivative one,
I would sayif it is a further, or distinct reason at all. If there is normative force here, it comes up from below, not down from above.
We will return to these issues. But for the moment we are only considering the fortunes of strong holism, which goes further than anything I was prepared to defend in Ethics Without Principles. As Crisp
points out, strong holism is in danger of being refuted by the discovery
of a single invariably relevant consideration, complex or simple. (I say
in danger of because I remain worried that the complex consideration
at issue may not itself be a reason, but rather a mere guarantee that
there is a reason.) So we pass to weak holism, which allows for the possibility of invariance. Crisp opens his enquiry by noting that, according
to the weak holist, we should not start off with the assumption that our
reasons must somehow function universally, as the atomist claims they
must. But, he says, the relation between justification and explanation
suggests otherwise. I found it quite hard to work out why he says this.
There are two points on which Crisp and I agree. First, he accepts the
crucial distinction between favourers and enablers. Second, he allows
that a combination of favourers and enablers should not necessarily
itself be a favourer. With these things in hand, as I see it, Crisp is thinking that, though we start from some favourer (perhaps that I promised
to do it), and add an enabler to it (that my promise was freely given),
the demands of justification and explanation drive us onwards until we
reach a specification of some complex of features which is an invariant
reason, even if the things from which we started, and what we have at
each intermediate stage, are not. What we have at the end of this process, I urged, was not a reason of any sortnot, that is, some sort of
favourerbut a guarantee that something favours the action, a guarantee which will contain a mention of the favouring consideration, but
which will not give it any pride of place. This would not respect Crisps
earlier thought that the normative force of the reason derives, by subsumption, from that of the guarantee. The guarantee has normative
force merely because of the force of the reason whose presence it guarantees; again, things are the other way around.
Crisp has seen this all before, and does not think it is a correct
account of what is going on. One way in which things have got distorted is that, he says, he does not think that every enabler must feature
in the relevant justification. For instance, I may have a strong justification for doing something which I could not do if God were to strike me
dead, but that God has not struck me dead is no part of the justification
for my doing that thing. This, however, is not to the point. An enabler
is not a feature required for me to do the action; it is a feature required
for the reason to do it to be the reason it is. What is enabled is not the
action but the reason; it is not that I am enabled to do the action but
that the feature is enabled to favour the action. I am sure that Crisp
knows this crucial aspect of my notion of an enabler perfectly well, but
he seems to have forgotten it at this point. However, having forgotten it,
he is in a position where he can say that only those enabling conditions
need to be part of the justification that are themselves justificatory. But
this was just the position I was complaining about. I dont feel that enabling conditions justify anything. One might call the relevant reason
justificatory, or one might call the whole story justificatory, but a
required enabler is only justificatory in a weak sense. We see this sense
emerging when Crisp writes It is the fact that you have freely promisedthat makes the difference.3 The difference it makes, I would say,
is the difference between a promise that gives us a reason and one
Crisp, Ethics Without Reasons?, p. 47, this volume.
110
jonathan dancy
thatdoesnt. But this difference is not enough to call the enabling condition justificatory. The only thing it justifies is seeing a reason to keep
this promise when one sees no reason to keep another. If the question
raised by someone who seeks justification is why there was a reason to
keep the first promise and no reason to keep the second, then this is
what the fact that the enabler was present in the first case and not in the
second will provide. But in that case what it is justifying in the first case
is not doing the action, but rather doing it when one is not going to
keep ones promise in the second, similar case. What I would say, then,
is that the enabler only succeeds in justifying ones comparative choice.
But in that role, it is acting as a favourer. It favours this rather peculiar
object: keeping ones promise in the first case and not in the second.
But this does nothing to show that it is part of the justificatory story for
the first case alone.
As I see things, it is the comparative question that drives the demand
to include within the relevant reason, properly understood, everything
whose presence or absence can make a difference to the normative
score, whether reason, enabler, intensifier, disabler or attenuator.
There remains a worry about the White Rabbit. Crisp writes: If some
feature that is claimed to be an ultimate reason for action in one case
fails as a reason in some other case, then either the two features must
be allowed to differ or we have followed the White Rabbit into a world
where anything can happen.4 Now I would have thought that it is
straightforwardly impossible for something that is an ultimate reason
in one case to fail to be a reason, indeed an ultimate one, in another.
But suppose instead that something that is claimed to be an ultimate
reason in one case turns out not to be any such thing in another. Then,
I think, by the definition of ultimate reason we know straight off that
that feature is not an ultimate reason anywhere. The worry about the
White Rabbit only emerges if we have a feature which is a reason in one
case and not in another, though there is no explanation of this at all.
This is at least conceptually possible, I would claim, though it is not
possible if we are dealing with ultimate reasons, which are defined in
such a way as to have to be invariant.
So what I want to say about this is that if we are dealing with a feature
which is a reason in one case and not in another, there will be an explanation of this, which will consist in the presence of some disabler or the
4
112
jonathan dancy
Others arrive switched off, but can be switched on by appropriate contexts. And the ones that arrive switched on can be likewise switched off,
if the situation so determines. So all this is formally compatible with
the constraints of holism, and it is therefore a picture that is available to
particularism.
As well as the metaphor of switching on and off, I offered a somewhat less metaphorical way of thinking about the default. If a default
reason is not switched off, but functions here at the default setting, as it
were, there is nothing to explain. It is only when things are not as they
are default-set to be that we begin to ask questions. If a default reason
is not here the reason it is antecedently set up to be, there must be an
explanation of the matter. The reverse of all this is true for the nondefault reason; here explanation is called for if these are reasons at all,
while no explanation is needed if they remain normatively inert in the
background.
But even this less metaphorical way of capturing the notion of a
default is not something of which one can be very proud, and it is easy
to see that there may be difficulties involved in trying to work the general idea out in more detail, as we would all wish to do. What I had not
foreseen is how easy it would be for the whole idea to be undermined,
as McKeever and Ridge have shown it to be. The first move in this process involves the link between defaults and enablers. A non-default reason is one that needs something to switch it on, and I have a term for
this sort of switcher-on: an enabler. So the default reasons must be ones
that need no enabler. If they are like the non-default reasons in needing
an enabler, it can hardly be true that they arrive switched on. They
would only be switched on if the enabler is there, which it might not be.
Of course one might try to say that what they need is not so much an
enabler as the absence of disablers. That sounds much better; there are
things that can switch them off, but so long as none of those disablers
is present they will stay switched on. The trouble is that there seems to
be no difference between the presence of an enabler and the absence of
a disabler, nor between the absence of an enabler and the presence of a
disabler. Our understanding of these terms is run entirely by appeal to
certain subjunctive conditionals, and subjunctive conditionals are not
capable of drawing distinctions fine enough to distinguish the presence
of something that would stop a default reason from being a reason
from the absence of something required for that reason to be a reason.So the absence of a disabler is an enabler, and so default reasons,
which certainly must have disablers if holism is to be true, must also
114
jonathan dancy
indeed dependent on experience, in the way that Quine claims, but this
is negative dependence; a belief is negatively dependent on experience if
its retention is dependent on the absence of awkward experience.
A belief is positively dependent on experience if experience acts as its
ground, that is, if the acceptability of getting the belief depends on the
presence of supporting experience. Giaquintos point is that a belief
can be negatively dependent on experience without being positively
dependent on it. And to constitute a priori, or conceptual, knowledge,
the belief only needs positive independence. Quines point that it is
negatively dependent on experience is irrelevant.
This seems to be another instance of the distinction between the
absence of an impediment (a disabler) and the presence of something
more positive, in this case a ground. The absence of an impediment is
nothing like the presence of a ground, nor is the presence of the impediment anything like the absence of a ground. Could we therefore try
saying, as it were, that a default reason is negatively dependent on the
absence of a disabler, but not positively dependent on anything?6 If we
could make this out, something would have been salvaged from the
flames. What we are after is a distinction between something worth
calling support and the mere absence of an impediment. We find versions of this distinction in pairs of the following sorts:
I can bicycle from home to campus in 15 minutes if there isnt a strong
head-wind. (The absence of the headwind does not produce any of the
power required.)
I can bicycle from home to campus in 10 minutes with a strong following
wind. (The wind helps; it produces part of the power required.)
I can get over the stile if you dont get in my way, hinder me. He cant get
over the stile without your help.
I can get over the stile if I am not too tired. My not being tired doesnt
help me get over the stile; my being tired prevents me from doing what
otherwise is in my power, which amounts to saying that it deprives me of
the power.
116
jonathan dancy
In fact there are two suggestions here, the link between which is
obscure. The first is that a reason for action is a consideration that
makes it better to act in one way than in another. This seems to differ
from the standard account of reasons as favourers in two ways. First, it
thinks of favouring as making good; I take this to be a distinct shift,
and will say why below. Second, it thinks of reasons as always comparative in their focus; a reason does not just favour an action; it favours
doing this one rather than doing that one.
118
jonathan dancy
but the action for which it is a reason may not, for all that, be done. This
is just as well, because there are reasons against as well as reasons in
favour, and we would not want to understand all reasons against as
some-how ensuring that an action of the sort that they are against still
gets done.
There is a problem here, which was, as far as I know, first noticed by
H.A. Prichard, and which I call the Prichard problem. How exactly are
we to make sense of the idea of a relation that can be instantiated even
when only one of its relata in fact obtains (or occurs, or is the case)?
One way in which the Prichard point is important is that it reminds us
that choosing an action is not at all like choosing a chocolate. It is not
as if the possible alternative actions are laid out for us on a tray, in their
rich particularity. Only one of those possibilities is going to have a rich
particularity, and the others are not even going to exist.
It is important, of course, that in addressing the Prichard problem
we focus on the case before us. It might not be enough, for instance, to
appeal to intensional contexts, since it is not at all clear that reasonclaims are intensional. But my purpose here is not to solve the problem
but to use it to point out a difference between favouring and goodmaking. The Prichard problem does not really apply to the good-making relation, any more than it does to the right-making relation. One
can refrain from doing an action for the reason that it would be wrong.
This is properly expressed in terms of a subjunctive conditional. To do
it would have been wrong, and our reason for not doing it can be thatit
would have been wrong; we dont need to say that the reason was that
it was wrong. And we can take a similar approach to good-making.
It would have been better to have given the car keys to someone else,
given the amount of beer I consumed that evening. I didnt do so, and
so there was no action of the relevant type that was made better by the
amount of beer consumed. But that doesnt raise a Prichard problem,
because the subjunctive reading (it would have been better ) is perfectly satisfactory. Things are different with the favouring relation. It is
not that an action of the relevant type would be favoured if one were to
do it. Doing an action of that type is already favoured, even if one
doesnt do it. Otherwise we would be restricted to saying that there
would have been some reason to do it if we had done it. This is not the
same as saying that if we had done it, we would have done something
that there was reason to do.
There are other things to be said about the difference between
favouring and good-making, some of which I tried to say in Ethics
Without Principles. I was discussing there the relation between favouring and right-making (or ought-making)and in my discussion
I made some mistakes, which I hope here to correct. For a while, therefore, my discussion will focus on right-making rather than on goodmaking.8 What stands on the right-hand side of the favouring relation
is an action (or a belief, or a feeling), the action favoured by the reason.
What stands on the right-hand side of the right-making relation? In
Ethics Without Principles (pp. 2223) I suggested that it is a judgement,
the judgement that something is right. But that suggestion involves a
cross-categorization. The contrast I am after is between two substantial
relations, the relation of favouring and the relation of right-making.
Though there will be points to be made about judgement, because
the rightness made is a proper object of judgement while the action
favoured is not, those points are consequent on the distinction between
two relations, which itself needs to be made in other ways. So: what
stands on the right-hand side of the right-making relation? It looks as
if it is the rightness made (though of course it is an action, belief or feeling that has that rightness, in the sort of cases we are thinking about).
And if so, we do indeed have a contrast; one relation has an action as
one of its relata where the other has a property. But here again I am
uneasy. Surely what is favoured is an action, and what is made right is
an action; what is the difference here? We might say that what is made
is the rightness of an action. But couldnt we equally say that favouring
is making favoured, so that our distinction can only be between making
right and making favoured? That would not be much of a distinction.
I had wanted to say something like this, that the favouring relation
is intrinsically normative; it is the relation of calling for. The rightmaking relation is normative only because of the normativity of the
rightness made. Rightness-making, after all, is just one form of the relation of making-it-the-case-that, which is not itself normative at all; it is
the form in which what is made the case is that the action is right. But
couldnt one equally think of favouring as making it the case that the
action is favoured, so that neither relation is more intrinsically normative than the other?
Things do not seem to be going well for the distinction at the
moment. But there is something salvageable here. There is nothing
more to right-making than making it the case that the action is right.
I am grateful here to Douglas Farland for various salutary corrections.
120
jonathan dancy
That side of the story is sound, and so is the claim that any distinctive
normativity here derives from the normativity of what is made. But
though it is true that where there is favouring, what favours makes it
the case that what it favours is favoured, this does not yet establish a
perfect parallel. It is true that by loving Sarah, I make it the case that
Sarah is loved, and loved by me. But that does nothing at all to show
that loving is just one form of the general making-it-the-case-that relation. It is not. It is a relation whose right-hand side is a person (let us
assume). And I want to say something similar about favouring: it is the
relation of calling for, and calling for something is not the same relation
as that of making it the case that something is called for.
Further, the concept of right-making assumes as given the concept
of rightness, and that concept is not (in the relevant sense) relational;
that is to say, if rightness is a relational concept, this is not because
whatever is right is made right by something. We understand this form
of the making-it-the-case-that relation by appeal to an independent
understanding of what it is that is made. Things are the other way
round with the favouring relation. We understand what it is to make it
the case that something is favoured in terms of a prior and essentially
relational notion of favouring, just as we understand what it is to make
it the case that someone is loved in terms of a prior and essentially relational notion of loving.
Understood in this way, we can say that what lies on the right-hand
side of the right-making relation is a proper object of belief, something
capable of being the case, while what lies on the right-hand side of the
favouring relation is an action, which is not a proper object of belief but
something capable of being done. A second difference is that the righthand side of the favouring relation is something that is not capable of
being the conclusion of an inference. One can conclude an inference by
doing something, but one cannot infer an action from anything. The
right-hand side of the right-making relation seems, by contrast, well
set up to be the conclusion of an inference. In this sense the favouring
relation is more essentially practical than is the right-making relation.
There is a spin-off benefit of these thoughts. It seems to me that
Christine Korsgaards complaints about what she calls substantive realism derive largely from failing to see the difference between the two
relations. She writes, for instance:
According to substantive realism, then, ethics is really a theoretical subject. When we ask ethical questions, or practical normative questions
more generally, there is something about the world that we are trying to
122
jonathan dancy
11
I call this a buck-passing view by analogy with the now standard buck-passing
view that to be good is to have reason-giving features. The view in the main text passes
the deontic buck; the standard view passes the evaluative buck. For all this, see my
Should We Pass the Buck?, in A. OHear (ed.), The Good, the True and the Beautiful
(Cambridge: Cambridge University Press, 2000), pp. 15973.
part III
moral realism
126
russ shafer-landau
in the mouths of the men on the Clapham omnibus, so as to more precisely capture popular thinking in this area. In short order, the most
influential of these arguments are the following:
1.If morality is objective, then it cannot be the result of human
creation.
2.If morality cannot be the result of human creation, then it must be
authored by God.
3. Therefore if morality is objective, then it must be authored by God.
1.Genuine moral requirements entail categorical reasons of obedience.
2.Categorical reasons are possible only if underwritten by God.
3.Therefore genuine moral requirements presuppose Gods existence.
1.Genuine moral requirements must exert a reliable motivational
influence on those who are bound by them.
2.Moral requirements can exert such an influence only if they are
backed by divine sanction.
3.Therefore genuine moral requirements must be backed by divine
sanctions.
4. Therefore genuine moral requirements exist only if God exists.
Each of these arguments contains at least one premise that expresses a
philosophical claim accepted by most philosophers today. Still, I think
that each argument is unsound. And so do most other philosophers.
Given that, it may seem that theres little here worth discussing. The
received view nowadays is that the fate of moral realism is not tied to
that of theological realism, and so we can entirely leave aside matters in
the philosophy of religion when doing metaethics. I am not so sure.
Thats not because I am worried that objective morality may after all
require Gods existence. Rather, I am concerned about the pressure that
arguments for religious skepticism place on arguments favoring moral
realism. For the considerations in favor of such skepticism also seem to
cast strong doubt on the merits of moral realism.
There are threereally, fourarguments in this connection that
seem especially worrisome. I have space here to discuss only the last of
these, but I want to sketch them each so as to give a fair impression of
the scope of the relevant concerns.
The first argument is a form of genealogical critique. It does not seek
to vindicate atheism, or moral nihilism, but rather to show that any
positive theistic or moral belief is epistemically unjustified. On such a
view the origins of our beliefs in a given area are directly relevant to
their epistemic merits. Most people hold the religious or moral beliefs
they do because of the way that they have been raised, or because of the
company they now keep. But there is no reason to suppose that such
influences track whatever truth there might be in these domains, especially since such influences have led people to contradictory beliefs. If
there are moral or religious facts, then given how our beliefs in these
areas have arisen, it would be sheer luck were they to land on the truth.1
But such luck vitiates the epistemic credentials of the associated beliefs.
Thus even if there is a God, and even if there is a set of strongly objective moral truths, our beliefs about them are epistemically unjustified.
The second argument is one from disagreement. The argument
really, a pair of argumentsbegins with the observation that the
breadth and depth of moral and religious disagreement is far greater
than that found within the natural and mathematical sciences. The first
version of this argument seeks to draw an ontological point from this
observation. On this line, the best explanation of such disagreement is
that there is no objective reality awaiting our discovery.2 Theists and
moralists disagree so often and so intently because they are projecting
their conflicting commitments or sentiments on a world that contains
no god, and no moral facts. The alternative accountthat half of those
involved in such disagreements have faulty belief-forming mechanisms
which stand in the way of their appreciating realityis both less parsimonious, and requires the presentation of a defensible religious and
moral epistemology, neither of which, it is claimed, is forthcoming.
The second version of the argument from disagreement is more
modest, and claims that even if there is a god, or objective moral facts,
the widespread disagreement on moral and religious matters undermines any justification we might have for our moral and religious
beliefs.
The last of the arguments says that we have most reason to deny
Gods existence, because God is explanatorily superfluous. God must
go the way of the ether, and the Aztec divinities. We havent any need of
1
The argument in the moral domain has been pressed in especially forceful ways by
both Sharon Street and Richard Joyce. See Street, A Darwinian Dilemma for Realist
Theories of Value, Philosophical Studies 127 (2006), pp. 10966, and Joyce, The Evolution
of Morality (Boston: MIT, 2006), ch. 6.
2
See J.L. Mackies so-called argument from relativity for a now-classic statement of
this worry, in his Ethics: Inventing Right and Wrong (New York: Penguin, 1977),
pp. 3840.
128
russ shafer-landau
these things to explain our encounters with the world. Ditto for God
all that he was once invoked to explain can nowadays be more simply
accounted for by reference to other, wholly natural phenomena. Belief
in God is therefore epistemically unjustified; further, belief that God
does not exist is highly justified.
The parallel with moral facts, realistically construed, is straightforward. Moral facts are unable to explain our observations of the way the
world works; an actions being morally required, or there being a moral
reason opposing some policy, does not explain why anything occurs as
it does. Moral facts are certainly unable to explain the goings-on of the
inanimate world. And even where they have most promise to explain
the animate worldnamely, in revealing why people do what they
dowe can more parsimoniously account for such goings-on by citing
non-moral facts about human psychology. Since moral facts are
explanatorily superfluous, therefore, we have most reason to believe
that they do not exist.
Lets begin with a more careful presentation of this last argument,
which will be the focus of the ensuing discussion. It is a variation on a
familiar argument, owing to Gilbert Harman,3 that the explanatory dispensability of moral facts entitles us to deny their existence.
Here is an initial go:
Something exists only if it is required in the best explanation of what we
believe, do or observe.
Neither religious nor moral facts are thus required.
Therefore there are no religious or moral facts.
We neednt consider the second premise just yet, for the argument is
unsound even if it is true. The first premise is false. The worlds contents
are not limited by what manages to satisfy our explanatory needs. What
serves such needs is a function of our interests and our ignorance. But
the true contents of our world are not fixed by such factors. Much of
what exists in the universe fails to figure in our best explanations, either
because we are ignorant of it, or ignorant of or uninterested in the
explananda that it would account for.
It is better, then, to conceive of the argument as having an epistemological, rather than an ontological, point. Thus conceived, it would go
something like this:
3
See his The Nature of Morality (Oxford: Oxford University Press, 1977), pp. 310,
and his Moral Explanation and Natural FactsCan Moral Claims be Tested against
Moral Reality?, Southern Journal of Philosophy, suppl. 24 (1985).
130
russ shafer-landau
8.There is no reason to believe in moral or divine facts, and some reason to deny their existence.
9.Therefore there is most reason to deny the existence of divine and
moral facts.
I hope it is clear by implication that the reasons throughout are epistemic reasons, and not other types, such as prudential ones. If it isnt,
let me make it explicit now.
My strategy here will be to present what I consider to be the best
replies on behalf of the moral realist, and then see whether these replies
allow the theist to resist this extended chain of reasoning. Call that
chain, from (1) to (9), the Explanatory Argument.
Ill spare you a detailed examination of each premise, and instead
sort all moral realist responses into two large groups: that of the
naturalist, and that of the nonnaturalist. Naturalists have a ready
reply to these arguments. They will accept the explanatory requirements asserted in premises (1) and (4), but claim that moral facts can
meet them. By the naturalists lights, moral facts are ordinary, run-ofthe-mill, natural facts, i.e., empirically discernible facts. Such factsas,
for instance, that a great deal of pleasure is generated by an
actionmay indeed play an explanatory role in accounting for our
beliefs and experiences. So, on the naturalist line, premises (2) and (5)
would be false. That would explain why we can resist these skeptical
arguments.
Thus if moral naturalism is true, then moral realists will have a satisfactory reply to the Explanatory Argument. Since the reply relies on
invoking moral naturalism in defense of moral realism, a parallel route
is unavailable to the theist. For that parallel route would be to enlist
theological naturalism in defense of theological realism, and that
ongoing, Western conceptions of Godjust doesnt make sense.
Theological naturalism would seek to vindicate the explanatory necessity of divine facts by revealing them to be a species of natural facts. But
that simply runs counter to the essence of the monotheistic tradition,
which depicts God and his attributes as supernatural.
Theists might nevertheless accept the explanatory requirements of
(1) and (4), and target the negative claims of (2) and (5). Theists cannot
implement this strategy by invoking naturalism. If theists accept the
explanatory requirements, they will have to show that natural explanations are incomplete, and that they require supplementation from the
supernatural for their success. I think that most of our naturalistic
explanations are in fact incomplete. But that gives no reason to suppose
132
russ shafer-landau
sort, say, that condemn genocide and slavery. On this line, those who
are appropriately sensitive to mathematical or moral reality can explain
the event of their believing, as well as the content of their beliefs, by
citing the mathematical or moral facts that have caused these doxastic
events and contents.
This may be true. But to show it so, we would have to show that alternative explanations are not as good. And the problem, at least in the
moral case, is that there are quite plausible alternative explanations
(those offered by error theorists and expressivists), ones that do not
invoke a moral reality that agents can accurately appreciate. Why
require moral truths to best explain our moral views, if we can cite the
social, parental and psychological factors that appear to so heavily
influence their content? This is the essence of Harmans challenge.
There may be a good answer to it. After all, we do need to go beyond
such factors to best explain why physicists hold the physical views they
do, biologists the evolutionary views they do, etc. But it seems that we
are licensed to move beyond purely genetic accounts of doxastic practices in the natural sciences because the phenomena under study possess recognized causal powers. We cannot assume that moral facts
enjoy the same status without begging the question in this context.
So at this juncture we are not in a position to vindicate the causal
efficacy of moral facts by claiming that they play an ineliminable role in
our holding the moral beliefs we do. We can successfully defend such
an argument only by revealing the flaws in anti-realist accounts. I dont
think that there is any short way to do this.
Consider, then, a different line of argument, meant to demonstrate
the causal efficacy of moral facts, that in its essentials can be run in a
way that is compatible with either moral naturalism or nonnaturalism.
The basic idea is this.4 Many nonmoral facts are counterfactually
dependent on moral facts. If certain moral facts had not obtained,
other nonmoral facts would not have obtained. This dependence may,
with qualifications, suffice to establish genuine causal power. Since
moral facts can often pass such a test, moral facts may possess genuine
causal power.
4
See David Brink, Moral Realism and the Foundations of Ethics (Cambridge:
Cambridge University Press, 1989), pp. 19296; Nicholas Sturgeon, Moral
Explanations, in D. Copp and D. Zimmerman, Morality, Reason and Truth (Totowa,
NJ: Rowman and Allanheld, 1985), and Sturgeon, Contents and Causes: A Reply to
Blackburn, Philosophical Studies 61 (1991), pp. 1937; Brad Majors, Moral Explanation,
Philosophy Compass 1 (2006), pp. 910; Graham Oddie, Value, Reality and Desire
(Oxford: Clarendon Press 2005), pp. 198203.
As all proponents of this strategy have acknowledged, counterfactual dependence has its limitations as a sufficient condition of causal
efficacy. Relata with common causes, and instances of backtracking,
pass the test but rarely reveal genuine causal relations. Exceptions can
also arise when the test is used as a necessary condition for causal
power. It fails, for instance, in cases of preemption and overdetermination. But barring aberrant cases, a counterfactual testhad X not
occurred, Y would not have occurred as it didis a reliable measure of
somethings causal (and hence explanatory) power. Lets proceed on
this assumption and see where it gets us.
Suppose that we invoke a moral fact to explain a nonmoral fact. The
employee pension fund is now drained; what accounts for this? It can
be perfectly natural to cite the venality, greed and moral corruption of
the corporate executives who perpetrated the fraud. And we might say
that we should respect the appearances, until we have good reason to
doubt them.
Unsurprisingly, there is a reason to doubt the appearances here, and
to deny that the citation of moral facts in everyday causal talk reveals
anything about the real contents and structure of the world. The reason
is that moral facts, by all accounts, supervene on causally efficacious
nonmoral ones, and since we should be loath to countenance widespread causal overdetermination, it is (say the critics) really the subvening nonmoral facts that are doing the causing, rather than the moral
facts that supervene on them.
That might be so. But there is reason to resist the criticism just
described. We are assuming that in standard cases, X causes Y if and
only if Y would have been different had X been different.5 Not only do
moral facts often pass this test, but, importantly, the nonmoral facts on
which they supervene often fail it. If we kept everything else fixed, but
It is important to note that almost every standard case is one that involves
supervenient phenomena; the only ones that dont are cases of causation at the fundamental physical level (assuming there is such a thing). All of the facts we inquire
aboutunless we are physicistsare supervenient facts. So we shouldnt seek to discount the applicability of the counterfactual test in the moral case just because moral
facts are supervenient ones; motivational, intentional, social, economic, biological, historical, political (etc., etc.) facts are all supervenient, too. It might, of course, be that all
moral cases are non-standard, but it is hard to see how to defend such a claim in a
non-ad hoc way. Or it might be that no supervenient phenomena really existthat
independent causal power is a prerequisite for ontological credibility, that supervenient phenomena lack such power, and that they therefore ought to be expunged from
the ontology. But if that were so, then the absence of moral facts would be no surprise,
given the absence of any people, actions, intentions or motivesall supervenient phenomena themselves.
5
134
russ shafer-landau
assumed that the executives were not greedy, venal and corrupt, then
the outcome would have been quite differentthe pension fund would
presently be well-endowed. So these moral failings do seem to be causally relevant. But the pension fund would remain depleted in the nearest natural possible world, one in which the specific natural facts that
constituted the corruption were changed only very slightly. This argues
for the greater causal-explanatory power of moral facts over certain
nonmoral ones.
Since moral properties are multiply realizable, there is a good deal of
variation in the make-up of the natural facts that can instantiate a given
moral property. This variability is what explains why citation of exclusively natural features often fails to satisfy a counter-factual test of
causal efficacya small change to the subvening facts, one that brings
us to the closest natural possible world, would not alter the relevant
outcomes in any way. So if we have to choose between these options in
fixing causal power, then whenever (keeping all else the same) alteration of moral facts would, and change of subvening natural facts would
not, yield a different outcome, we should be willing to credit the moral
facts with enough causal power to allow them to retain a place in our
ontology. After all, we extend the same courtesy to many other types of
supervenient facts, on precisely the same grounds.
There is a worry, of course. It was expressed a good while back, by
David Zimmerman,6 and his critique remains, I think, the best expression of doubt about this sort of strategy. His criticism is that the greater
generality achievable by adverting to supervening facts in causalexplanatory contexts is a double-edged sword. Rather than explain
(say) Hitlers deeds by reference to his depravity or evil nature, we
might more informatively account for his actions by citing the specific
nonmoral intentions and motives that prompted them. There are many
ways of being evil and depraved, some of which Hitler did not exemplify, and so some of which are irrelevant to accounting for his actions.
We therefore gain greater understanding of his behavior if we cite the
more specific nonmoral facts that realized the evil and depravity at particular times and places. In this way we can always replace a moral
explanation with a nonmoral one in order to gain greater understanding of the phenomenon in question.
6
David Zimmerman, Moral Realism and Explanatory Necessity, in D. Copp and D.
Zimmerman (eds.), Morality, Reason and Truth (Totowa, NJ: Rowman and Littlefield,
1985), pp. 79103, at p. 87.
136
russ shafer-landau
And yet there are facts of these kinds that cause and explain things in
our world. Thus reliance on the causal test is insufficient to vindicate a
realistic understanding of the facts or entities that pass it.
I dont think that this points to a special problem for the counterfactual test. It, or something very like it, seems essential to our understanding of causal efficacy. And I dont see any natural amendment to
the test that would enable us, through its use alone, to discriminate
between facts that are best construed realistically, and those that are
not. That is because many conventional facts seem, in the examples
above, and in so many others, to be causally relevant to outcomes in
our world.8
Where does this leave us? We reached this point by casting doubt on
the tenability of the strategy that had us first assuming the existence of
moral facts, and then asking after their explanatory efficacy. Such a
strategy, endorsed by Harman, makes it far too easy to vindicate the
explanatory necessity of moral facts, as becomes clear when we consider the analogous case that might be made for theistic facts. Once we
grant the existence of the disputed class of entities or facts (theistic,
moral, etc.), and make them the relevant explananda, it wont be difficult to establish the need to invoke other facts of the same kind in
accounting for them.
So perhaps we can, without assuming the existence of the disputed
class, defend its existence by means of a counterfactual test of causal
efficacy. Yet as we have seen, though we may be able to do this for moral
facts, that does not suffice to vindicate their realistic status. Those
moral realists who rely on the test to underwrite their position thus
need to look elsewhere.
This isnt to say that the counterfactual test, or a near-cousin, is irrelevant in these debates. It might be, for instance, that counterfactual
dependence ordinarily supplies a necessary, but not sufficient, condition for somethings earning its way into the ontology. If that were so,
then passage of the test would be insufficient to vindicate moral realism, but failure would suffice to refute it.
8
One could deny that conventional facts are causally efficacious, and so claim that
we have good reason to regard only nonconventional facts that pass the causal test as
best construed realistically. Since moral facts are nonconventional, and pass the causal
test, we therefore have good reason to think that moral realism is true. The problem
with this is that we cant simply assume that moral facts are nonconventional. Passage
of the causal test was supposed to show that they are. But it cant do that, as the examples from the law and etiquette have shown.
If we insist on causal relevance as a necessary condition of ontological credibility, then a good deal more must be done to more precisely
fix the content of a causal test, and to show that moral facts are (un)able
to pass it. Those who enjoy strenuous exercise are encouraged to take
this path. But I think there is good reason to sidestep it.
If we proceed on a neutral basis, and assume neither that moral facts
exist, nor that they dont, then it is as yet unclear whether they are
explanatorily indispensable. If they are, then moral realism is so far in
the clear. But I want to proceed here on the assumption that we can do
all of the relevant explaining without invoking moral facts. My goal is
to show that even on this worst case scenario, the moral realist has a
defense against the Explanatory Argument that is largely successful.
This defense, I think, is unavailable to the theist. And so the parity that
is asserted at the heart of the Explanatory Argument can be resisted.
If we grant, for purposes of argument, that moral facts will fail the
explanatory tests, then nonnaturalists are faced with a real worry. For
theres no doubt that we have used these explanatory tests to very good
effect in accounting for why were no longer warranted (if ever we were)
in believing in ghosts, UFOs, or snake gods. I think, in fact, that the
explanatory tests are excellent devices for assessing the epistemic merits of our ontological commitments in a very broad, but not unlimited,
way. Specifically, their good work is limited to the natural realm, the
one whose contents are (at least in principle) empirically discernible.
How convenient for me! But there is a case to be made for such a
limitation. For the explanatory requirements, if strong enough to eliminate moral facts, will almost certainly eliminate all normative facts. If
we assume that moral facts lack explanatory power, there is every reason to make a similar assumption regarding all normative facts. That
Iought to believe that So Paulo is in Brazil would not explain why I do.
Testimony, and my past acquaintance with maps of South America,
would explain my belief. That I would be irrational for failing to dress
warmly in the snowy season would not explain why I have done so.
Rather, my wanting to stay warm, and knowing how to do it, could do
the needed explaining. That I am morally required to support my children would not explain why I do. That I believe that I am thus required,
in addition to the fact of my direct love for my children, and my ability
to support them, would together explain my support. As a general matter, its quite implausible to suppose that normative facts are going to
directly account for non-human events. But even when it comes to
human agency, there are explanations in terms of our psychological
138
russ shafer-landau
profiles and opportunities for action that might be capable of doing all
the needed explaining.
Normative properties are individuated by the standards they set, the
reasons they represent. Predictive or causal-explanatory failure is no
strike against the credibility of a normative standardthat people dont
adhere to it is not, by itself, good evidence that its content is mistaken.
Now these considerations, even if correct, do not amount to an argument for jettisoning moral and other normative facts from our ontology. We get that result only if we supplement these considerations with
the explanatory requirements in (1) and (4). Since I want to argue for
our justified belief in moral facts, I have to showagain, on the assumption that moral facts dont pass the explanatory requirementsthat
these requirements dont apply to them. I think that there is a direct
way to do this:
10.If the explanatory requirements (i.e., premises (1) and (4)) are both
correct, and perfectly general in application, then we have most
reason to deny the existence of all normative facts.
11.We dont have most reason to deny the existence of all normative
facts.
12.Therefore the explanatory requirements are either false, or, if true,
are limited in their application.
Let us proceed by assuming the truth of (10), which represents a
generalization of the concession I have been making for purposes of
argument. Whatever grounds there are for thinking that moral facts are
explanatorily dispensable will, so far as I can tell, generalize to all normative facts. If (10) is false, then whatever explains its falsity will also,
presumably, justify the explanatory necessity of moral facts. And while
(per our earlier discussion) this is not enough to vindicate moral realism, it is enough to insulate moral facts from the Explanatory Argument.
So anti-realists can have no quarrel with the concessive assumption
that (10) is true.
In support of (11), consider first some hopefully uncontroversial
examples: it is a fact that it is sometimes appropriate or legitimate or
justified to pursue ones happiness, a fact that one ought to believe in
ones own existence, a fact that one often should pursue efficient means
to ones chosen ends. More importantly, we can show that (11) is plausible by showing how implausible its negation is. The negation of (11)
that we have most reason to deny the existence of all normative
factsis, if true, a normative fact. So, by its own lights, we would have
most reason to deny it. If we have most reason to deny the negation of
(11), then we have excellent reason to affirm (11).
In other words, the explanatory requirements, and the conclusions
they are meant to help establish, are all themselves normative claims.
They tell us what we have reason to do or believe. But if we deny the
existence of all normative facts, then the conclusions of the skeptical
arguments are ones that we have most reason to deny. Since I am
assuming that the proponents of those skeptical arguments would find
such a view unacceptable, even they should accept the truth of (11).
Here is where we stand. I have tried to vindicate (12), by granting to
anti-realists the truth of (10), and then arguing for (11). If I have succeeded, then we either have to say of the explanatory requirements that
they are false, or that they do not apply to the normative realm. I prefer
to say the latter, just because of the excellent work they have done in
justifying the rejection of beliefs in phlogiston, the Greek pantheon,
Easter bunnies, etc. In other words, I think we are right to apply the
explanatory requirements when seeking to determine the natural contents of our world.
And this provides a basis for drawing a disanalogy between the
moral and the divine. The job description of normative facts does not
include the possession of explanatory power. They may indeed possess
such power, though I am assuming here that they do not. But because
their functional role does not require that they explain nonnormative
phenomena, but rather that they specify ideals, requirements, or standards that in some way must be met, an explanatory failure does not
license their expulsion from the ontology. By contrast, the job description of God does include explanatory power. Indeed, such power is
causal-explanatory power, which, moreover, must sometimes be exercised. God must, at the least, get the universe going, and will further, to
all but deists, intervene in our affairs at least occasionally in such a way
as to vitiate the causal closure of the natural.
Yet it is far from clear that supernatural explanations are superior to
natural ones in accounting for the goings-on in our world. Supernatural
explanations are certainly less parsimonious than natural ones. When
a battle is won, for instance, we can say, if we are so inclined, that God
had a hand in the victory, but what we must say is that these particular
soldiers, acting in these particular ways, in these specific circumstances,
carried the day. Perhaps a divine hand was guiding the combat, but
there is no need to invoke such an explanation when the natural one is
apparently complete. The lesson is easily generalizable. Since theistic
140
russ shafer-landau
142
russ shafer-landau
the ontological argument were sound, and we could know this, then
that would suffice to justify belief in Gods existence, even if we were
unable to show that divine intervention was the best explanation of
anything that we had experienced.
This strategy bears an obvious resemblance to the sorts of intuitionist epistemological strategies sometimes adopted by moral realists. The
crucial difference is that intuitionists are not claiming to gain knowledge of any existential truths in an a priori way. If there are any a priori
moral propositions, they are restricted to ones that express moral principles (which, as I understand them, are conditionals), rather than
moral facts (which, by my lights, are instantiations of moral properties). So, for instance, one might be able to know a priori that killing
solely for enjoyment is immoral, though one cannot know a priori
whether any given instance of killing is immoral. And that is because
one cannot know a priori whether moral properties have been instantiated. One must know something of the worlds contingencies in order
to have such knowledge. Thus moral facts are not knowable a priori,
even if some moral principles are.
So, although some moral realists are perfectly comfortable invoking
the possibility of conceptual truths and a priori knowledge in moral
matters, they deny that we can know a priori either (i) that there are
moral facts, or (ii) that the specific moral facts are as we take them to
be. We must have some experience of the world in order to know that
there are moral features in it. We must know, at the very least, that there
are beings with intentions and deliberative powers, vulnerable beings
who can suffer, etc. There can be no moral facts in the absence of moral
agents. And whether there are any moral agents about is not something
we can know a priori.
Compare the situation for the theist. It might be a conceptual truth,
and hence a priori knowable, that if there is a God, then God possesses
more knowledge than human beings do. But the theist needs more
than this, of course, since atheists can sign on to such principles. What
the theist needs is a justification for believing that God exists, and such
a justification must, if I am right, proceed in one of two ways. Either the
theist must show, a priori, that God exists, or the theist must show, a
posteriori, that God is explanatorily indispensable.
Both such argumentative burdens are different from those inherited
by the moral realist. Moral realists cannot show, a priori, that moral
facts exist. (Nor do they need to.) Nor does the realist have to show that
moral facts are essential in explaining why the nonmoral world is as it
is. Thus the argumentative burdens on the moral realist differ very
importantly from those of the theological realist.
This concludes my assessment of the explanatory argument that
seeks to establish parity between the vulnerability of moral and theological realism. The take-home message here is that there is an important disanalogy between the cases that can be made against moral facts
and theistic ones. The explanatory requirements are best construed as
limited in their application to entities whose job description includes
having a discernible impact on the natural world. Since moral facts do
not, but many divine facts do, have it as part of their job description to
causally intervene in that world, moral facts are excluded from the purview of the requirements. Divine facts are not. And therefore divine
facts must earn their way into our ontology by passing the explanatory
requirements; their failure to do so would (absent a sound a priori
argument for Gods existence) provide us most reason to deny their
existence. Moral facts, by contrast, are immune to the explanatory
strictures. They might nevertheless manage to satisfy such demands.
But even if they fail such tests, we do not, on that account, have most
reason to deny their existence.
Postscript
I havent yet said anything that would give us reason to think that there
are moral facts. Indeed, on the very generous assumption that everything Ive argued for thus far is correct, we are left with a very large
question: if passing an explanatory requirement is not the appropriate
path to normative vindication, then what is? Why are we licensed (if we
are) in believing that there are moral facts?
Lets distinguish two questions that might tend to get entangled here.
The first is the one just asked: what reason do we have for thinking that
there are moral facts? The other, narrower question is: what reason do
we have for thinking that the specific moral facts are as we take them to
be? The former question is the more fundamental. If we cant supply a
positive answer to it, then theres no point in considering the second
question.
The answer to the first question is, at a sufficiently general level, the
same answer that we return to the question of what licenses belief in
the existence of natural facts. We believe that there is an external world,
a world of natural facts, because such a belief unifies a great deal of our
144
russ shafer-landau
148
michael ridge
the very idea of an irreducible normative property begins to look somewhat obscure. The challenge put to anti-reductionists is to explain how
there could be normative properties which both necessarily supervene
and are irreducible. In this paper I want to consider and rebut two
attempts to deal with the challenge.
The first strategy for dealing with this challenge has recently been
put forward (in unpublished work in progress) by Nicholas Sturgeon.
Sturgeon argues that while most metaethical views can accommodate
some form of supervenience, there is no single formulation of supervenience which is hospitable to all prima facie plausible metaethical
views. Just which form of supervenience one endorses will on this
account depend on ones background meta-ethical theory. If this argument succeeds then it puts pressure on the anti-realist who argues that
non-naturalists cannot explain supervenience. For perhaps the only
form of supervenience which non-naturalists can be expected to accept
without begging any central questions is indeed a form of that doctrine
which they can, after all, explain.
In section one, I respond to Sturgeons interesting line of argument
in a concessive way. I think we do indeed learn something important
from Sturgeons critique, but the lesson is not that there is no form of
supervenience which can be used as a dialectical weapon in these
debates. Instead, Sturgeons own argument actually suggests a formulation of supervenience which does not beg any central questions, and
which therefore could reasonably be used in a premise against one or
another metaethical view. In section one I try to show how a suitably
disjunctive characterization of the subvenient base avoids any reasonable controversy of the sort Sturgeon has in mind. In the remainder of
this paper I try to show why anti-reductionist views will have problems
explaining even this form of supervenience.
The second strategy for dealing with the challenge to explain
supervenience takes a more direct approach. Here I focus on Russ
Shafer-Landaus recent attempt to meet the challenge by explaining
supervenience in anti-reductionist terms. In the remainder of the
paper, I argue Shafer-Landaus strategy fails, and that the failure is
instructive. For Shafer-Landaus strategy seems like one of the most
promising ones available to the anti-reductionist. Shafer-Landau tries
to show how what he takes to be the most promising strategy for
explaining supervenience in the philosophy of mind can be carried
over successfully to the normative domain. In particular, he argues that
in both cases, we can adequately explain supervenience by appealing to
a thesis about how each instantiation of any of the supervening properties is fully constituted by some concatenation of the relevant subvening properties.
I argue here that even if this strategy works in the philosophy of
mind, it remains problematic in the normative case. The possibility of
such a divide between the normative realm and other cases of supervenience is itself very instructive. For anti-reductionists (Shafer-Landau
included) very often rely on companions in guilt style arguments.
These arguments rely on the idea that a supposed problem in the normative case is really just a variant on a problem encountered in other
areas of philosophy, and that whatever solution works in that other area
can be carried over to the normative case. The more general lesson of
my argument against Shafer-Landaus proposed explanation is that
such companions in guilt arguments should be viewed with considerable suspicion. Strategies which are perfectly plausible in other domains
may not carry over plausibly to the normative, or so I shall argue. First,
though, we need to clarify just how the supervenience of the normative
should itself be understood.
1. Supervenience: Formulating the Challenge
Supervenience theses come in all different shapes and sizes, though.
Debates in metaphysics and the philosophy of mind have led to the
development of a wide array of different supervenience theses.2 How
exactly should we formulate supervenience in the context of challenging the normative anti-reductionist to explain it?
Supervenience theses can be global or local. Global supervenience
theses claim that no two possible worlds can differ in the supervening
respect without also differing in the subvening respect. By contrast,
local supervenience claims hold that two numerically distinct objects
in one and the same world cannot differ in supervening respects without also differing in the subvening respects.
One problem for local supervenience claims is that if the subvening
base is widened far enoughsay, to include spatio-temporal location
then supervenience may be trivially true for the reason that no
2
For a useful survey, see K. Bennett and B. McLaughlin, Supervenience,
Stanford Encyclopedia of Philosophy (2005), http://www.seop.leeds.ac.uk/entries/
supervenience/#4.3.
150
michael ridge
umerically distinct things can be precisely the same in all of the subn
vening respects. In which case it becomes trivial that no two things can
differ in the supervening respects without also differing in subvening
respects as well. This will be trivially true for the simple reason that no
two genuinely distinct things can fail to differ in some of the subvening
respects. Global supervenience theses skirt this problem by being formulated in terms of whole possible worlds, rather than in terms of individuals within the same world. The issue then becomes which kinds of
worlds really are genuinely possible, and which are not. I propose to
understand supervenience here as a global thesis, so as to avoid these
sorts of triviality worries.
The other issue is how we should understand the subvening
base. Here I want to consider a worry recently discussed by Nicholas
Sturgeon.3 The subvening base is sometimes characterized in terms of
the natural properties, but the question of just which properties count
as natural is itself deeply vexed. Moreover, on many ways of understanding the natural it will not be true that the normative supervenes
on the natural. Plausibly, supernatural facts about the next life can give
us reasons for action, for example, in which case two worlds could differ normatively without differing naturallysupernatural differences
in ones fate in the afterlife, or differences with respect to whether one
even has an afterlife, could be enough to explain differences in ones
reasons for action.
A seemingly more promising approach, often taken in these debates,
is to maintain that the normative supervenes on the non-normative.
This is indeed very plausible, but it is not clear that it is analytic. Here
the issue seems to be too theory-dependent for analyticity. For some
forms of descriptivism are committed to denying this sort of supervenience thesis. Let me explain why this way of defining supervenience
may be inhospitable to certain forms of descriptivism.
Consider the descriptivist view that being worthwhile just is
being pleasant. This may not be a very plausible view, but it is also not
3
In what follows, my thinking has been influenced by an as-yet unpublished paper,
currently in draft, by Nicholas Sturgeon on supervenience. See Sturgeon (unpublished
manuscript), Doubts About the Supervenience of the Evaluative. To be clear, as of the
time of my writing this article, Sturgeons own views are still potentially in flux, as he
tells me that his paper is not in his view ready for publication. I therefore do not attribute any of the ideas discussed in the text to him in the sense of representing his considered view, but only in the sense of being his ideas, regardless of whether he still endorses
them as true or not.
4
See F. Jackson, From Metaphysics to Ethics (Oxford: Oxford University Press,
1998), p. 120.
152
michael ridge
154
michael ridge
Shafer-Landau actually aims to explain a slightly different supervenience thesis than the one discussed here, though. He is interested in a
supervenience thesis which holds within a single possible world, and
maintains that no two things within that world have the same normative status without differing in some descriptive respects. It seems clear,
though, that he would extend this strategy to deal with the global
supervenience thesis advanced here.
Shafer-Landaus positive explanation of supervenience dovetails
very nicely with the companions in guilt strategy. For the appeal to
constitution is supposed to be borrowing the best strategy for the companions. Shafer-Landau gives the examples of a pencil and its material
constitution and that of a subjective feeling and its neurophysiological
(and perhaps intentional) constituents.
Crucial to the success of this strategy is the idea that constitution is
distinct from identity. Otherwise, the view on offer would collapse into
a form of reductionism, with normative properties being identical with
certain constellations of descriptive properties. In itself this is not a
problem, since constitution by is indeed distinct from identity with.
Constitution is asymmetric; identity is symmetric. A clay statue is
constituted by the lump of clay, but the lump of clay is not constituted
by the statue; the clay is instead constituted by certain molecules, which
themselves are not constituted by the lump of clay, but by various
other subatomic particles, and so on. Whereas if a = b then it follows
that b = a.
However, matters are not so straightforward. One obvious disanalogy between the normative case and the case of companions in guilt is
that the supervenience of the normative on the non-normative and
descriptive is plausibly a priori, and indeed analytic. By contrast the
supervenience of the mental on the non-mental is hardly a priori, much
less analytic. In fact, the majority of human beings at various points in
history may well have denied the supervenience of the mental on the
non-mental out of an antecedent commitment to a certain sort of dualism about the mental. For many religious traditions would naturally
lead one to such a view of the mental. The same point holds about the
material constitution of macroscopic objects by subatomic particles.
Perhaps even more obviously, this constitution relation is not a priori,
much less analytic.
Shafer-Landau does not explicitly address this disanalogy in his
discussion, but there is at least one interesting move that an antireductionist can make here. The anti-reductionist can argue that the
156
michael ridge
preceding objection simply assumes that the analyticity of the supervenience of the normative on the non-normative/descriptive must be
taken as a theorem, rather than an axiom in the following sense. The
argument assumes that whatever metaphysically explains the supervenience of the normative on the non-normative/descriptive must
itself be analytic. For otherwise the explanation of supervenience will
not explain the analyticity of the supervenience thesis.
The anti-reductionist can simply reject this premise, though. The
anti-reductionist can insist that the analyticity of supervenience is axiomatic, and not derived as a theorem. It is, the reply continues, simply
a brute fact about our normative concepts that something cannot count
as rightness, for example, unless it obeys a suitable supervenience constraint. Crucially, on this view the explanation of how any given property could actually satisfy the supervenience constraint need not be
analytic, or even a priori. Indeed, at some points, this seems to be
Shafer-Landaus own view of the issue of analyticity.9
Here is an analogy that might help clarify the defence I am proposing on behalf of the anti-reductionist. Suppose that I just stipulate that
by shmong I shall refer to that property which obeys such-and-such
supervenience constraint, and which is also such that someone who
fails to do that which is shmong is therefore blameworthy. There shall
now be no mystery about why facts about which actions are shmong
will conform to the relevant supervenience constraintthat follows
trivially from the stipulated meaning of shmong. How any given property can count as being the property picked out by shmong, on the
other hand, is another matter, and that will involve explaining how the
property in question satisfies the relevant supervenience constraint.
The explanation of how a given property satisfies the relevant supervenience constraint need not itself be a priori or analytic, though. For
we do not at this stage need an explanation of why it is analytic that the
shmong facts supervene in such-and-such way; that follows from a
semantic axiom about shmong. All we need is an explanation of how
shmong so defined could manage to refer to a given property, and that
explanation need not itself be a priori or analytic.
9
See Shafer-Landau, Moral Realism, p. 88. See also Shafer-Landaus reply to a critique by Robert Mabrito in R. Shafer-Landau, Replies to Critics, Philosophical Studies
126 (2005), pp. 31329. Mabritos critique of Shafer-Landau is also itself worth reading
for those interested in this debate (see R. Mabrito, Does Shafer-Landau Have a Problem
with Supervenience?, Philosophical Studies 126 [2005], pp. 297311).
158
michael ridge
10
Actually, Shafer-Landau lumps length together with weight and claims that both
are constituted by molecular composition; see Shafer-Landau, Moral Realism, p. 77.
Weight, though, is relative to gravitational field, and so is partly constituted by more
than the objects molecular composition. Mass would be a much better example for his
purposes. This in no way undermines the argument, but it is worth getting the examples just right.
11
I take this account from D. Robb, The Properties of Mental Causation,
Philosophical Quarterly 47 (1997), pp. 17894, though it seems to be fairly standard
fare amongst trope theorists.
160
michael ridge
Within this framework, we can reformulate Shafer-Landaus hypothesis. He suggests that we can best explain the supervenience of the
normative12 on the descriptive by invoking the claim that every normative property is fully constituted by the instantiation of some cluster
of descriptive properties. Within the framework of trope theory, this
turns out to be the claim that every normative trope is constituted by
a suitable cluster of descriptive tropes. Presumably we can here read
constituted by literally as meaning something like made up of , in
which case the idea of constitution is still understood mereologically.
So far, so good.
How does the thesis that every normative trope is constituted by a
suitable cluster of descriptive tropes explain supervenience? By itself, it
does not logically entail supervenience. The thesis is logically consistent
with a given set of descriptive features constituting the instantiation of
a normative property in our world, and yet failing to constitute a normative property in another world which is otherwise identical to ours.
In order to explain supervenience, we need some auxiliary premises.
The same sort of point in the philosophy of mind emerges nicely in
David Robbs discussion of how trope theory can be deployed to explain
the supervenience of the mental on the physical. Robb offers a thesis
he calls trope monism which is very close to the philosophy of mind
analogue of Shafer-Landaus thesis about the normative. In Robbs
sense, trope monism is the thesis that every mental trope just is a physical trope, though mental types are not identical with any physical types.
This is not precisely the analogue of ShaferLandaus thesis, which
claims only that every normative trope is constituted by a descriptive
trope.
The point is that on either way of proceeding, some further assumptions are needed to explain supervenience:
Trope monism, as I have formulated it, entails Supervenience (which,
I shall assume, concerns types) so long as we are allowed two plausible
assumptions: (a) if x = y, then necessarily x = y, and (b) if a trope is an
12
Here I take some liberties with Shafer-Landaus text. He focuses exclusively on the
moral, and not on the normative more generally. However, I strongly suspect that he
would want to generalize his account to include non-moral reasons for action too, and
to defend a kind of anti-reductionist realism there too. Rather, he would have wanted
to extend the strategy in question when he still believed in this whole approach. ShaferLandau now rejects the strategy discussed in the text on the strength of some of the
arguments in the textmost specifically, the argument that constitution alone does not
explain supervenience.
13
162
michael ridge
anti-reductionist terms, are as much as possible, given the modal properties attributed to them by supervenience. The fact that supervenience
itself is a trifling and analytic truth does nothing to dispense with this
very real explanatory burden.14
In fairness to Shafer-Landau, I should note that he has indicated (in
correspondence) that he is not sure he wants to adopt the Robb strategy for explaining supervenience. In fairness to myself, I should note
that he has not suggested a particular alternative to Robbs approach.
Moreover, the general appeal to whatever works in the philosophy of
mind seems to make it fair game to argue that Robbs approach is what
works in the philosophy of mind, yet will not work in the case of normative properties. In which case, the appeal to companions in guilt
should not ease our worries about normative supervenience. Finally, of
course, apart from Shafer-Landaus own view, it is interesting in its own
right to see whether a variation on Robbs approach could be made to
work in the case of normative properties.
I now present four challenges for the defender of this approach. The
first challenge stems from the fact that the bulk of the explanatory work
here is done by the thesis that (b*) if a trope constitutes an F-trope,
then it necessarily constitutes an F-trope. Why, though, should we take
(b*) to be true? Constituition is, after all, not identity; that is part of
Shafer-Landaus point. Why, though, could we not have the same trope
elsewhere without its constituting the same F-trope it constitutes here?
Some further metaphysical work needs to be done to vindicate this
hypothesis, or as yet we have no explanation of how normative properties could supervene. Note that Robbs version of the argument has an
advantage on this score. For Robb can explain the analogous thesis that
(b) if a trope is an F-trope, then it necessarily is an F-trope. For this
14
Many thanks to Brad Majors for drawing me out on this aspect of the dialectic,
which is, in my view, all too often submerged as implicit in ways that are confusing
in the existing debates. Here I also find it instructive to note the analogy with the philosophy of mind. If dualism in the philosophy of mind were true then it would be
plausible to infer that the mental does not necessarily supervene on the physical, as
supervenience in this case is not analytic. This shows how we can have irreducible
properties without supervenience. This contrast with the normative suggests an
explanatory itch which deserves to be scratched. How could there be irreducible properties which nonetheless supervene in the relevant ways? Perhaps the challenge can be
met, but it cannot be dismissed as not posing a genuine challenge once we distinguish
the issue of supervenience itself qua trifling analytic truth from the metaphysical question of how there could be normative properties with the modal profile entailed by
supervenience.
15
16
164
michael ridge
one of the main virtues of anti-reductionism. Because it is entirely neutral on the first-order questions of the day, it can much better accommodate the sorts of intuitions that underwrite the force of G.E. Moores
Open Question Argument.
Why do I say that Shafer-Landaus explanation potentially rules out
certain forms of monism? Because according to trope nominalism, a
property just is a set of resembling tropes. Whether a given property is
normative or descriptive depends entirely on whether the relevant
resemblances are normative or descriptive (the or is inclusive). Antireductionism just is the thesis that normative properties are not identical with descriptive properties, however. In the context of trope
nominalism, this means that the tropes that make up a given normative
property cannot resemble one another in ways that would be salient
from a purely descriptive perspective.
Somewhat surprisingly, this is enough to rule out certain forms of
normative monism. For normative monism will hold that just one kind
of fact can constitute a reason, and the tropes that make up this sort of
fact may well be ones that resemble each other in important ways from
a purely descriptive point of view. For example, take a simple form of
egoistic hedonism, according to which a fact is a reason for an agent to
perform an action just in case it is a fact to the effect that the action
would promote the agents pleasure. The property of promoting ones
own pleasure may, though, be highly salient from a purely descriptive
point of view; in which case the tropes which make up this property
will indeed resemble one another in ways that are salient from a purely
descriptive point of view. After all, empirical psychology may well
make heavy use of this property in explaining the actions of intentional
agents, for a start. This, in turn, would entail that the property of being
a reason is, after all, a descriptive property, which is inconsistent with
anti-reductionism.
The anti-reductionist has at least two replies. The first is just to bite
the bullet, and perhaps be so bold as to claim as a virtue for his theory
that it allows us to refute egoism! This, though, would be to bite the
dialectical hand that feeds you. For ruling out egoism on purely metaethical grounds seems incompatible with the spirit, if not the letter, of
G.E. Moores Open Question Argument, which is itself perhaps the
strongest argument for antireductionism in the first place.
The second anti-reductionist reply to the objection from monism is
to reject the trope nominalists account of what makes a property count
as being of one type rather than another (physical versus mental;
17
See, for example, J. Dancy, Ethics Without Principles (Oxford: Oxford University
Press, 2004).
166
michael ridge
The key move for our purposes comes in the sentence I have italicized.
This is enough to entail supervenience, but it hardly looks like an explanation of supervenience. For the italicized sentence just asserts a form
of superveniencethat there can be no moral difference without some
non-moral difference conspiring to make the difference. You cannot
explain supervenience by stating it.
Dancy, Ethics Without Principles, p. 89; emphasis added.
18
19
See W.D. Ross, The Right and the Good (Oxford: Oxford University Press, 2003
[1930]).
168
michael ridge
(namely, Shafer-Landaus attempt) to explain supervenience in genuinely anti-reductionist terms. By way of summary, those four challenges are:
1.A dilemma: Either explain why we should reject trope nominalism
(and hence why the Robb strategy for explaining supervenience in
the philosophy of mind is not the best strategy here too) or
explainwhy trope nominalism plus the proffered explanation does
not implausibly entail the rejection of monism. Either of these
options threatens to bite the dialectical hand that feeds the antireductionist. On the one hand there is the spirit of the Moorean
intuitions behind the Open Question Argument, which would
counsel against a direct metaphysical argument against monism.
On the other hand, there is the fact that the main force of the original argument was an appeal to companions in guilt; once we are
more selective about which companions we invoke, though, we
need an argument that we are not hanging out with the wrong
crowd. In any event, note how little ice the this is just like in the
philosophy of mind explanation is cutting now. We need a lot more
argument about just which of the available stories in the philosophy
of mind is the right one for the anti-reductionist in metanormative
theoryand the anti-reductionist may be forced to say not the
trope nominalism theory and not a standard universals theory
either, but instead a universals plus tropes theory. Needless to say,
its not obvious that this is the right story in the philosophy of mind,
in which case companions arguments really do not advance the
anti-reductionist position very far.
2.Provide some explanation of why we should reify tropes. Why not
go for a more elegant metaphysic that posits only universals and
objects which instantiate them, and doesnt bloat its ontology with
property instantiations as well? This challenge may be especially
forceful if the reply to point (1) above is to take the horn of the
dilemma which rejects trope nominalism. For one of the main arguments for tropes is that given trope nominalism they allow us to
avoid the problems associated with a theory of universals.
3.Mount a defence of the assumption that if a property constitutes
another property then it necessarily does. This assumption seems to
do the lions share of the explanatory work, but is in Shafer-Landaus
discussion a suppressed premise which gets no defence.
4.Explain why this explanatory strategy does not lead all too quickly
to the rejection of holism, or why its so leading is not problematic.
part IV
virtue ethics
174
eric l. hutton
If Han Feizi really did study with Xunzi, then one might expect Han
Feizis criticisms of the Confucian tradition to be especially incisive,
because they would be coming from someone with inside knowledge
of Confucianism. Ofcourse, there is no guarantee that this is the case
upon investigation, we may ultimately decide that Han Feizi simply
misunderstood Confucianism, or that he simply did not present good
arguments against it, either unintentionally or on purpose. Indeed,
some have called his arguments question-begging and outright blind
to the values espoused by Confucians.3 However, here I will contend
that when we reconstruct the debate between Han Feizi and the
Confucians, trying to be sympathetic to both sides and trying not to
make astraw man out of either, we can see Han Feizi as a thoughtful
and powerful critic, one whose arguments are not based simply on misunderstanding, erroneous reasoning, or lack of effort, and who therefore poses a genuine difficulty for the Confucians that has no simple
answer. Let me stress that in making this claim I do not mean to show
that Han Feizi ultimately wins the debate, because that project is far
too long and complicated to undertake here.4 My aim is rather the
more modest one of demonstrating that Han Feizi has put his finger on
a serious problem for Confucian views and that he cannot be easily
dismissed.
In addition, a growing number of scholars have proposed that
Confucianism should be regarded as a form of virtue ethics,5 and this
view makes possible interesting new approaches to understanding not
and other non-English publications, there is more treatment of this matter, though still
less than one might perhaps expect for a topic of such prominence in the text. Cf.
Liangshu Zheng , Han Feizi Zhijian Shumu (Taipei: The
Commercial Press , 1993), pp. 12328, for a list of some of these writings.
3
Cf. Cheng, Legalism versus Confucianism, p. 332. As will become clear, I disagree
with Chengs assessment. However, in order to leave enough space to set out my proposed interpretation, I will not be arguing against him directly. I leave it to my readers
to judge whose reading of Han Feizi is more persuasive.
4
Accordingly, here I am also not going to discuss the plausibility of Han Feizis
advocacy of fa (, conventionally rendered in English as laws) as the crucial element
of government. Stress on fa is one of the most distinctive aspects of Han Feizis philosophy, and so any judgment of the ultimate value of Han Feizis views would have to
include treatment of that issue. As noted in the main text, my concern here is solely his
criticism of Confucianism, and the matter of whether Han Feizis own positive proposals are any better than those of his rivals is really a distinct issue that deserves separate
consideration.
5
Cf. Nicholas Gier, The Dancing Ru: A Confucian Aesthetics of Virtue, Philosophy
East and West 51.2 (2001), pp. 280-305; Edward Slingerland, Virtue Ethics, the
Analects, and the Problem of Commensurability, Journal of Religious Ethics 29.1 (2001),
pp. 97125; Bryan Van Norden, Virtue Ethics and Confucianism, in M. Bo (ed.),
175
176
eric l. hutton
the worries raised by these scholars, using such terms need not always
be gravely distorting, either, and this is particularly the case as applied
to the Han Feizi. For at the beginning of chapter fifty of the text, we find
the following statement as part of a tirade against fools and
charlatans:
Among the most prominent kinds of learning in this age are the ru The
greatest of the ru was Confucius Since the death of Confucius, there
have been the ru of Zizhang, the ru of Zisi, the ru of the Yan family, the
ru of the Meng family [i.e. Mencius], the ru of the Qidiao family, the ru
of the Zhongliang family, the ru of the Sun family, and the ru of the
Yuezheng family. Thus, after Confuciusthe ru split into eight factions
andeach faction claims that they are the true representatives of [the
way of] Confucius.7
Here, not only does Han Feizi quite clearly use the term ru to refer to
Confucius and those who follow him, but moreover he labels it a kind
of learning ( xue), and thus insofar as we are trying to understand
Han Feizis views in this essay, it is appropriate to speak of Confucians
and Confucianism, since he himself identifies his targets (or at least
one set of them) in this way.8
Accordingly, I will use Confucian to refer to Confucius and his
immediate followers, as well as those who subsequently championed
Confucius ideas about how to live, and Confucianism will refer to the
views of these people, but with the following qualification. The passage
just cited shows that Han Feizi also recognizes disagreements among
these Confucians, so that Confucianism is not a highly unified body
of thought for him (nor for me). Given this diversity within
Confucianism, a study of Han Feizis criticisms of Confucians should
Religions 17.34 (1978), pp. 303330; Michael Nylan and Mark Csikszentmihalyi,
Constructing Lineages and Inventing Traditions through Exemplary Figures in Early
China, Toung Pao 84 (2003), pp. 59-99; and Kidder Smith, Sima Tan and the Invention
of Daoism, Legalism, et cetera, The Journal of Asian Studies 62.1 (2003), pp. 12956.
7
All references to Han Feizi are according to the numbering for the version that
appears in D.C. Lau and F.C. Chen (eds.), A Concordance to the Han Feizi
(Hong Kong: The Commercial Press , 2000). This volume is part of
the Chinese University of Hong Kong Institute of Chinese Studies Ancient Chinese
Texts Concordance Series, and hereafter, this book and all other works from this series
will be referred to as HKCS for convenience. This passage is from Han Feizi, HKCS
50/150/16-19. The translation is adapted from Philip J. Ivanhoe and Bryan Van Norden
(eds.), Readings in Classical Chinese Philosophy (Indianapolis: Hackett, 2nd edn, 2005),
pp. 35152. All further translations here are my own, unless noted otherwise.
8
The same applies, mutatis mutandis, to my use of the terms Mohists and Mohism
here.
177
178
eric l. hutton
seeing further e vidence concerning particular texts are invited to consult the notes.
With these clarifications out of the way, let us begin by examining a
passage from Han Feizi that is fairly typical of some of his complaints
against Confucianism:
When a sage governs a state, he does not wait for people to be good in
deference to him. Instead, he creates a situation in which people find it
impossible to do wrong. If you wait for people to be good in deference to
you, you will find that there are no more than ten good people within the
borders of your state. But if you create a situation in which people find it
impossible to do wrong, the entire state can be brought into compliance.
In governing, one must use what works in most cases and abandon what
works in only a few cases. Therefore, the sage does not work on his virtue,
he works on his laws.12
179
180
eric l. hutton
One begins, when one serves the lost cause, to discover that, in some sense, one
ought to devote ones highest loyalty to causes that are too good to be visibly realized at any one moment of the poor wretched fleeting time world Loyalty
seeks, therefore, something essentially superhuman In its highest reaches it
always is, therefore, the service of a cause that was just now lostand lost because
the mere now is too poor a vehicle for the presentation of that ideal unity of life of
which every form of loyalty is in quest.
I should note that, as a psychological matter, one might dispute the claim that people
can really pursue ideals that they consider unachievable. For instance, one might claim
that, even when such people say they know their ideals are unachievable, deep down
they really do not believe this, and it is only because they think there still might be
some glimmer of hope for realizing their ideals that they continue to pursue them.
William James seems to adopt this sort of position (cf. William James, The Will to
Believe, in J. McDermott (ed.) The Writings of William James: A Comprehensive Edition
(Chicago: University of Chicago Press, 1977 [1896])). Determining which psychological account is correct goes beyond the scope of this paper. However, my primary point
concerns the kind of justification that might be offered for pursuing an ostensibly unrealizable ideal, and as I go on to argue, the Confucians seem to adopt a position of the
kind I describe in the main text. Even if they happened to be deceived about their own
motivations, we can still consider the logical force of their position as a reply to Han
Feizi. I thank Erin M. Cline for discussion of these latter issues.
181
18
182
eric l. hutton
183
To flesh this out a bit more, imagine two people, a healthy woman who
possesses the virtue of temperance, and a man having serious weight
problems due to binge eating, who with great difficulty exercises only
tenuous control over his gluttonous appetites. Imagine further that
both are invited to a party and arrive to discover that the host has laid
out an extensive buffet in the d
ining room. For the woman with temperance, it is perfectly appropriate to wander in and out of the dining
room as she eats and chats with others, since there is no temptation to
over-indulge, but for the man with gluttonous appetites, it is better to
stay out of the dining room altogether, lest he revert to his old ways. In
such a case, it seems that the right thing for the man to do is not the
same as what the virtuous person would do.
20
Rosalind Hursthouse, On Virtue Ethics (New York: Oxford University Press,
1999), p. 28.
21
For an essay that raises a very serious challenge to Hursthouses view and others
like it, see Robert Johnson, Virtue and Right, Ethics 113.4 (2003), pp. 81034.
22
Bernard Williams, Replies, in J. Altham and R. Harrison (eds.), World, Mind, and
Ethics: Essays on the Ethical Philosophy of Bernard Williams (Cambridge: Cambridge
University Press, 1995), p. 190. One might label this the Kids, dont try this at home!
objection.
184
eric l. hutton
Such criticism calls into question the use of the virtuous person as an
ideal, by showing how that ideal provides the wrong kind of guidance.
That point in turn brings us back to the issue from which we began,
namely how to construe Han Feizis criticisms of Confucianism. As
mentioned earlier, since the Confucians may allow that their ideal is
not fully realizable, merely accusing them of being too idealistic is not
a good criticism, and instead, if Han Feizi is going to make his attack
stick, he must claim that their ideal itself is the wrong kind of ideal to
follow. In my view, Han Feizi does exactly this, arguing along lines similar to those just described. Let me now review the evidence for such a
reading.
Han Feizi makes copious use of examples from history in his arguments, but there is one incident to which he refers repeatedly, and
which obviously made a deep impression upon him.23 This is the story
of Lord Zikuai of Yan. As he is depicted in the text, Lord Zikuai seems
a decent man, in many ways better than most other rulers of his age:
Lord Zikuai of Yan was the descendent of Duke Shi of Shao. He possessed
land several thousand li square, with several hundred thousand spearmen. He did not indulge in the pleasures to be had with young boys and
girls. He did not listen to the sound of bells and stone chimes. Within the
palace grounds, he did not build pools and towers, and outside the palace
grounds, he did not engage in trapping and hunting. Moreover, he himself took up the plough and hoe to cultivate the fields. The way Zikuai
belabored his body so as to care for the people was thus so great that even
those whom the ancients called sage kings and enlightened rulers, in tiring their bodies and caring for the world, were not greater than this.24
The idea that the king worries about the peoples welfare andat least
occasionallytakes up the implements of farming to aid in working
the fields were traits commonly attributed to the ancient sage kings
23
The incident is mentioned several times over five different chapters: HKCS
7/10/11-13, 35/109/16 35/110/13, 38/123/9-10, 39/127/17-24, and 44/134/27-30. The
most extended discussion is in chapter 35. This story may have been especially memorable for Han Feizi, since according to one reference (ch. 38), the king of Yan turned
down Han Feizis own teacher, Xunzi, in favor of Zizhi, a fatal mistake.
24
HKCS 44/134/27-30. In conversation, Masayuki Sato has pointed out to me that
the description of the king eschewing music in this passage implies that Zikuai was
trying to be a good king on the model of the Mohists, rather than the Confucians. This
may be so, but it does not particularly affect my argument here, because the criticism
of trying to follow the ancient sage kings, and especially imitating their actions of
handing rule over to a sagely minister, will apply just the same to the Confucians as to
the Mohists.
185
(especially Yu)25 at the time. The fact that Zikuai does these things
seems to show that he was consciously trying to imitate them, and
apparently with some success, since his accomplishments are said to
rival those of the ancient sages.
Despite these qualities, Zikuai is not a sage in Han Feizis eyes, but
rather provides an example to be avoided, for he had a fatal flaw that
cost him both his state and his life. The passage just cited continues:
Nevertheless, Zikuai died and his state perished after being taken away
from him by Zizhi, and he was ridiculed by all under Heaven. What is the
reason for this? It is because he did not understand the way to employ
ministers.26
In typical fashion, Han Feizi does not explain here what passed between
Zikuai and Zizhi, but rather assumes that his readers will know the
story, because itwas a quite infamous affair in recent history.27 The particular events he references are as follows. Zizhi, who was one of
Zikuais ministers, managed to dupe Zikuai into handing over power to
him. Zizhi, however, turned out not to be a competent ruler. The state
grew more chaotic, and after a few years, the state of Qi invaded (around
314 BCE). During the conquest Zikuai was killed, and Zizhi met with
an especially grisly fate, being made into meat paste.28 Here, Zikuais
fault is said merely to be that he did not understand how to choose and
employ ministers, and insofar as Zizhi was neither as honest or capable
as Zikuai thought, it is clear that this is a fault in Zikuai.
25
Yu supposedly engaged in physical labor personally to help clear away floods and
prepare land for cultivation. Versions of the tale can be found at: Han Feizi, HKCS
49/145/29-30; Zhuangzi, HYIS 91/32/25-27 (William Hung (ed.), A Concordance to
Chuang Tzu (Beijing: Harvard-Yenching Institute, 1947)); Xunzi HKCS
25/121/12-13, HYIS 93/25/23-24; and Analects 14.5.
26
HKCS 44/134/30-31.
27
The affair was infamous enough to be mentioned in bronze inscriptions of the
time. Cf. the discussion of the Zhongshan Wang Cuo vessel by Gilbert Mattos, Eastern
Zhou Bronze Inscriptions, in E. Shaughnessy (ed.), New Sources of Early Chinese
History: An Introduction to the Reading of Inscriptions and Manuscripts (Berkeley: The
Society for the Study of Early China and The Institute of East Asian Studies, University
of California, 1997), pp. 104111, esp. p. 109. The event is also mentioned in the
Mencius explicitly at 2B8, and implicitly at 1B10, 1B11, and 2B9.
28
The story of Zikuai and Zizhi is recorded in various places, but a fairly standard
account is contained in the Shi Ji , in the biography of Su Qin (cf. Sima Qian,
Shiji (Beijing: Zhonghua Shuju, 1998), p. 2268). Zizhis gruesome end is reported
by the Zhushu Jinian book five , Annals of King Yin , year two (for
a translation, cf. James Legge, The Shoo King (Taiwan: SMC Publishing, Inc., [1893]
1991), prolegomena p. 175).
186
eric l. hutton
Here Pan Shou refers to a common legend that, because the sage kings
were truly concerned for the well-being of their people and the state
rather than for keeping power to themselves, as they grew older they
sought out worthy ministers with whom to entrust the government.
These worthy men, though, were likewise not interested in power, so
they either refused or accepted only with much reluctance.30 Either
way, the state was well governed, and both sides were acclaimed for the
way they displayed great virtue. In the case of Yan, Zizhi initially made
a show of refusing the throne, butcontrary to poor king Zikuais
expectationshe eventually accepted it, and things only went downhill
from there.
In analyzing the kings downfall, we should see that part of the kings
mistake was his faith in Pan Shou, and in a sense it is the same mistake
as the one Han Feizi notes in the previous passage cited, namely that
Zikuai did not know how to choose and employ appropriately people
to serve him. For according to the stories, Pan Shou was secretly a
henchman of Zizhi, and the two conspired to bring about the kings
resignation.31 More generally, one broad theme that the various stories
of Zikuais downfall share in common and which Han Feizi emphasizes
is that a ruler must exercise great care when choosing whom to trust,
HKCS 35/109/28-31.
For a telling of the Xu You tale, cf. Zhuangzi HYIS 2/1/22-26. For a case of reluctant succession, cf. Mencius in 5A5.
31
HKCS 35/110/1-7.
29
30
187
188
eric l. hutton
push-government [i.e. government methods as crude as pushing a cart
by hand]. When one dwells in a time of many problems, to use implements suited for few problems is not the kind of preparation made by a
wise person. To face an age of great contention, but follow in the tracks of
those who yielded and gave way, is not the way a sage brings about order.
And so, the wise person does not ride in a push-cart, and a sage does not
practice push-government.34
First, although Han Feizi does not mention the case of Zikuai yielding
the throne to Zizhi here, given that he shows strong interest in the story
in other places, it seems most likely that he has that case in mind when
attacking those who would follow ancient ways by yielding and giving
way. Second, note that Han Feizi does not criticize the action of giving
away the throne per se; here, at least, he seems to accept that it was a
sagely thing to do for the ancients.35 Rather, his point is that what was
appropriate for the early kings is not appropriate for modern rulers,
and if we think of this in terms of the lesson learned from the case of
Zikuai and Zizhi, then the thrust of his argument seems to be that
modern rulers should not imitate past sages, lest disaster follow.36
Now the idea that one should follow the ways of the ancients was a
central part of Confucianism, and the Confucians were fond of the stories about sagekings ceding their rule to a worthy minister as a manifestation of their virtue.37 Hence, in making such remarks Han Feizi is
HKCS 47/140/20-24. Commentators disagree about how to construe parts of this
passage. Here I follow the reading of Qiyou Chen , Han Feizi Jishi
(Taipei: Huazheng Shuju, 1987).
35
Han Feizis exact view on this is difficult to pin down. He nowhere denies that the
sage kings of old ceded the throne to others, and in one place he seems to describe it as
the natural thing to do, given the circumstances (e.g. HKCS 49/145/28 49/146/6).
However, elsewhere he seems to say that it set a bad precedent which has led to chaos
ever since (HKCS 51/153/11-21), but even there he does not seem to deny that it was
an effective decision for the ruler at the time. It is worth noting that Xunzi displays a
certain hostile attitude toward the legends about the sages abdicating, so perhaps Han
Feizi picked up part of Xunzis view here (cf. Xunzi HYIS 67/18/53 68/18/72; HKCS
18/86/6 18/87/5).
36
It is worth noting that in the Zhuangzi, at HYIS 43/17/34-35, there is a passage
that is very similar to the arguments we find in Han Feizi. However, given the many
problems in dating the later chapters of the Zhuangzi, it is difficult to tell whether Han
Feizi is influenced by the Zhuangzi author or vice versa.
37
Cf. e.g. Analects 20.1 and Mencius 5A1 (esp. ), 5A4 (esp.
), 5A5, 5A6. Although Mencius denies that the king can directly give the
empire to anyone, he accepts the basic idea that the king chose a worthy person to rule
in his place. I discuss Mencius further in the main text, p. 441. For a nice example of
how Mencius encourages imitating the sages, see 6B2. As mentioned above (cf. note
35), Xunzi seems suspicious of the legends about sages abdicating, but the basic point
of the objection, as I note in the main text, has to do with the idea of imitating the
34
189
190
eric l. hutton
In contrasting Yao and Shun with Jie and Zhou, Han Feizi seems clearly
to be thinking in terms of the traditional view accepted by the
Confucians, on which Yao and Shun ruled with benevolence and righteousness, and Jie and Zhou ruled with cruelty and debauchery.
Therefore, when he then considers what would happen when another
Yao or Shun arrives (as opposed to just a nameless sage), I take him to
be referring to a sage who is like Yao and Shun precisely in the respect
of ruling with benevolence and righteousness. Since he says that when
such a Yao or Shun comes, there will indeed be order, Han Feizi thus
seems to acknowledge that someone might be able to rule effectively
using the virtue-based techniques of the past sages, though it would
require sagely character and intelligence to make this work (for another
passage of thissort, see p.442 below). Yet, as this same passage indicates, Han Feizi regards such people as exceedingly rare, and for that
41
HKCS 40/129/7-9. Translation from Ivanhoe and Van Norden, Readings in
Classical Chinese Philosophy, pp. 33031.
191
reason he proposes a different method that can work for the far more
common mediocre rulers.42
While ultimately Han Feizi would probably prefer not to rest the fate
of the government on any particular strengths of the ruler, but instead
depend on a system of laws and administrative techniques that will
work for all but the most kind or most cruel rulers, insofar as Han
Feizis claim is thus that the mediocre rulers incapable of governing
successfully by virtue should avoid trying to follow the past sages
approach, his thought does seem much closer to Williamss argument
that agents of lesser ability and virtue should not imitate those of
greater ability and virtue, because disaster will likely result. In this light,
the story of Zikuai and Zizhi serves as an especially powerful case for
Han Feizi against the Confucians, because on a very plausible reading
of the passage cited earlier, in yielding the throne Zikuai was motivated
more by concerns for enhancing his reputation and keeping his power
than he was by any genuine desire to find the most worthy person to
administer the state. To that extent, he was trying to make a show of
virtue that he did not really possess, which helps explain his failure
but such morally mediocre rulers as Zikuai are precisely the norm,
according to Han Feizis view, and hence one should avoid those ideals
(e.g. that of the Confucians) that will normally lead to ruin. While Han
Feizi is not so explicit as all this, one can perhaps see at least a hint of
this Williams-style concern about the abilities of agents in his treatment of the story of Zikuai and Zizhi in this vein, given how he sometimes stresses Zikuais lack of understanding.43
42
In conversation, Li Chenyang has suggested to me that in this regard Han Feizi is
suggesting yet another kind of criticism of the Confucians, namely that they have no
acceptable fallback plan for when it turns out that their ideals cannot be realized. This
is an interesting and plausible idea that deserves more consideration. However, I will
not be discussing it further here, in orderto be able to give sufficient attention to the
line of argument I am already tracing out in the maintext.
43
There is another passage that seems to support understanding Han Feizis worry
in this way. At HKCS 38/121/24 38/122/5, Han Feizi relates the tale of Duke Wen of
Jin and the eunuch Pi. When Duke Wen was early on exiled from his state, Pi was sent
to try to assassinate him twice, but failed both times. When Duke Wen eventually manages to return to Jin and become ruler, the eunuch Pi requests an audience with him.
Through an intermediary, the Duke asks why the eunuch acted with such alacrity in
trying to kill him previously, implying that the eunuch must harbor some hatred
against him. Pi answers that he was merely following orders and that he has no personal hatred toward the Duke. He also cites the story of Guan Zhong, who nearly shot
and killed Duke Huan of Qi, but was later employed by Duke Huan, who realized that
Guan Zhong would make an effective minister. Commenting on this story, Han Feizi
192
eric l. hutton
193
Mencius answered, It may. It was not allowable for Zikuai to give Yan
to another person, nor was it allowable for Zizhi to receive Yan from
Zikuai. Suppose there were a well-bred man here whom you liked, and to
whom you privately gaveyour official rank and salary without telling the
king, and suppose this well-bred man likewise received these things from
you without a mandate from the kingwould that be permissible? How
is that any different from this [i.e. the affairs in Yan]?
Mencius clearly disapproves of Zikuai and Zizhi, and even thinks that
they deserve punishment, though the remainder of the passage (not
translated here) shows that he thinks Qi was wrong to have undertaken
the task of punishing Yan. His reasons for disapproving of Zikuai and
Zizhi, which are merely hinted at above, are made more explicit at 5A5
and 5A6. According to those passages, the ancient sage kings did not
directly give the empire to their successors. Instead, they recommended
those men to Heaven, to whom the empire really belongs, and Heaven
accepted them as successors. Accordingly, Mencius view would seem
to be that the state of Yan, as a landholding originally granted by the
Zhou dynasty king, belongs to the Zhou king (and ultimately to
Heaven), not to Zikuai individually, so Zikuai had no authority to give
it away on his own, but at most could only recommend to the Zhou
king that Zizhi succeed him.45 Confucians might then follow out
Mencius thought by saying that Zikuai does not represent a case where
someone succeeded in imitating the sages but met with disaster.
Such a response to Han Feizis criticism, however, would not serve
the Confucians well. For one thing, Han Feizi has other examples that
seem less subject to the same complaint. Consider the following:
In ancient times, King Wen dwelled between Feng and Hao, with a territory of only a hundred li square. He practiced benevolence and righteousness and was friendly to the Western Rong [a barbarian tribe], and
subsequently came to be king over all under Heaven. King Yan of Xu
lived to the east of Han, with a territory of five hundred li square. He
practiced benevolence and righteousness, and thirty-six states cut off
pieces of their territory [i.e. to present as a tribute and symbol of submission] and came to pay him court. [The king of Chu] feared being harmed
by him, so he raised troops and attacked Xu and subsequently destroyed
it. Thus, King Wen practiced benevolence and righteousness and came to
be king over all under Heaven, but King Yan practiced benevolence and
For more discussion of Mencius views on this matter, cf. David Nivison, Mengzi
as a Philosopher of History, in A. Chan (ed.), Mencius: Contexts and Interpretations
(Honolulu: University of Hawaii Press, 2002), especially pp. 29698.
45
194
eric l. hutton
righteousness and lost his state. This shows that benevolence and righteousness were useful in ancient times, but are not useful in present
times.46
Here, the problem does not seem to be that King Yan did not succeed
in imitating the sage king Wen. On the contrary, the problem seems
rather to be that he was too successful. For as the Chinese scholar Lin
Weiyi has astutely noted, in this passage Han Feizi seems to admit that
virtue does have suasive power even in his own contentious age, since
he says that thirty-six states were won over by King Yans benevolence
and righteousness.47 King Yan was thus succeeding quite well in following the path of King Wen, but this very success made him a threat to
Chu and so brought about his destruction. Now we are not told the
details of King Yans rule, so it might perhaps still be open to the
Confucians to argue that King Yan was not really properly imitating
King Wen. Yet, for the Confucians to dispute in this fashion every such
example that Han Feizi raises does not seem a promising strategy,
because it will begin to look like special pleading if it turns out that
any time someone ostensibly acting as the Confucians direct meets
with disaster, it is because they were not really correctly imitating the
sages. Furthermore, such a tactic would threaten to undermine the
Confucians own views, since in some cases they seem to admit that
truly virtuous people may fall victim to disaster through their good
behavior.48
Apart from these worries, though, there is another, more fundamental problem with trying to escape Han Feizis criticism by arguing that
HKCS 49/146/8-11.
Cf. Weiyi Lin , Fa Ru Jian Rong: Han Feizi de Lishi Kaocha :
(Taipei: Wenjin Chubanshe , 2004), pp. 3133. It is an
interesting question to what extent this admission may undermine Han Feizis other
remarks about the importance of virtue for good government. I cannot pursue that line
of inquiry here, however. At theveryleast, this seems to show that when Han Feizi says
in the passage that benevolence and r ighteousness are not useful ( yong) for governing, what he means is that they are not sufficient for governing well, not that they are
completely ineffective.
48
For example, Xunzi seems to think of Bi Gan as a genuinely virtuous person, and
Bi Gan tried to save his lord, the tyrant Zhou, from impending doom by remonstrating
with him, but Zhou responded by having his heart cut out (cf. Xunzi, HYIS 50/13/1019, 66/18/24-29; HKCS 13/63/28 13/64/9, 18/84/14-18). In this case, to say that
whenever doom befalls a seemingly good person, it is because that person is not really
good, would threaten to undermine the belief in Bi Gans goodness, and likewise for
other figures whom the Confucians esteem. For similar cases in the Analects, cf. 18.1,
which mentions Bi Gan, and also 15.9, which notes that death may be the price of virtue. This last idea is also expressed in the Mencius in 3B1, for example.
46
47
195
Zikuai did not successfully imitate the sages. In particular, such a tactic
seems to miss the point of Han Feizis criticism, for (subject to certain
qualifications already noted above) Han Feizi is aiming to discourage
people from even trying to imitate the past sages, not just from successfully imitating the past sages. Indeed, since Han Feizi thinks we cannot
be sure about the distant past, he seems committed to a high degree of
pessimism about peoples ability to follow in detail the actual behavior
of sages like Yao and Shun.49 His repeated criticisms of imitating past
sages would then be pointless, if he thought almost no one could truly
imitate them anyway. Accordingly, the only charitable way to understand his criticisms of imitating the past is to construe them more
broadly, namely as applying to those attempting to imitate the past,
regardless of whether or not they actually do it correctly. On this understanding, the Confucian strategy considered in the previous few paragraphs will not make a difference even if successful, because Zikuai was
clearly trying to follow the model of the sages, even if he did not succeed perfectly, and to the extent that this attempt led him into disaster,
that provides reason enough against trying to imitate the past sage
kings.50 Thus, if the Confucians are going to defend themselves against
Han Feizis attack, they must adopt a different strategy.
Now there is an alternative available to the Confucians. Namely, they
can deflect the attack launched by Han Feizi against their ideals by
claiming that such criticism misconstrues the role that these ideals are
supposed to play in guiding peoples actions. Since I have been suggesting a parallel between Bernard Williams and Han Feizi, let me go back
to Williams for a moment in trying to explain this idea. His argument
gets its bite by taking the idea of what the virtuous person would do in
the circumstances to refer to fairly specific actions and then pointing
out the dangers of imitating those actionsin terms of my earlier
example, what the temperate person does at the party is to go in and
out of the dining room partaking of the buffet, but that is not what the
person fighting gluttonous appetites should do. Similarly, Han Feizi
Cf. note 16 above.
It is worth noting that although Williamss stand on this issue is unclear, a similar
idea seems to be suggested in his original critique, in particular in his analogy with
weight-lifting (cf. p. 433 above). Namely, I may hurt myself by being unable to stay
steady once I have the weights over my head, but if the weights are too heavy for me,
I may also tear muscles and injure my back just trying to lift them, while still not budging them an inch. Thus, if I am not up to the task, I have good reason not even to
attempt it.
49
50
196
eric l. hutton
takes it that the Confucians want rulers to govern as the ancient sages
would, but then points out how the specific actions taken by the past
sages, such as ceding the throne, result in ruin when copied by others.
In both cases, it is presumed that the point of upholding the virtuous
person or sage as an ideal is to encourage people to imitate the specific
things they did.
However, one might reject this understanding of the role of the ideal
figure. If so, there are at least three alternatives to consider. Before discussing those options, though, let me note what this approach means in
terms of the story of Zikuai. Specifically, it means that the Confucians
would be saying that Zikuais actions do not in any way represent what
they are advocating. For Zikuai was trying to imitate the actions of the
sages, but that is not what they are advising people to do. Rather, on the
alternative view, what the Confucians espouse is not the simple imitation of actionssomething which perhaps even they could admit51
would indeed be subject to the kind of critique that Han Feizi articulatesbut rather the imitation of something else, and since Zikuai was
not even trying to imitate (much less succeeding in imitating) this, his
downfall and other similar stories about the perils of imitating past
sages actions do nothing to impugn the Confucian ideal.
If one does not take what is most admirable and worthy of imitation
in the virtuous person or sage to be the particular things he or she does,
then what should one be trying to imitate instead? The first possibility
to consider is that it is the ideal figures good judgmentthe ability to
pick the course of action that is truly best. On this view, the advice to
imitate the sages amounts to the idea that one ought to do what they
did, but internally and mentally, rather than externally and physically,
or in other words, to think the same way they thought. For evidence
that Confucians might adopt this approach, we may look to the Xunzi,
which in counseling the novice states, When the teacher explains thusand-so, and you also explain thus-and-so, then this means your understanding is just like your teachers understanding.52 Since here the text
51
The discussion at Mencius 4A17 of whether or not to save ones drowning sisterin-law in violation of ritual propriety might serve as an example of how the Confucians
recognized this problem. Similarly, Analects 9.3 suggests an acknowledgment that
strict adherence to the dictates of ritual may not always be best, and the Xunzi repeatedly advocates an ability to adapt in response to changing circumstances, e.g. HYIS
7/3/15, 17/6/50, 23/8/87, 28/9/48, 45/12/25, 52/14/17, HKCS 3/10/2-4, 6/25/2, 8/32/11,
9/37/19, 12/58/3, 14/67/8.
52
Xunzi, HYIS 5/2/38, HKCS 2/8/2. Compare Analects 13.5, which makes it clear
that in learning the Odes one is supposed to go beyond merely being able to recite them
197
198
eric l. hutton
199
there was nothing to be done about it, they still insist that in order to
model oneself after Confucius and take him as ones teacher, one must
compose and transmit writings to pass on to future generations. But how
can one say that they have no other choice? Why do they so disregard the
people of their own time, while being so dedicated to future generations?
Thus, those who study Confucius should study what it was that Confucius
studied [i.e. how best to respond tohisown time], and should not study
what Confucius had no other choice but todo.56
Hence, for Zhang, the way to follow the sages is not to try to do what
they did, but rather to learn to appreciate how they made the right
decision for their situation, then look to ones own circumstances and
take appropriate action for ones own time.
Although Zhang lived many centuries after Han Feizi, the roots of
Zhangs view go way back in the Confucian tradition. In fact, one can
see hints of it in Han Feizis onetime teacher, Xunzi.57 Indeed, one might
even interpret the remarks from Xunzi about imitating the teachers
understanding cited earlier along such lines. Chapter twenty-one of
the Xunzi, which discusses the causes and cures for fixation or obscuration ( bi) that prevents people from truly grasping the Way, contains the following remarks:
Thus, among the cases of fixation, onecan be fixated on origins, or one
can be fixated on ends. One can be fixated on what is far away, or one can
be fixated on what is nearby One can be fixated on the ancient past, or
one can be fixated on the present. In whatever respect the myriad things
are different, they can become objects of fixation to the exclusion of each
other. This is the common problem in the ways of the heart The sage
knows the problems in the ways of the heart So, he is neitherfor the
origins, nor for the end results, is neither for what is near, nor for what is
far away,is neither for the ancient past, nor for the present.58
Here Xunzi seems to say explicitly that the sage is neither tied to past
ways, nor does he simply follow present customs. Instead, the sage
picks what is best based on the circumstances, and in some cases this
may mean following ancient ways, and in other cases departing from
them. Now this is not to say that Xunzi holds the exact same position
56
Cf. Xuecheng Zhang , Yu Chen Jianting Lun Xue (Letter
to Chen Jianting Discussing Learning), in X. Zhang, Zhang Xuecheng Yishu
(Beijing: Wenwu Chuban She , 1985), p. 86 (, ).
57
David S. Nivison, The Life and Thought of Chang Hseh-cheng (17381801)
(Stanford: Stanford University Press, 1966), p. 166, notes a similarity between Xunzi
and Zhang that is much like what I suggest here.
58
Xunzi, HYIS 78/21/6 79/21/29, HKCS 21/102/12 21/103/18.
200
eric l. hutton
201
61
For example, consider Confucius statement in Analects 2.4 that it took him seventy years to get to the point where he could follow his desires without going astray,
Mencius warnings at Mencius 2A2 neither to neglect nor rush the process of moral
cultivation, and Xunzis remarks at HYIS 2/1/17-18, 89/23/68-69; HKCS 1/2/9-10,
23/116/13-15 about the need for slow accumulation ( ji) of efforts in order to
achieve sagehood.
202
eric l. hutton
To give an example of this worry, suppose I want to imitate a strongman. On the view under consideration, I want to imitate him in being
strong, rather than the particular things he does, such as lifting 300 lbs,
which would be extremely dangerous for me even to try, if I am a weakling. (Here I rely on Williamss formulation of the problem rather than
Han Feizis, because Williamss is more straightforward, but one could
easily imagine versions of the problem, mutatis mutandis, in terms of
Han Feizis worry.) Nevertheless, it may turn out that the only way for
me to become strong is by lifting heavy things, which is still quite
dangerous.
Now there is an objection that could be raised to this last point, in
order to save the Confucian position. Namely, just as in the way that, as
a matter of fact, many people do manage to become strong by lifting
heavy things without seriously injuring themselves, so by analogy it
would seem possible to imitate the virtuous persons virtue without a
high probability of falling into disaster. Note, however, that at this point
the relation between action and the ideal has now become rather tenuous; the actions required in such imitation may bear only a very vague
resemblance to the actions of the ideal figure (e.g. the strong person
pulls airplanes along the ground with his bare hands, but the beginner
lifts 5-lb dumbbells). If so, then it will be difficult to use the actions of
past sages to justify particular regimens of practice for current imitators, and that would again undercut the Confucians ability to justify
their preferred form of government against Han Feizis proposalsthe
problem of vacuity has thus returned. In terms of the analogy, if my
goal is to be strong, and I could accomplish this in more than one way,
say by lifting heavy things or by taking drugs, etc., then the goal of
being strong will not serve to justify one over the other. Similarly, Han
Feizi sometimes seems to suggest that if one really wants to be a benevolent ruler, as the Confucians espouse, one should follow his proposals
rather than theirs,62 and if the Confucians admit that there might be
many ways to be benevolent in order to avoid Han Feizis initial criticism, they will again have trouble ruling out this possibility.
The argument of the preceding three paragraphs still leaves open
one further move, the last I will consider. The previous argument works
by assuming that the beginner must imitate the sages actions in order
to become virtuous, but then notes how, in order to avoid Williamss
E.g. HKCS 42/130/24 42/131/1, 53/156/12-15, 54/156/19-27.
62
203
204
eric l. hutton
205
67
For example, cf. John McDowell, Virtue and Reason, in idem, Mind, Value, and
Reality (Cambridge, MA: Harvard University Press, 1998), pp. 5073, and Jonathan
Dancy, Moral Reasons (Cambridge, MA: Blackwell, 1993).
68
To put this point in jargon that may be more familiar to contemporary philosophers, there seems to be a tension between using the good person criterion as a guide
for action and adopting a particularist stance toward moral judgments.
206
eric l. hutton
actions should be imitated, and which not, is a difficult task. In attempting it, the advocates of virtue ethics would perhaps also do well to distinguish something like a practice model from a goal model. Since the
Confucians (or at any rate, some of them) hope to occupy this kind of
middle ground, and since their notion of ritual may help in building an
appropriate practice model, this suggests that those interested in virtue ethics could perhaps learn something from how the Confucians
tried to do this, though that is, as I have said, a task I must leave for
another time.
208
maria w. merritt
209
210
maria w. merritt
5
Kamtekar, Situationism and Virtue Ethics on the Content of Our Character,
pp. 48285.
211
individual virtue in rare cases.6 Eric Hutton, in sympathy with traditional virtue ethics, has recently interpreted early Confucian thought
as compatible with situationist social psychology in just this way. While
the Confucians recognize that ordinary people are susceptible to the
vagaries of situational influence, at the same time they believe in rare
gentlemen and sages whose good behavior is utterly stable and
consistent.7
Yet the social function of moral norms is to regulate actual conduct
among ordinary men and women in human populations. And part of
the point of philosophical moral theory is to reflect critically on what
the content of moral norms should be, given their social function. If
traditional philosophical ideals of virtue are so lofty that even their
advocates expect them to be realized only rarely at best, how can they
be relevant to practical morality?
One possibility is that individuals might strive toward such ideals as
aspirational goals.8 Even when the goal is unattainable you can hope,
through the effort of trying to reach it, to become better than you might
otherwise have been. A second possibility is to adopt a division-oflabor model and hold out hope for a moral meritocracy to keep the rest
of us in line. Hutton suggests that in early Confucian thought, reflection on ideals of virtue could be crucial not despite the rarity of virtue,
but because of it. Given a traditional hierarchical society, it is ideally
the virtuous gentlemen and sages who will establish the social structures that ordinary (ex hypothesi, mostly non-virtuous) people inhabit:
if there are people who do have robust character traits and are resistant
to situational variation, they can design and reliably maintain the broad
range of institutions and situations that facilitate good behavior for
everyone else.9 Transposing the idea to modern liberal democracies,
we might hope that the individuals who lead particular communities,
212
maria w. merritt
organizations and institutions will turn out to be the ones who prove
capable of virtue.10
These two possibilities are not mutually exclusive. Individuals could
be encouraged to aspire to virtue whether or not they hold positions of
social or political leadership. Those who do hold such positions could
be given to understand that their institutional roles carry an extra burden of responsibility to strive for virtue, commensurate with the importance of their influence on situational settings that guide the behavior
of men and women subject to their authority. But even if a few virtuous
super-individuals exist, there is no particular reason to be confident
that they will regularly occupy positions of power (not least because in
order to secure such positions they would have to compete against
individuals and organizations who value power above all). In general,
if virtue is so rare as traditional virtue ethicists are willing to maintain,
and if our behavior is influenced by morally arbitrary factors so extensively as they are willing to admit, we have good reason to doubt that
we should entrust the effective social functioning of moral norms to
the personal resolve and conscientious striving of individuals (ourselves or others, highly placed or not).
My concern in this article is with the psychology of interpersonal
processes as they affect morally important personal dispositions over
time. This issue is distinct from the problem highlighted by situationism, which is that morally irrelevant factors affect morally important
behavior at the time of action. Focusing on the influential Aristotelean
ideal of virtuous character as firm and unchangeable, I argue that the
evidence about interpersonal processes and their effect on personal
dispositions over time provides a distinct and additional reason, over
and above situationism, to predict that ordinary men and women will
run into trouble in the attempt to realize such an ideal.11 The safest bet
is to assume that however well-intentioned and conscientious we may
be, empirical generalizations about interpersonal processes apply to
10
Or third, as Hutton (Character, Situationism, and Early Confucian Thought,
p. 57, n. 41) suggests in passing, insofar as a society is democratic, and it is the ordinary people who direct the government and thereby their own lives, it may seem
equally imperative to make their behavior as little situationally-dependent as possible.
But to take this third option is to back off from the assumption that virtue is rare, thus
incurring greater exposure of moral norms to empirical generalizations about what
most people are like.
11
In principle, my point about interpersonal processes should stand on its own,
whatever ones view of the situationist problem.
213
214
maria w. merritt
NE 1140b2930.
16
215
216
maria w. merritt
noble and despising the base, we will energetically seek to learn more
about how to make practical choices in accordance with those attitudes, progressing in the gradual acquisition of practical wisdom.
Abilities that we develop on the way to moral virtue, such as impulse
control and other forms of self-control necessary for keeping emotional and appetitive activity in line with our values, are in due course
supposed to reach beyond mere self-control to culminate in the highly
integrated self-mastery of virtuous character.
Meanwhile the process of acquiring practical wisdom teaches us
more about the nature of the nobler objects of choice, thereby adding
depth, intensity, and discerning precision to our love for the noble.
What constitutes practical wisdom is our commitment to our values in
reflective understanding, manifested in our intelligent recognition of
when and how they are in play on occasions of deliberation and choice,
including circumstances of adversity. The continued accumulation of
experience gradually reveals more and more of the ways in which the
nobler objects of choice really are superior, keeping us in the attitude of
caring about getting our choices right.
Interpersonal Aspect of Virtuous Character
Aristotle allows that even mature virtuous character needs at least
some s upport from social relationships. We can begin by remembering
that in his view the whole value of virtue lies in its active exercise, as
distinct from its mere possession.19 Now add the consideration that a
great many of the Aristotelean virtuescourage in battle, justice, liberality, magnificence, pride, due ambition, good temper, friendliness,
truthfulness, ready witrequire for their active exercise a social world
inhabited by a community of peers. This is so for reasons over and
above Aristotles argument that since human beings are social animals,
we can live the best human life only in society. More interestingly for
our purposes, it is so because Aristotles conception of virtue, founded
in love for the noble and contempt for the base, is highly sensitive to
standards of honor.
For Aristotle as for Homer, whom he quotes approvingly on this
point in his discussion of courage, honor is an essentially social phenomenon.20 In order for a man to be fully assured that in his conduct
NE 1098b33-1099a5; see also NE 1114a7-10.
NE 1116a20-26.
19
20
217
he meaningfully satisfies a standard of honor, he needs to win the recognition and respect of other men who are honorable themselves. And
for their part, each of them needs the same.21 The standard of honor
must be located in the judgment of fellow men of practical wisdom, as
opposed to the judgment of anyone at all, however wise or foolish. The
value of honor is in its guidance and confirmation of the honored mans
understanding of himself as virtuous; only honor bestowed by men of
practical wisdom can deliver this result.22
Similarly, in Aristotles account of friendship, friends are supposed
to love each other on the basis of virtue, each loving the other for his
good character. They choose to live together and they coordinate their
activities to pursue various specific ends that they share. The overall
end that unites them is the end of living a good life. One reason they
can help each other to pursue this larger end is that each admires and
tries to emulate the other. Given that Iregard my friend as someone
generally of good character, a person practiced in living well and better
at it in some respects than I am, I can learn from my friend how to
improve in those respects. And in the best kind of friendship there will
be reciprocity: my friend can learn from me in the same way, mutatis
mutandis.23
So, in Aristotles view, the virtuous man needs a social world for the
expression and exercise of the virtues. He needs not just any social
world, but one that prominently features social bonds with others who
share his evaluative sensibilities and priorities.
III. Empirical Psychology: Interpersonal Processes and
the Discontinuity of Character
It is a commonplace of Aristotelean moral psychology that in early life
the agent will, ideally, begin to adopt the values characteristic of virtue
by way of training and education in a supportive social community.
With increasing maturity and reflective understanding comes increasing independence in the agents commitment to the right values. The
goal of living up to the right values, initially a goal assigned by others,
NE 1095b26-29.
See also NE 1159a21-24. For a discussion of the general point see Bernard
Williams, Shame and Necessity (Berkeley: University of California Press, 1993).
23
NE 1172a11-14.
21
22
218
maria w. merritt
is transformed into a goal we set for ourselves. But even self-set goals
may continue to be influenced by the continuing normative expectations of selected others: those who share our values, and with whom we
choose to maintain interpersonal relationships. As noted above, this is
what we find in Aristotles ideal of the virtuous mans social world.
Using the resources of present-day scientific psychology, I shall suggest how interpersonal processes operate through social bonds with
like-minded others, with the effect of helping individuals to sustain virtuous normative commitments over time (section A); I shall then give
an empirically informed account of how certain aspects of the same
interpersonal processes also threaten to break the self-reinforcing
cycle between moral virtue and practical wisdom, prompting occasions of discontinuity that may over time become lasting disruptions
(section B). Rather than simply serving to strengthen character, as the
Aristotelean view would have it, social bonds with others who share
our evaluative sensibilities and prioritieseven assuming we have got
those right in principlemay also tend to undermine character.
A. Self-Evaluation and Self-Presentation
The agents values are the psychological ground of virtue. For
Aristotelean virtue, we may more accurately speak of positive and negative values, or values and anti-values, since virtue requires getting
it right not only about the noble as an object of choice, but also about
the base as an object of avoidance.24 This motivational structure brings
with it a self-evaluative concern to be, or to become, a certain kind
of person.
In the first place, the very activity or stance of valuing draws some
of ones attention to self-evaluation. As Agnieszka Jaworska has
pointedout, A person values herself in terms of how well she lives up
to her values.25 On top of that, the standard Aristotelean tradition
specifically encourages the use of broadly characterological attributions in evaluating ones own and others actions. Indeed, for Aristotle
the normative criterion of virtuous choice, the orthos logos, is set by the
idea of choosing as the man of practical wisdom would choose.26
NE 1104b30-1105a1.
Agnieszka Jaworska, Respecting the Margins of Agency: Alzheimers Patients
and the Capacity to Value, Philosophy and Public Affairs 28 (1999), pp. 105138,
at p. 115.
26
NE 1106b36-1107a2.
24
25
219
220
maria w. merritt
32
In the routines of everyday activity and interaction, there is simply too much
information about oneself to do anything but provide an edited, packaged version that
is relevant to the goals at hand. B.R. Schlenker and M.F. Weigold, Self-consciousness
and Self-presentation: Being Autonomous versus Appearing Autonomous, Journal of
Personality and Social Psychology 59 (1990), pp. 82028. See also B.R. Schlenker and
M.F. Weigold, Interpersonal Processes Involving Impression Regulation and
Management, Annual Review of Psychology 43 (1992), pp. 13368; M.R. Leary, Selfpresentation: Impression Management and Interpersonal Behavior (Boulder, CO:
Westview Press, 1996).
33
Leary, The Self We Know and the Self We Show, pp. 46263.
34
To be sure, even when there is some incentive to satisfy such expectations, there
might also be other incentives to opt out. The advisability of opting out will depend in
part on the cost of severing the relationships in play.
221
222
maria w. merritt
them the treatment to which they are entitled by law. She does so in the
right spirit (caring that justice be served), and for the right reason (that
justice requires equal treatment). In short, Angelas v alues, choices, and
reasons for action are unimpeachable with respect to justice.
But over time, the groups caseload gradually increases. In their
efforts to keep up, they begin to cut corners, but in each case they manage to rationalize their actions as ethically unproblematic. Without
quite realizing it, Angela too makes more and more choices that are in
fact inconsistent with what justice requires, going in for her share of
well-meaning but unwittingly self-serving rationalizations. Angela
continues to understand herself as dedicated to justice, but in her
actual conduct she is no longer just. Indigent defendants suffer the
consequences.
If there is a single way to characterize the downward spiral of mistakes that Angela and her colleagues have fallen into, it is misapprehension of the objects of choice and avoidance. In terms of Aristotles
account of acting for the sake of the noble with respect to the objects
of choice and the objects of avoidance, they have become prone to
choose actions they ought to avoidactions that are advantageous
but wrong (base)under misapprehensions of them as advantageous
and right (noble).
A sadly compelling real-life example is the scandal that recently surrounded Patricia Dunn, former Chairman of the Board at the HewlettPackard Company (HP). Dunn is an expert in corporate governance.
When she joined the HP board, she brought with her an impeccable
record of ethical responsibility in powerful positions such as sole global
CEO of Wells Fargo Investment Advisors, where she had been the
principal fiduciary for more than a trillion dollars in assets.37 As board
chairman at HP, in a well-intentioned campaign to contain high-level
leaks to the press, Dunn trusted her in-house legal and ethics team to
determine the lengths to which hired private detectives could go in
their investigations of board members.38 Unfortunately, the detectives
went so far as to draw felony counts of fraud and identity theft, not only
against themselves but also against Dunn.39 At issue was the practice of
37
James B. Stewart, The Kona Files, The New Yorker, 19 February 2007,
pp. 15267.
38
Stewart, The Kona Files, p. 159.
39
A judge later dropped all charges against Dunn. But by then the episode had
imposed a heavy cost in organizational turmoil, as well as an enormous personal toll
223
upon Dunn herself. Benjamin Pimentel, Court Dismisses Charges against Former HP
Chair, The San Francisco Chronicle, 15 March 2007, p. C1.
40
H.M. Bazerman and M.R. Banaji, The Social Psychology of Ordinary Ethical
Failures, Social Justice Research 17 (2004), pp. 11115; M.R. Banaji, Ordinary Prejudice,
Psychological Science Agenda 14 (2001), pp. 811.
41
D.A. Moore and G. Loewenstein, Self-Interest, Automaticity, and the Psychology
of Conflict of Interest, Social Justice Research 17 (2004), pp. 189202; N. Dasgupta,
Implicit Ingroup Favoritism, Outgroup Favoritism, and their Behavioral
Manifestations, Social Justice Research 17 (2004), pp. 14369; L.A. Rudman, Social
Justice in Our Minds, Homes, and Society: The Nature, Causes, and Consequences of
Implicit Bias, Social Justice Research 17 (2004), pp. 12942.
42
Jason Dana and George Loewenstein, A Social Science Perspective on Gifts
to Physicians from Industry, Journal of the American Medical Association 290 (2003),
pp. 25255.
224
maria w. merritt
225
The initial error of choosing an advantageous action under a misapprehension of it as morally acceptable can be seriously compounded
by overconfident faith in ones own good intentions.47 Broad characterological self-attributions further increase the hazard. The belief
I am a good person, or even I am trying my best to be a good person,
encourages the expectation I would not deliberately do anything
wrong. Even if this expectation were accurate, its content is quite distinct from that of a further expectation that can too easily slip into its
place during the everyday flow of the first-person perspective on our
experience and conduct: I would not do anything wrong, deliberately
or not. Even though we might not openly, in full awareness, credit ourselves with moral infallibilitythat would be immodestwe are prone
to limit our field of vision unwittingly, so that the effective reach of
critical self-monitoring extends only to those occasions when we do
take time to think carefully. Left out of consideration are occasions
when everything seems fine, and we are not in fact doing wrong deliberately, but had we deliberated we would have noticed decisively
wrong-making features of our course of action. Instead, if and when we
consider the latter type of occasion in retrospect, the conviction of
being a good person tends to displace deliberative effort into after-thefact rationalization, a selectively confirmatory survey of all the reasons
that seem to support the past decision. In hindsight, we may even
enhance our description of the action as not only morally neutral, but
downright laudable.48
Gigerenzer. See, e.g., Gerd Gigerenzer, Peter M. Todd, and the ABC Research Group,
Simple Heuristics That Make Us Smart (New York: Oxford University Press, 1999).
Gigerenzer has identified a form of intuitive decision-making, one-reason decisionmaking, which is ubiquitous in everyday reasoning and may often be useful, but which
might occasionally lead to moral misapprehension.When there are too many relevant
considerations to take into account in a limited amount of time, we recruit a single type
of consideration to stand in for the harder-to-determine, all-things-considered best
choice. See also Walter Sinnott-Armstrong, Moral Intuitions as Heuristics, unpublished manuscript.
47
Returning to the real-life example mentioned above, Patricia Dunn has reportedly said, There has never been a whiff of scandal or taint related to my activities,
particularly any issues concerning my integrity or ethics. Indeed, in the roles I have
held, any such taint would be an instant career-ender, and for good reason. Quoted in
Stewart, The Kona Files, p. 154. My point in quoting Dunn here is not to accuse her of
moral conceit, but rather to note that even seemingly well-founded confidence in the
quality of ones character can sometimes be ethically counterproductive.
48
John Darley, The Cognitive and Social Psychology of Contagious Organizational
Corruption, Brooklyn Law Review 70 (2005), pp. 117794; Elliot Aronson, The Social
Animal (New York: Worth Publishers, 9th edn, 2003), pp. 14399; R.M. Jones, Law,
226
maria w. merritt
227
228
maria w. merritt
Psychological studies of accountability in judgment and decisionmaking open a revealing window onto differential accessibility.
Accountability is defined, for the purposes of assessing these studies,
as the implicit or explicit expectation that one may be called upon to
justify ones beliefs, feelings, and actions to others.55 It turns out that
accountability simpliciter may or may not awaken the effortful cognitive activity of deliberation. Whether it will do so depends, in part, on
the expected characteristics of the audience to whom one is accountable. Numerous experimental and field studies have shown that expecting to be accountable to an audience whose views one knows ahead of
time will stultify cognitive effort, as compared with expecting to be
accountable to an audience whose views one does not know.56
When experimental subjects are asked to make judgments or decisions, and they know the preferences of the audience who will evaluate
the result, their behavior conforms to the use of cognitive shortcuts
guided by their beliefs about what the audience will find acceptable,
bypassing the cognitive effort required to assess the object of judgment independently.57 Subjects who do not know the preferences
of their audience, by contrast, anticipate the need to justify their judgment or decision, prompting a critical, reflective internal dialogue that
awakens deliberation. These subjects more readily tolerate evaluative
complexity and exercise care in considering complex information,
weighing the reasons for and against alternative options, and recognizing trade-offs. Accountability to an audience with unknown views has
been shown significantly to reduce subjects tendency to proceed with
undue haste and substitute oversimplified, superficial, or misleading
decision procedures for effortful, systematic, self-critical deliberation.
It has also been shown to attenuate the effects of a wide variety of cognitive biases, such as oversensitivity to the order in which information
appears.58
55
J.S. Lerner and P.E. Tetlock, Accounting for the Effects of Accountability,
Psychological Bulletin 125 (1999), pp. 25575, at p. 255.
56
Lerner and Tetlock, in Accounting for the Effects of Accountability, provide a
comprehensive review and analysis of the extensive body of research on accountability
and its impact on judgment and decision-making.
57
Lerner and Tetlock, Accounting for the Effects of Accountability; J. Pennington and
B.R. Schlenker, Accountability for Consequential Decisions: Justifying Ethical Judgments
to Audiences, Personality and Social Psychology Bulletin 25 (1999), pp. 10671081.
58
Lerner and Tetlock, Accounting for the Effects of Accountability; see especially
Table 1, p.260.
229
230
maria w. merritt
231
guidelines and under the law of many countries, it is subject to prospective and ongoing independent review, prompting investigators to
think more sensitively and systematically about the ethical aspects of
their research.64 Similarly, the empirical literature on decision-making
can inform the design of better mechanisms to promote accountability
on the part of corporate leaders.65
IV. Conclusion: Virtue Reconsidered
For people who care about living up to their own values, the activity
known to psychologists as self-regulationmonitoring ones movement toward or away from the desired self is practically inescapable.
The psychological evidence suggests that self-regulation influences
conduct partly through its engagement with the ongoing social
demands of self-presentation, often through cognitive processes that
are not accessible to first-person introspection or control.
The influence of these social demands may edify personal character,but may also disrupt iteven when exerted through relationships
with the same people on both counts, and even if both you and those
other people (let us assume for the sake of argument) have the right
values in principle. When your internal sense of ethical accountability is pervasively influenced by your relationships with close associates whose values you know, admire, and sharealong the lines of
the ideal Aristotelean social worldyou run the risk of living in an
incubator for your own cognitive complacency.66 Even if the shared
values are actually right, what threatens to dull your deliberative edge
is your comfortable familiarity with the attitudes held in common.
Familiarity breeds a pattern of overconfident, low-effort, intuitive
decision-making, which may sometimes fail to track important moral
64
Henry K. Beecher, Ethics and Clinical Research, New England Journal of Medicine
274 (1966), pp. 1354360; Allan M. Brandt, Racism and Research: The Case of the
Tuskegee Syphilis Study, Hastings Center Report 8 (1978), pp. 2129; National
Commission for the Protection of Human Subjects of Biomedical and Behavioral
Research, The Belmont Report, U.S. Government Printing Office, 18 April 1979; U.S.
Code of Federal Regulations Title 45, part 46, Protection of Human Subjects; Council
for International Organizations of Medical Sciences (CIOMS), International Ethical
Guidelines for Biomedical Research Involving Human Subjects (Geneva: World Health
Organization, 2002).
65
Jones, Law, Norms, and the Breakdown of the Board.
66
Contrast the contemporary ideal of social diversity in educational environments.
232
maria w. merritt
233
234
maria w. merritt
67
Cf. Hursthouse, On Virtue Ethics; Christine Swanton, Virtue Ethics: A Pluralistic
View (Oxford: Oxford University Press, 2003).
235
Aknowledgments
Credit for significant financial, institutional, and collegial support in
the production of this article is due to the Edmond J. Safra Foundation
Center for Ethics at Harvard University, where I held a Faculty
Fellowship in 2005-06. For thoughtful reading and discussion of an
earlier draft, I thank Dennis F. Thompson and the Faculty Fellows
seminar, and especially Renee M. Jones for extensive discussion,
exchange of ideas, and fresh insights into the psychological literature.
Precious Eboigbe provided excellent research assistance at Harvard.
Grateful acknowledgment for discussion and comments at various
stages of writing is also due to the following individuals and groups:
Neera Badhwar, Lorraine Besser-Jones, Alan Fuchs, Maximilian de
Gaynesford, Eric Hutton, Mark Leary, and Gopal Sreenivasan; the
Moral Psychology Research Group, especially John Doris, Gilbert
Harman, and Walter Sinnott-Armstrong; the Department of Philosophy
at Wake Forest University; and the audience at the Rocky Mountain
Virtue Ethics Summit, especially Rosalind Hursthouse, Karen Stohr,
Christine Swanton, and Kit Wellman. Some financial support for work
at earlier stages was provided by a Summer Research Grant at the
College of William and Mary.
238
jens timmermann
and noble deeds as high-flown emotional nonsense.3 It is indeed difficult to see how within Kants general doctrine of duty any act could be
morally admirable or commendable but not at the same time required.
Acts of wide or imperfect dutya sub-class of which are seenastheclos
est thing to supererogation by one prominent modern interpreter4
are no exception. Kant makes no provision for going beyond the call of
duty because duty leaves the agent little room for choice:
[W]here the moral law speaks, there is, objectively, no further room for
free choice with regard to what there is to be done (Critique of Judgement,
V, p. 210).
Such considerations give rise to the fear that actions intuitively classed
as morally commendable but not required must be re-classified as
commands of duty by Kant, making his ethical theory as unbearably
demanding as certain strict variants of direct utilitarianism.5
3
Cf. Critique of Practical Reason, V, pp. 8485: By exhortation to actions as noble,
sublime and magnanimous, minds are attuned to nothing but moral enthusiasm and
exaggerated self-conceit; by such exhortations they are led into the delusion that it is
not dutythat is, respect for the law whose yoke (though it is a gentle one because
reason itself imposes it on us) they must bear, even if reluctantlywhich constitutes
the determining ground of their actions, and which always humbles them inasmuch as
they observe the law (obey it), but that it is as if those actions are expected from them,
not from duty as bare merit; and V, p. 157. Kants warnings are mainly psychological
and do not as such settle the question of how much duty demands. Preaching romantic
ideals distracts from plain and simple duty and encourages a view according to which
some things in morality are less stringent than others, which directly conflicts with the
categorical nature of all moral commands.
4
Thomas Hill in his seminal article Kant on Imperfect Duty and Supererogation, in
idem, Dignity and Practical Reason (Ithaca: Cornell University Press, 1992), pp. 14575
(first published in Kant-Studien 61 [1971]).
5
I.e. variants of direct utilitarianism that do not admit of supererogation.
It should be noted that even in pre-philosophical moral thought the ascription of
an act as supererogatory is one often not shared by the agent, especially in cases of
so-called saintly or heroic acts. The doctor referred to in the opening paragraph
is unlikely to say that practising in a plague-ridden town is a particularly fine thing
to do for a doctor who goes beyond the call of duty. Rather, she would probably refer
to the importance of saving peoples lives, or that she could not just leave them to
die, etc. Similarly, the soldier would probably say that he could not let his comrades
down, and that any of them would have done the same thing in his situation. Such
actions are regarded as heroic or saintly only from the third-person perspective. This
seems at least some indication that supererogation is not as firmly grounded in
common sense as some people think. The more ordinary case of buying a toy for
the child of a friend seems to allow of more room for judgement. It is also worth notingthat the phrase that a certain act cannot be demanded of a person is ambiguous.
Those who deny the possibility in supererogation are not committed to saying that
individuals therefore have a right to request the performance of acts of duty from
others.
240
jens timmermann
Moral goodness always takes precedence over any other kind of value in
action, and not just that. Both subjectively and objectively, it silences or
eclipses any other claim to value, i.e. in a given situation nothing conflicting with moral value can have any value at all. The morally good
action is not just better, it is the only good option available to the agent.
There is nothing to be said for the prudent course of action. Kant
states that in deliberation I must first be certain that I am not acting
against my dutyonly afterwards am I permitted to look around for
happiness (ibid.).
In his Critique of Practical Reason Kant makes a similar point, this
time put forward in terms of the rationality of actions. This is his evocative chemical example of calcareous earth. Like a chemist, the philosopher can conduct an experiment with any mans practical reason
in order to distinguish the pure determining grounds of morality from
the empirical ones of inclination by adding the moral lawto the
7
What is the person motivated by who, contrary to both duty and momentary
desire, acts in accordance with his or her long-term interest? Even if prudence generates a motive analogous to respect in the moral case, as Kant sometimes seems to
The details of the chemical analogy need not concern us.9 The general
idea is sufficiently clear. Rationality determines the value of any kind of
action. As long as an agent considers possible courses of action from a
purely prudential point of view, reason points in the direction of his or
her advantage, of what is good for him or her. Moreover, he or she can
weigh and compare the various options: there can be a more and a less
in prudential matters.10 But moral reasons are different. They do not
acknowledge (cf. e.g. Refl. 1028), it cannot be effective in this case. Kants theory of
happiness-based motivation remains obscure.
8
Reading nun gnzlich for nur gnzlich.
9
Most of us will have a rough impression of what happens in experiments like
thatcountless chemistry lessons at school prove useful at last.
10
As one of the sentences leading up to the passage quoted indicates, the reason
why prudential judgements, unlike moral judgements, admit of degrees is to be found
in the empirical aspects involved in the pursuit of happiness: The distinction of the
doctrine of happiness from the doctrine of morals, in the first of which empirical principles constitute the whole foundation whereas in the second they do not make even the
smallest addition to it, is the first and most important task incumbent upon the
Analytic of pure practical reason, in which it must proceed as precisely and, so to speak,
as scrupulously as ever the geometer in his work [als je der Geometer in seinem Geschfte.
M. Gregor erroneously seems to read als jeder Geometer, as any geometer, and mistranslates the phrase accordingly, JT] (V, p. 92). The realization of our desires depends
on our knowing technical laws, which feed into hypothetical imperatives. They tell us
how to realize ends we happen to have. This is a highly complex empirical task.
Knowing what will make you happy in the long run is a matter of experience, and so is
the search for suitable means to that end. These means can be more or less suitable to
satisfy desires that can be more or less urgent. Kant optimistically contends that there
242
jens timmermann
just outweigh other, non-moral considerations. As soon as conflicting moral reasons enter the stage, the scene changes dramatically.
Moral considerations are accompanied by the characteristic rational
necessity of categorical judgements and consequently do not leave even
a single particle or appendage (Anhngsel) of rationality with what
previously appeared to be the prudent course of action favoured by
reason.11 The moral action alone is rational and good; all other value
depends on the unconditional value of moral goodness. Moral goodness does not just trump any other kind of practical value but completely annihilates it. (Note that makes Kants thesis of the necessity of
moral commands much stronger than standard modern overridingness.)12 If so, things are not looking good for supererogation. An action
correctly judged to be morally good cannot rationally be rejected in
favour of one that is not.
This may seem to leave open the possibility of rejecting a morally good
action (one that is not required) in favour of another action that is less
worthy but still morally good. But we have already encountered some
evidence that for Kant this is not a viable option either. The above
quotations from the Common Saying and the second Critique also
express the thought that the moral goodness defined by the categorical
imperativeas opposed to the prudential and presumably technical
goodness of hypothetical imperativesdoes not admit of degrees. It is
unconditional and makes an action necessary. There cannot be a lesser
kind of goodness that those engage in who do not opt for the extravagance of supererogatory acts.
There is further confirmation of this in a well-known passage from
the Metaphysics of Morals discussing the possibility of conflict amongst
different morally good courses of action. On strictly conceptual
can be no such vagaries in the moral sphere if moral judgement rests on pure a priori
principles.
11
The option that maximizes satisfaction of the agents desires in the long term presumably does not even deserve to be called the prudent choice any more (if the term
is supposed to imply normative approval).
12
This point is often overlooked, and rarely endorsed, by todays Kantians. A good
example is David Brinks Kantian Rationalism: Inescapability, Authority, and Suprem
acy, in Garrett Cullity and Berys Gaut (eds.), Ethics and Practical Reason (Oxford:
Clarendon Press, 1997), pp. 25591. Brink carefully discusses Kantian ethics in terms
of the three characteristics of moral judgement mentioned in the title of his paper but
stops short of the strong supremacy of moral value that renders all conflicting claims to
value null and void.
grounds, Kant rules out that there can be a conflict of duties. As duty
and obligation generally are concepts which express the objective practical necessity of certain actions, and two rules opposing each other
cannot both at the same time be necessary,13 it follows that a collision
of duties and obligations is unthinkable. However, there may be a conflict of grounds of obligation, one of which is not sufficient for obligation. If two such grounds conflict, the stronger groundnot the
stronger obligation!wins out to constitute a duty all things considered: fortior obligandi ratio vincit (Metaphysics of Morals, VI, p. 224).
The judgement of reason is unequivocal, even in the case of practical
conflict.14 It is true that other kinds of duty may follow from the vanquished ground. If I am unable to help a friend because my duty lies
elsewhere I may well incur a duty to explain the situation to her. But
when one ground of obligation loses out against another stronger
ground the latter does not constitute a lesser duty, but rather no duty
at all. Strictly speaking, the term duty should thus be used in the singular only, for all-things-considered token duties. This is the sense in
which a case of supererogatory action would have to go beyond the call
of duty. But there is no such case in Kantian ethics.15
In sum, the passages discussed present us with the following picture:
By virtue of the necessity of categorical judgement, moral value, all
things considered, (i) is absolute and does not admit of degrees; (ii) it
takes precedence over other kinds of practical value to the extent
that (iii) it even renders all other claims to practical value null and void;
and (iv) duty can be unequivocally determined even in cases of apparent moral conflict. We must conclude that there are no token actions
within Kants moral theory which are morally good but not required.
Good but not enforceable?Yes, that is just what duties of wide obligation are. Good prima facie or for the most part, but not all things
considered?Certainly, that will frequently be the case. Morally good
but not required?No, that would be a contradiction in terms.
13
In parentheses he adds: rather, if it is a duty to act in accordance with the one it is
not only not a duty to act in accordance with the opposed one, but even contrary to
duty.
14
It is, of course, conceivable that two actions are equally permitted. Two (or more)
actions can be permissible while it is obligatory to do either.
15
Morality may leave things undecided. In that case, two or more actions can be
permissible, but there would still be no opportunity for the agent to distinguish himself
as a saint or a hero. He would still have to do the one action or the other.
244
jens timmermann
2. Strict Duty, Wide Duty, Supererogation
This has been taken as proof for the thesis that imperfect duties, as
opposed to perfect duties, do admit of exceptions in favour of inclination.19 If so, maybe imperfect duties tell us to do actions that somehow
are morally good but not required. However, Kant does not say that
imperfect duties may sometimes be set aside in the interest of inclination. Rather, he denies that perfect duties allow of any such exceptions.
Saying the latter could admittedly be taken to imply the former; but
this conclusion is by no means necessary.
16
Also, wide duties are said to be meritorious, whereas strict duties are owed.
In the case of wide duties, the object of my duty (e.g. a person in need of help) does
not have a right to my action (my helping him) even though I do have a duty to help
him.
17
Cf. Hill, Meeting Needs and Doing Favors, pp. 203ff.; Imperfect Duty and
Supererogation, pp. 149ff.
18
Here the word beliebig does not mean arbitrary (Paton), as it does in standard
German today. Kant does not wish to say that the present classification is without foundation. That would be highly surprising. Rather, the present classification is provisional
and inadequate for the purpose of a metaphysics of morals because it does not contain
the distinction between duties of right and duties of virtue.
19
Cf. Hill, Imperfect Duty and Supererogation, p. 148; idem, Meeting Needs and
Doing Favors, p. 214 and n. 34; and Mary Gregor, Laws of Freedom (Oxford: Basil
Blackwell, 1963), p. 96.
246
jens timmermann
deny them).22 Again, you would like to employ double standards, making an exception for yourself without any good, impartial justification.
The reason why duties of the wider variety are subordinate to duties
of strict obligation is disclosed in the passage immediately following
the four illustrations. Kant there states the distinction between the two
types of duty in the following way:
Some actions [actions contrary to strict/perfect duty, JT] are so constituted that their maxims cannot even be thought without contradiction as
a universal law of nature, far less could one will that they should become
such. In the case of others [actions contrary to wide/imperfect duty, JT]
that inner impossibility is indeed not to be found, but it is still impossible
to will that their maxim be elevated to the universality of a law of nature
because such a will would contradict itself (IV, p. 424).
Far less could one will that they should become such! Kant reveals the
reason for the primacy of strict duty in this sentence: all actions contrary to duty cannot possibly be willed to be universal laws, they all
generate what has been called a contradiction in the will; some actions,
in addition, also generate a contradiction in conception because their
maxim cannot even be conceived as universal if we want it to be successful. Maxims that, if universalized in the Kantian manner, generate
a contradiction in conception also generate a contradiction in the will,
whereas the reverse is not true. Actions contrary to wide duty only generate a contradiction in the will if we try to make them universal in
accordance with the test of the categorical imperative. They are therefore restricted by duties of strict obligation, the violation of which creates the additional difficulty of a contradiction in conception. In fact,
the categorical imperative in its first and primary formulation addresses
the question whether we can will the maxim to become a universal law,
covering all four possible cases, not whether we can think it as such.
Act only on a maxim that you can think a universal law! would be the
categorical imperative of strict duty.23
Similarly, the second variant formulation of humanity as an end
in itself provides us with two partially overlapping criteria: treating
persons as mere means to ones own ends (the strict criterion) and not
22
There is arguably even a particularly strong rational commitment for you to
accept other peoples help should you require it because you endorse your own
selfishness.
23
The details of this need to be spelt out. Not everything that seems to be a strict
duty turns out to be one. There may be cases of suicide or self-killing which should not
be subsumed under morally bad self-murder. Judgement must decide between maxims and imperatives that conflict.
248
jens timmermann
neighbour in general by love of ones parents), by which in fact the field
for the practice of virtue is widened (VI, p. 390).
250
jens timmermann
252
jens timmermann
do more than I owe them: these are duties of virtue. In their case, I impose
a new duty on the others, rather than merely doing my duty as is the case
with duties of right. The other person does not have a right to duties of
virtue (Moral Mrongovius II, XXIX, p. 631).
254
jens timmermann
a patron) that he will not count it as an obligation towards him (Moral
Mrongovius, XXVII, p. 1559; cf. the briefer account in Moralphilosophie
Collins, XXVII, p. 442).
Accepting favours can destroy the equality required for friendship; and
the fact that an act of assistance would be a burden is a very good
moralreason against it. We must similarly avoid being intrusive, patronising or paternalistic. Other-regarding action may be morally bad
even if the recipient happens to benefit. Partly with the intent of meeting these difficulties, Kant suggests earlier in the lectures that some
basic acts of assistance be re-classified as acts of strict duty:
One may have a share in the general injustice, even though one does
nobody any wrong by civil laws and institutions. So if we now do a kindness to an unfortunate, we have not made a free gift to him, but repaid
him what we were helping to take away through a general injustice. For if
no-one might appropriate more of this worlds goods than his neighbour,
there would be no rich people, but also no poor. Thus even acts of kindness are acts of duty and indebtedness, arising from the rights of others
(Moral Mrongovius, XXVII, p. 1539; Moralphilosophie Collins, XXVII,
p. 416).37
37
Cf. Even the civil constitution is pre-arranged so that we participate in public and
general oppressions, and thus we have to regard an act we perform for another, not as
an act of kindness and generosity, but as a small return for what we have taken from
him in virtue of the general arrangement (Moral Mrongovius, XXVII, p. 1552;
Moralphilosophie Collins, XXVII, p. 432).
256
jens timmermann
40
The empirical elements of prudential deliberation make it much more complicated. Cf. Groundwork, IV, pp. 401402.
41
There seems to be no exact decision procedure: How far should one expend ones
resources in practicing beneficence? Surely not to the extent that one would finally
come oneself to need the beneficence of others (Metaphysics of Morals, VI, p. 454). If
we go to extremes, rational inconsistency looms, as this quotations clearly shows. Still,
apparent moral value can only be outweighed by moral value, and what remains is
duty full stop yet making other people happy is not an unrestricted or unconditional duty.
260
jonathan webber
of stimuli, and that it is compatible with the traditional theory of character if traits are conceived as dispositions to be inclined with a certain
strength to behave in a certain way in response to a certain kind of
stimulus. We will see in section five that this latter conception is prominent in Aristotles Nicomachean Ethics and in recent characterological
ethical discourse influenced by it. Section six presents a pragmatic reason to prefer this conception of traits to the behaviourist one favoured
by proponents of the fragmentation theory, and the final section considers some implications of this discussion for future philosophical
investigations of character and virtue.
2. Specious Evidence
There are five kinds of experiment cited in favour of the fragmentation
theory, all of which are cited by Doris, and two of which are cited by
Goldie and Merritt.5 Of these five experiments, as we will see,twoshould
be discounted. There is also a sixth experiment discussed by Doris, but
although he has been criticized for taking this experiment as evidence
for his theory, he does not actually use it in this way, as we shall see.
One of the experiments to be discounted seems to indicate that people are more likely to exhibit helping behaviour if they have recently
had good luck. The subjects of this experiment were unaware that the
experiment was taking place. Having made calls from a telephone
booth, some found a dime in the coin return slot and others looked but
found nothing. Each subject was soon walking behind a woman who
dropped a folder full of loose papers and began to gather them up. Of
those who had just found a dime, 96 per cent helped her. Of those who
had not, just 13 per cent helped.6 All three advocates of the fragmentation theory cite this experiment.7 The results are taken to challenge the
5
Gilbert Harman presents a similar but distinct argument in his paper Moral
Philosophy Meets Social Psychology: Virtue Ethics and the Fundamental Attribution
Error, Proceedings of the Aristotelian Society 99.3 (1999), pp. 31531. He employs some
of the data employed by Doris, and argues that character is illusory and so characterological discourse should be abandoned altogether. I cannot do justice to his argument
here, but see my Character, Common-Sense, and Expertise (forthcoming).
6
Summary of Alice M. Isen and Paula F. Levin, Effect of Feeling Good on Helping:
Cookies and Kindness, Journal of Personality and Social Psychology 21.3 (1972),
pp. 38488, Study II.
7
Doris, Lack of Character, pp. 3032; Merritt, Virtue Ethics and Situationist
Personality Psychology, p. 366; Goldie, On Personality, pp. 3032.
262
jonathan webber
264
jonathan webber
whether they had cheated in those tests, fake a record of their athletic
performance, steal a small amount of unattended money, and so on.
The study found a high correlation between the proportion of a group
of children behaving dishonestly in one situation and the proportion of
the same group behaving dishonestly in another situation of the same
type. Children who cheated in an exam, it seems, are likely to cheat in
the next exam. But it also found very little consistency across types of
situation. A child who cheats in an exam seems no more likely than any
other child to steal unattended money or fake a record of athletic performance. So among children at least, there seem to be no unified traits
of honesty or dishonesty, only narrower traits such as being an examcheat or an unattended-money-thief.
Rachana Kamtekar raises a methodological concern about the use of
this data in the debate over character, asking whether it is legitimate to
make inferences about adult character traits from observation of children who might be more impressionable, less committed to particular
ideals of conduct, or less integrated than adults.12 But this is simply
unfair to Doris, who does not use the data in this way. He is aware that
children with developing personalities are plausibly thought to exhibit
less behavioural consistency than fully formed adults, and discusses
these experiments not so much for their evidential role as for the interpretive perspective they provide.13 In this, he is right. This data might
illuminate the development or structure of the trait of honesty, but cannot provide evidence against the idea that adult behaviour is regulated
by such general character traits as honesty, generosity and courage.14
We will therefore set it aside.
3. Significant Evidence
The three remaining kinds of experiment do provide reliable evidence
concerning adult character traits. Probably the best known, and
12
Rachana Kamtekar, Situationism and Virtue Ethics on the Content of Our
Character, Ethics 114.3 (2004), pp. 45891 (466 n. 30).
13
Doris, Lack of Character, p. 63.
14
Other objections to the use of this data are raised by Gopal Sreenivasan in Errors
About Errors: Virtue Theory and Trait Attribution, Mind 111.441 (2002), pp. 4768,
and by John Sabini and Maury Silver on pp. 54044 of their Lack of Character?
Situationism Critiqued, Ethics 115 (2005), pp. 53562. These criticisms are misplaced,
as Doris rightly does not use this data as evidence. See also note 23 below, and my
Character, Consistency, and Classification (forthcoming).
266
jonathan webber
20
Summary of Bibb Latan and John M. Darley, The Unresponsive Bystander: Why
Doesnt He Help? (New York: Meredith, 1970), pp. 5760. See also Doris, Lack of
Character, pp. 3233.
268
jonathan webber
and of those who had been told that they would be early, 63 per cent
stopped.21
4. Interpreting the Data
It is unclear exactly why proponents of the fragmentation theory take
this data to be evidence in their favour. Part of the problem is that all
three give ambiguous definitions of the position they are opposed to,
the understanding of character purportedly prevalent in ethical discourse. Doris writes that a trait as traditionally construed would be
reliably manifested in trait-relevant behaviour across a diversity of
trait-relevant eliciting conditions that may vary widely in their conduciveness to the manifestation of the trait in question. Merritt writes that
a trait as traditionally construed will reliably give rise to the relevant
kind of behaviour, across the full range of situations in which the
behaviour would be appropriate, including situations that exert contrary pressures. Goldie writes that we tend to think of character traits
asstable and consistent in their manifestation in thought, feeling and
action across a wide range of different situationshonest people can
be relied on to act honestly wherever honesty is appropriate.22 What is
meant by a situation being conducive to a certain kind of behaviour, or
that kind of behaviour being appropriate in that situation?
These could be read as moral terms: the proponents of the fragmentation theory could be taken to be arguing that the data indicates that
whether or not one will respond correctly to a moral demand for compassionate behaviour depends not on whether one is disposed to do so,
but on further situational features such as instructions from an authority figure, ones degree of hurry, and the passivity of bystanders. Some
of their critics certainly take this to be their argument. Michael DePaul
and Christian Miller, for example, have both claimed that the data is
21
Summary of J.M. Darley and C. Daniel Batson, From Jerusalem to Jericho:
A Study of Situational and Dispositional Variables in Helping Behaviour, Journal of
Personality and Social Psychology 27.1 (1973), pp. 100108. See also Doris, Lack of
Character, pp. 3334; Merritt, Virtue Ethics and Situationist Personality Psychology,
p. 366; and Goldie, On Personality, pp. 6263.
22
Doris, Lack of Character, p. 22; Merritt, Virtue Ethics and Situationist Personality
Psychology, pp. 36566; Goldie, On Personality, p. 50. Doris provides much the same
understanding of traits in his earlier paper, where he first puts forward his argument
against the traditional conception of character: Persons, Situations, and Virtue Ethics,
Nos 32.4 (1998), pp. 50430 (506507).
270
jonathan webber
virtues, but that they do not have any character traits as traditionally
conceived. This is also the claim made by Merritt and Goldie.
That character traits and virtues are indeed conceptually distinct will
be illustrated in more detail in the next section. Here it is sufficient to
note that the idea that one can be habitually too courageous or inappropriately honest is a philosophical commonplace. We should therefore take the terms conducive and appropriate in these definitions of
character traits as non-moral, descriptive terms: the proponents of the
fragmentation theory take the data to indicate that whether or not one
will respond to the apparent distress of another person by trying to
alleviate it depends not on whether one is disposed to do so, but on
situational features such as instructions from an authority figure, ones
degree of hurry, and the passivity of bystanders.
It is unclear just why they understand the data this way. Goldie does
not formulate an argument. Merritt does so, but only briefly. She claims
that the data shows behavioural dispositions to be related to such ethically arbitrary situational factors as ones degree of hurry or the presence of passive by-standers, and that the traditional understanding of
character should lead us to expect that this is not the case.26 But without further elaboration, it is difficult to see why the traditional theory
of character should lead us to expect that people do not have commitments to punctuality or to peer-approval strong enough to preclude
their offering help in certain types of situation.
Doris is more explicit. He hinges his argument on the claim that: If
a person possesses a trait, that person will engage in trait-relevant
behaviours in trait-relevant eliciting conditions with markedly above
chance probability p.27 We can make sense of this notion of chance
behaviour only as a comparison between the agent and the population
at large: if a person possesses a certain trait, then the probability of that
person behaving in a way that manifests that trait in a given situation is
significantly greater than the probability of a randomly chosen member
of the population doing so. But this seems to imply that traits are comparative notions, and that it is incoherent to hold that a whole population could possess any given trait. This notion is therefore distinct from
the notion traditionally employed by theories that encourage us all to
develop certain traits. But Doris does not want to provide such a comparative theory of traits: he agrees that in principle, every individual in
Merritt, Virtue Ethics and Situationist Personality Psychology, p. 366.
Doris, Lack of Character, p. 19.
26
27
272
jonathan webber
of about the same strength whenever the subject is confronted with the
relevant kind of situational feature where this is described in nonnormative terms, though as we shall see in the next two sections this
inclination need not necessarily be consciously felt in order to play the
explanatory role required of it.
If we employ this regularity theory when interpreting the data, we
find the positive evidence in favour of differing character traits that
Doris urges us to look for: as we have seen, the subjects of the Samaritan
Experiment differ in their inclinations towards helping a distressed
stranger and towards punctuality; the subjects in the Bystander
Experiment differ in their inclinations towards helping an apparently
distressed stranger and towards winning or maintaining peer-approval;
and the subjects in the Obedience Experiment differ in their inclinations towards obedience to authority and their inclinations against
inflicting pain.31
As we will see in the next section, moreover, the regularity theory
underlies key aspects of Aristotles ethics and of some prominent recent
works influenced by it. The argument against characterological ethical
discourse mounted by Doris therefore misses its target: so long as that
discourse retains the regularity theory, the data Doris cites will present
no threat to it. But this is not yet to show that there is any positive reason to prefer the regularity theory to the fragmentation theory, which
will be shown in section six.
5. Character in Ethical Discourse
Various key aspects of characterological ethical discourse involve this
notion of a trait as a relatively stable disposition to be inclined with a
certain strength towards a certain kind of behaviour in response to
a certain kind of situational feature. The wellspring of the tradition,Aristotles Nicomachean Ethics, contains a number of prominent
31
As we saw in section 3, the authority involved in the Obedience Experiment is
that of an institutional expert and moral guide, not simply that of the person in charge.
Sabini and Silver question whether the behaviour of the subjects in the Samaritan
Experiment should be described in terms of punctuality, preferring instead to invoke
potential embarrassment at arriving late (Lack of Character? Situationism Critiqued,
p. 558). But these explanations seem complementary rather than contrary. Potential
embarrassment could enforce ones acting on ones commitment to punctuality, and
without that commitment it is difficult to see why somebody would be embarrassed by
being late rather than proud of having stopped to help.
274
jonathan webber
e xamples, and descendents of these ideas prominent in recent literature retain the conception of trait they involve, as we shall see.
Aristotle makes it abundantly clear that the kinds of traits that he
recommends as virtuous are certainly not dispositions to behave in a
certain kind of way whenever or almost whenever a certain kind of
situational feature is present. He considers such traits to be vices
of excess: giving money away at every or almost every opportunity is
not generosity, but profligacy; never or hardly ever recoiling from danger is not courage, but foolhardiness; and so on. The vice consists not
just in behaving in a certain way too often, but in behaving in that way
on inappropriate as well as appropriate occasions. Possessing a specific
virtue, on the other hand, means that one will act in a certain way
when one should, towards the things one should, in relation to the
people one should, for the reasons one should, and in the way one
should.32 Someone might commit murder in the face of real danger, but
would not thereby display the virtue of courage, since the intrepid act
was inappropriate for other reasons. Someone who tells the truth even
in situations where the demands of compassion make it inappropriate
does not fully possess the virtue of honesty.
Underlying this theory is the idea that each trait leads one to be
inclined with a certain degree of strength towards certain kinds of
actions in certain kinds of situations, and that such traits are virtues
only if these inclinations are tempered by other inclinations that constrain the range of occasions on which they will result in action. This
explains why Aristotle takes the full possession of one virtue to require
full possession of all virtues, and to hold that virtue in the primary
sense consists in this possession of full virtue. There is, for Aristotle, a
single web of interdependent virtues: full possession of any one virtue
means habitually being inclined to behave in a certain way with the
right degree of strength in the presence of a certain situational feature,
where what is right is relative to the strength of ones other habitual
inclinations in response to other possible situational features.33
John McDowell has argued for a similar position on the grounds that
virtue issues in nothing but right conduct.34 Philippa Foot, on the
32
Aristotle, Nicomachean Ethics, translated by Christopher Rowe, with introduction
and commentary by Sarah Broadie (Oxford: Oxford University Press, 2002). Quotation
is from 1106b2123; see also 1115b1520.
33
Nicomachean Ethics, 1144b321145a2.
34
John McDowell, Virtue and Reason, Monist 62 (1979), pp. 33150. Reprinted in
his Mind, Value, and Reality (Cambridge, MA and London: Harvard University Press,
1998), 2.
other hand, has disagreed with such a stringent criterion for virtue,
arguing that a trait counts as virtuous so long as it usually issues in right
conduct. On the occasions when it does not, it is not operating as a
virtue because its manifestation ought to have been precluded by
another trait, which the agent unfortunately lacks. But one can, according to Foot, possess a virtue that does not always operate as a virtue.35
McDowell and Foot nevertheless agree that the traits in question are
dispositions to be inclined with a certain degree of strength to a certain
kind of action in response to a certain kind of situational feature; they
disagree only over the criteria for such traits to be counted as virtuous.
Where McDowell holds that the murderers trait of courage cannot be
virtuous because it is not accompanied by the right level of aversion to
killing, Foot holds that it is virtuous so long as it is usually manifested
in good behaviour.
Aristotles discussion of self-control and weakness of will also seems
to involve this notion of character trait. The difference between g enuine
virtue and mere self-control, he claims, is that only the latter involves a
struggle to overcome contrary inclinations; the difference between vice
and mere weakness of will is that only the latter involves such a struggle; and the difference between self-control and weakness of will is that
the inclination one believes one ought to act on wins the struggle in the
former case, loses it in the latter. The picture is complicated, of course,
by the fact that one might have an additional disposition towards being
inclined to do whatever one believes one should do. But even when
such a trait is in play, it seems that weakness of will results from the
inclination acted upon being stronger than the combination of the
inclination one believes one should act on and the inclination to do
whatever one believes one should do. Self-control results from this
combination being stronger than the inclination to behave otherwise.
McDowell has recently echoed this view, claiming that the virtuous
person acts for a reason that is apprehended, not as outweighing or
overriding any reasons for acting in other waysbut as silencing
them.36 The virtuous person pursues the right course of action without
consciously considering any other. One might disagree and argue that
an action is virtuous even when the right reason is apprehended
as more important than reasons for contrary actions, whereas merely
35
Philippa Foot, Virtues and Vices, in her Virtues and Vices and Other Essays in
Moral Philosophy (Oxford: Blackwell, 1978), III.
36
Aristotle, Nicomachean Ethics, book 7 chapters 110, esp. 1146b315; McDowell,
Virtue and Reason, 3.
276
jonathan webber
278
jonathan webber
further claim that belief in this deeper level of explanation is the reason
why predictions of behaviour often fail. One way to decide whether this
is true would be to undertake longitudinal studies. If consistent observation and classification of the actions of a few individuals over significant stretches of time were to yield reliable predictions of their
behaviour in novel situations, then the traditional notion would be
shown to be preferable to the behaviourist notion. If it were to yield no
such reliability, then the traditional notion would have no such
advantage.
Such longitudinal studies would require long-term detailed surveillance of the public and private lives of people unaware of being under
surveillance and are therefore unavailable to professional psychologists
for both logistical and ethical reasons.39 But these kinds of studies are
not necessary for settling this issue. Latitudinal studies, in which different subjects are tested in the same situation, can provide evidence of
this deeper level of explanation, and hence also evidence of character
traits as traditionally construed.
One study that provides such evidence was originally designed to
explain the following difference in homicide rates between the n
orthern
and southern regions of the USA: white males in large southern cities
are no more likely to commit homicide than are their counterparts in
large northern cities, but those who do commit homicide are significantly more likely to do so as a result of an argument; and white males
in rural areas and small urban areas in the southern states are no more
likely to commit homicide as a result of a crime than are their northern
counterparts, but are twice as likely to commit homicide as a result of
an argument.40
The experimenters tested the hypothesis that these differences were
due to a southern culture of honour, more pronounced in rural and
small urban areas than in larger, more cosmopolitan cities. Their subjects were white, non-Jewish, non-Hispanic, male citizens of the USA
studying at the University of Michigan. They were divided between
39
Some philosophers argue that longitudinal studies are available in the form of
novels and the writings of other acute observers of human behaviour. I defend this
view in Character, Common-Sense, and Expertise. But since the traditional notion of
a character trait can be shown to be preferable to its behaviourist rival by reference to
recent experimental data, there is no need to make a more controversial appeal to nonexperimental literature here.
40
See Richard E. Nisbett and Dov Cohen, Culture of Honor: The Psychology of
Violence in the South (Oxford and Boulder, CO: Westview Press, 1996), p. 21.
280
jonathan webber
In the third, chicken variation, the subject returning along the corridor was faced with somebody coming the other way, who was six feet
and three inches (one metre and 90 centimetres) tall, weighed 250
pounds (17 stone and 12 pounds; 114 kilograms), and played college
football. All subjects gave way. They differed in the distance at which
they did so. Insulted southerners yielded at an average of around three
feet (90 centimetres), control southerners at about nine feet. Insulted
northerners yielded at around six feet, control northerners at around
five and a half. The insults therefore made a significant difference in the
case of southerners, but not in the case of northerners.42
If we adopt the behaviourist notion of traits, these experiments
suggest only that none of the subjects have the trait of responding
violently to insults in experimental situations, that southerners are
more likely than northerners to have the trait of completing the story
violently after being insulted in an experimental situation, and that
southerners are more likely than northerners to have the trait of displaying increased bravado shortly after being insulted in experimental
conditions. These are distinct traits. The data concerning cortisol and
testosterone levels does not report any behaviour, and so does not indicate any difference in traits. The traits that are indicated, moreover,
cannot be linked to any data concerning homicide occurring outside
experimental conditions, since the traits manifested are relative to
those conditions. Indeed, since homicide is committed in a wide range
of situations, the behaviourist notion of traits precludes any explanation of the homicide data in terms of trait differences between southerners and northerners.
The traditional notion of traits, on the other hand, allows us to
explain the homicide data and the data from the story and chicken
variations of the experiment in terms of a trait prevalent among southerners but not northerners: a disposition to be strongly inclined to
respond violently when insulted. Evidence of the inclination elicited
by insults is provided by the saliva tests. The fact that nobody actually
did respond violently can be explained by the other relevant inclinations in play in the experimental situation. This information allows
predictions of results of future experiments that also test reactions to
insults or slights among northerners and southerners, but test these in
42
Nisbett and Cohen, Culture of Honor, ch. 4. For further data in support of the
theory of a southern culture of honour, see chs. 3 and 5; for more on what this culture
involves, see chs. 1 and 6.
282
jonathan webber
part V
*I owe thanks to Barbara Herman, Seana Shiffrin, Michael Otsuka, Frances Kamm,
Jeff McMahan, Derk Pereboom, David Christensen, Donald Loeb, Shelly Kagan, Kai
Draper, and my colleagues at the Philosophy Department at Oberlin for comments on
this manuscript: Al MacKay, Peter McInerney, Martin Thomson-Jones, Todd Ganson,
and Dorit Ganson. Thanks, too, to anonymous referees and to Thom Brooks for constructive criticisms and editing suggestions.
288
timothy hall
acting. Yet the example above involves action, and many respondents
will claim that Jones lets Smith die.1
Matters are further complicated by examples such as the following:
The Enemy: Smith, a terminally ill patient, is in great pain, but wishes to
continue to be kept alive by life support devices. Smiths enemy, Jones,
learns of Smiths incapacity and sneaks into the hospital. Jones disconnects Smith from the life support device, and Smith dies.2
of the m
etaphysics of withdrawals. These accountswith the exception
of Hansers, and Boorse and Sorensenscharacterize withdrawals as
the doing or allowing of harm, with the aim of employing that metaphysicalcharacterization in subsequent moral analysis of withdrawal
behaviors.
I wish to offer an alternative view. Withdrawals of aid have a greater
significance than as a proving ground for broader theories of doing
and allowing harm. I will argue that withdrawals of aid reveal that nonconsequentialism is best understood as consisting in three broad, morally distinct categories of behavior where commentators have ordinarily
identified two. I will call these categories standard doings of harm,
standard allowings of harm, and denials of resources.
I will argue that, with one exception, denials of resources are morally
distinct from standard doings or allowings of harm. They are morally
distinct because their moral status depends to a large extent on a right
to the resource denied. The moral statuses of standard doings and
allowings of harm, on the other hand, are not influenced to a great
degree by such rights. The exception is cases in which a person is herself the resource that she denies. Those cases are morally on a par with
standard allowings of harm.
A distinctive feature of my view is that the claim of moral distinction
is independent of whether denials of resources are the doing or allowing of harm. So the metaphysical question of whether denials of
resources are doings or allowings of harm can be left aside in understanding their moral status, I shall argue. I qualify standard doings and
allowings of harm as standard precisely to leave open the possibility
that denials of resources are also doings or allowings of harm.
I will proceed by considering Jeff McMahans analysis of the withdrawal of aid in section II. McMahans is the most sophisticated attempt
to analyze the moral status of withdrawals of aid in terms of doing and
allowing harm, but my comments will be applicable to other views of
this kind. In section III, I offer an argument for the moral distinctness
of denials of resources based on a comparison of well-matched examples. In section IV, I argue that the rights of persons conjoined with any
of a range of widely held justifications for the distribution of resources
in political society imply the moral distinctness of denial behaviors.
(New York: Oxford University Press, 1996), p. 28; Jonathan Bennett, The Act Itself
(New York: Clarendon Press, 1995).
290
timothy hall
Though McMahan has been concerned largely with lethal harms, and thus the narrow distinction between killing and letting die, the full body of his work makes clear
that he means to characterize a broader distinction between doing and allowing harm.
See McMahan, The Ethics of Killing, p. 381; Killing, Letting Die, and Withdrawing Aid;
A Challenge to Common Sense Morality.
6
See Foot, Euthanasia, pp. 85112.
7
I have simplified McMahans account here somewhat, in part for expository
purposes, but in part for other reasons. In Killing, Letting Die, and the Withdrawal of
Aid, and in The Ethics of Killing, McMahan offers a further condition of a withdrawals
being the allowing of harm: that the aid be in progress or continuing, and not selfsustaining. As far as I can tell, however, this condition is redundant. It is equivalent
to requiring that the harmed party must have been threatened independently of the
5
292
timothy hall
violinist, which will result in his death. You disconnect yourself, and the
violinist dies.10
Many people would agree that Jones lets Smith die. Jones takes back
what is his, what he would have been permitted to refuse to provide
initially. So, Jones allows a threat to run its course, rather than doing
harm to Smith.
McMahans analysis of the withdrawal of aid has two problems. The
first is that not all careful readers will agree with McMahans intuitions
about doing and allowing harm in the examples above. Some respondents will be unwilling to say that the behavior in question is either
doing or allowing harm. For example, I can attest that some readers
will claim that Jones neither kills nor lets die in the examples above.
They will limit themselves instead to the claim that Jones withdraws aid
from Smith and Smith dies as a result.11
Some readers will, of course, share McMahans intuitions. Rather
than attempt to resolve this issue, however, I will leave it aside. For I am
interested in an independent problem of moral explanation.
McMahans account is, strictly speaking, a metaphysical one.
However, it is clear that a large part of his motivation is normative.
10
Judith Thomson, A Defence of Abortion, Philosophy and Public Affairs 1 (1971),
p. 5.
11
On Hansers view, Jones prevents Smith from being saved, rather than killing
Smith or letting Smith die. Jones acts to deny to Smith what Smith would otherwise
save himself with. So, Hanser claims, Jones does not let Smith die. On the other hand,
Jones does not cause the threat to Smith that kills him. So, Jones does not kill Smith,
either, according to Hanser. See Hanser, Killing, Letting Die, and Preventing People
from Being Saved, pp. 27795.
294
timothy hall
question, however. Indeed, one must stretch to explain how the latter
two possibilities are cases of providing at all. Merely having a right to a
resource used by someone else is not obviously accurately described as
a case of providing. Nor is having ones presence be useful to someone
else obviously a case of providing.
Perhaps, though, one might claim that there is an attenuated sense of
providing that requires only that resources used by someone be owned
by a person to be provided by that personin something like the sense
it might seem correct, or at least not clearly wrong, to say that a victim
provides money to a thief. We might, then, notice that in the case in
which one is a resource one also has a right to the resource. For one has
a right to ones own body. Thus having ones body be useful to another
might be another way of providing.
However, the fact that an ownership right to a non-person resource
might well haveand on many views does havea very different
source than a right to oneself weakens the similarity between these two
ways of providing, even if they are really ways of providing. Furthermore,
having a right to a resource is still a very different matter than physically providing a resource. Physically providing a resource is a physical
act. Ownership is a moral claim.
One might claim that, in the appropriate conditions, physically providing a resource also generates a right to that resource. Then one
might claim that being a resource, owning a resource, and physically
providing a resource are all alike: they are ways one might have a right
to a resource.
However, physically providing a resource does not, in normal circumstances, give any right to the provider. The right to the resource is
usually predetermined by ownership or some other claim. I do not
come to have a right to Smiths life-support machine, for example,
merely by connecting Smith to it.
Furthermore, the resulting interpretation leaves aside McMahans
notion of provision altogether. Rather than there being three ways of
providing a resource, these three possibilitiesphysically providing a
resource, owning a resource, and being a resourcewould simply be
three ways of having a right to resource.
Finally, and most importantly, there would still be a gross dissimilarity between standard cases of allowing and withdrawal-allowings.
Withdrawal-allowings would be those allowings in which an agent has
some right to the resource he withdraws, whether through ownership,
being the resource, or having physically provided the resource.
296
timothy hall
and too different from the paradigm cases of doing or allowing harm to
permit a satisfying generality of moral explanation across all doings
and allowings. This problem is not just an idiosyncracy of his account,
however. As I shall now argue, there is a significant moral difference
between withdrawal behaviors and paradigm examples of doing and
allowing harm.
III
In this section, I will argue that a category of behaviors called the denial
of resources is morally distinct from paradigm doings or allowings of
harm. I argue that this moral difference is independent of whether or
not the denials are doings or allowings of harm.
Before going on, I need to make some phrases clear. A standard
doing of harm is an agent acting and creating a threat that harms someone. Examples include paradigmatic cases of harming others, such as
shooting, stabbing, beating, or poisoning. Standard allowings of harm
are cases in which there is some threat of harm to another person
uncaused by, and otherwise independent of, the agent. Though the
agent could act to eliminate, forestall, or otherwise protect against the
threat, the agent does not, and harm befalls the threatened person.
A paradigm example would be an agent capable of aiding standing by
while another person drowns. The third category is what I will call the
denial of resources. By a resource, I mean something apart from a person herself that offers the prospect of being useful in furthering or protecting her interests. Sometimes, other persons can be a resource,
though most often I will be considering cases of non-person resources.
If there is a person to whom there is a realistic prospect of some
resources being useful, and another person, an agent, who acts to make
the resource unavailable to the first person, the agent denies the
resource to the first person. Withdrawals of aid just are one way of
denying resources: others include consumption of resources, the imposition of barriers to resources, or destroying resources.
My first argument for the moral distinctness of denials of resources
begins with a comparison of matched examples. My method differs
from a common one in normative ethics. Typically, a writer interested
in arguing for the moral significance of some factor of interest will
compare behaviors in circumstances in which all other things are equal,
except for the factor of interest. If there is some all things considered
14
See the qualifications in the use of this method discussed in Kagan, The Limits of
Morality, pp. 1115; Kagan, The Additive Fallacy, Ethics 99 (1988), pp. 531; Kamm,
Morality, Mortality, vol. II, pp. 1786; Peter Unger, Living High and Letting Die
(New York: Oxford University Press, 1996), pp. 1023.
15
At least, Smith is not permitted to use violence if he is aware of the provenance of
the antidote.
298
timothy hall
17
Boorse and Sorensen call this effect the multiplier effect, as does Kamm in the
context of abortion. See Boorse and Sorensen, Ducking Harm; Frances Kamm,
Creation and Abortion (New York: Oxford University Press, 1992), pp. 2041.
300
timothy hall
Joness action in this case is wrong, most people would agree. Smith,
furthermore, would be permitted to use even lethal force to save his
own life, if he were able and if that were necessary for him to save his
life.
A second case:
Drive, Smiths Antidote: Jones is in a public park, and accidentally touches
a poisonous plant. Jones had made some provision for the possibility, but
sold his antidote some time ago. It happens that Smith is napping nearby,
and Smiths antidote to this poison is just a further, short distance away.
Because Smith is a sound sleeper, for Jones to get to Smiths antidote in
time, Jones would have to drive over Smith. Jones drives ahead, killing
Smith.
In this case too, most people would agree, Joness action is wrong and
Smith would be permitted to use even lethal force to save his life, were
he able and were it necessary.
In each of these standard doings of harm, Joness action is impermissible. The fact that Jones has a right to the antidote in one case and not
in the other does not make a difference in the permissibility of his
behavior. Yet these cases are comparable to the denial cases discussed
above. In both the denial and standard killing cases, Jones either acts
with the result that Smith dies or Jones himself dies. Joness action in all
cases is necessary for his using the antidote. In none of the cases need
Jones aim at the death of Smith to acquire the antidote.18
One might object that Drive, Joness Antidote is importantly dissimilar to Jones
Antidote. In the latter case, Smith is vulnerable because of his need for the antidote. Yet
Smith has no right or claim to the antidote. In Drive, however, Smith is vulnerable
because of his position in the park. Yet, Smith has a right to sleep in the park. So, in
Joness Antidote, Smith is denied only that to which he has no right, but he is killed in
Drive, Joness Antidote where he has a right to be.
We might imagine that there is a rule against sleeping in the park, however. Even
with Drive, Joness Antidote so modified, it is not permissible for Jones to kill Smith.
We might also offer more indirect considerations. Consider the following example:
18
Joness Ambulance: Jones is rushing to the hospital to save the lives of five, whom
he came across poisoned near his property. A person has innocently fallen unconscious across Joness private roadway, however. The roadway is the only path to the
hospital. So, should Jones stop, the five in the rear of the ambulance will die.
Should Jones proceed, the one lying unconscious in the road will be killed.
Most people will think here that Jones may not drive over Smith. Jones must stop, even
at the cost of the five lives in the rear of the ambulance.
If Jones may not drive over an innocent Smith, laid across Joness private roadway,
even to save the lives of five people, then surely Jones may not drive over Smith in
a comparable case in which one life is at stake, even if it is Joness own life at stake.
Most people would judge that Jones acts permissibly in this case. Most
people would agree that Smith would not be permitted to kill Jones, if
that were necessary to save his life, either. Imagine, for example, that
Smith and Jones were arranged such that Smiths shooting Jones would
cause Jones to fall out of his shelter, knocking the nearby antidote to
Smith. Smith would not be permitted to shoot.
Consider a second example:
Shelter, Smiths Antidote: A poison gas has been released from a volcanic
vent on the valley floor. Jones has found some natural shelter, where he is
safe from the poison. From his shelter, he can see Smith some distance
away. Smith, weak from poison, is struggling to reach his own antidote,which sits on the ground nearby. Jones can help Smith reach his
antidote only by leaving the shelter. Smith and Jones differ in their sensitivity to the poison in the following way, however. If Jones leaves his shelter now, his exposure would be too serious for the antidote to save him.
He must wait for the poison to dissipate a bit before the antidote will be
So, even if we assume Smith is innocently in a place he has no right to be in Drive, the
moral differences between Drive and Joness Antidote remain.
There may be an important qualification to be made here. If Smith knowingly places
himself in a position he has no right to be, many people will be sympathetic to the
claim that Jones may then drive over Smith to get to the life-saving antidote.
302
timothy hall
effective, whereas Smith could use the antidote now. So, Jones does not
leave the shelter until the poison dissipates enough for it to be safe, and
Smith dies.
19
There is a complication in Shelter, Smiths Antidote. It seems that Jones requires
that Smith be dead for the following reason. If Smith were alive, Jones would have no
right to use the antidote in preference to Smith. So, because Smiths being dead is a
necessary element in Joness plan, Jones must intend that Smith die. This intention
marks a difference between Shelter, Smiths Antidote and Smiths Antidote.
Even with the intention asserted by this objection, the argument for the moral
difference between denials and standard allowings of harm goes through. For the
intention that Smith die should, if anything, count against the permissibility of letting
Smith die. It is still permissible for Jones to let Smith die in Shelter, Smiths Antidote,
however. Jones need not surrender his own life to save Smiths. So, the difference in
permissibility between Smiths Antidote and Shelter, Smiths Antidote remains.
20
In some cases, the reasons in common between a denial case and a standard
doing or allowing might dominate. For example, an impermissible allowing of harm
might be impermissible because it is an easy rescue case. Denying resources in a comparable case might be impermissible, and largely for the same reason: aid would be low
cost to the benefactor but prevent great harm to a beneficiary.
21
The remarks below can be repeated for other elements of the moral status of a
behavior, such as whether the behavior licenses violent defense from a victim, whether
an agent is liable to compensate afterward, and the degree of wrongfulness or blame.
304
timothy hall
IV
I do not, however, necessarily assert the range of natural rights to property that philosophers like Locke or Nozick assert. See Robert Nozick, Anarchy, State, and Utopia
(New York: Basic Books, 1974), pp. 17482; John Locke, The Second Treatise of
Government, II:v:2551.
306
timothy hall
This medicine is the only way either sort of blindness can be treated.
The medicine is scarce, however, and can be employed to prevent blindness in the one group or the other, but not both.
Most non-consequentialists would forbid this way of realizing equality,as it would involve the violation of the rights of the healthy subclasswho would be made blind in one eye. Importantly, the fact that
the case involves a standard doing of harm, as opposed to a denial of
resources, is part of the explanation for the rights violation that occurs
in this case.
Similar points might be made in regard to an alternative, and widely
endorsed, standard of distributive justice: the assurance of a social
minimum. Should a standard doing of harm be necessary to realize a
minimal level of welfare, again as in a case like Blinding Action, that fact
would tell against realizing the minimum. No such constraints would
be recognized if resources had to be denied to some in order for a social
minimum to be realized.
So, the two features of non-consequentialism taken together providea strong case for a moral difference between standard doings and
allowings of harm on the one hand and denying resources on the other.
The fact that the personal rights tell against standard doings of harm
undertaken in service of equality, or some social minimum, but not
against the denial behaviors involved in a distribution of non-person
resources, indicates a moral difference between standard doings of
harm and denials of resources. Similarly, the fact that a person is permitted to refuse sacrifices imposed by standard doings of harm, and
thus allow another person to be harmed, even at the cost of equality or
some social minimum, indicates a moral difference between standard
allowings of harm and denials of resources.
V
I turn now to six objections to the two arguments just presented.
1. One might object that the second argument does not show that
denying resources is morally distinct from standard doings of harm,
despite differing assessments of cases like Two Groups of Potentially
Blind People and Blinding Action. In Two Groups of Potentially Blind
People, the better placed will appropriate the medicine, or the worse
placed will do so, or neither group will appropriate. If neither group
appropriates, then each group will suffer the blindness to which they
are susceptible. So, if either party appropriates, the other party will be
no worse off than if no one had appropriated. In these conditions, the
objection runs, it is permissible for the group whose denial would be
less bad than to deny the resource. It is permissible, the objection
claims, even if denials are morally comparable to standard doings of
harm.
To see the point, we might consider an example of a standard doing
of harm:
Falling People: Two people are falling toward one another, with a net
below them both. If each person does nothing, then they will collide and
both will be killed. Each person has a ray gun with him, with which he
might disintegrate the other person.24
24
Michael Otsuka, Killing the Innocent in Self-Defense, Philosophy and Public
Affairs 23 (1994), pp. 7494.
308
timothy hall
In this example, many people will agree that it would be permissible for
each person to try to disintegrate the other. Disintegration is plainly a
standard doing of harm, however. Since the only alternative to one person d
isintegrating the other is both people dying anyway, if one man
kills the other the dead man is no worse off than if neither had acted.
It is therefore permissible for each to try to disintegrate the other.
In response, we might note the moral differences between the appropriation of resources and a Falling People-type example. In Falling
People it seems as if the people still retain their rights against being
attacked, despite the permission of each to try to disintegrate the other.
It seems best to say that the person who is disintegrated has her right
infringed by the other person, rather than that she has no right at all.25
In the case of justified appropriation of resources, however, as in Two
Groups of Potentially Blind People, the people to whom resources are
not justly distributed are not naturally described as having their right
to the resource infringed by its denial. The principles of distribution
to which one appeals settle precisely the question of whether they have
a right.
2. One might also complain that the case made above by the second
argument is insufficiently general. I considered only two sorts of principles by which a non-consequentialist might decide the allotment
of non-person resources: equality, and the guarantee of a minimum
level of welfare. There have been other principles defended by other
authors, principles that are not so teleologically oriented toward welfare goals. For example, the principles of justice in distribution advocated by Robert Nozick are famously unconcerned with welfare
targets.26 Perhaps the case advanced above would not show that nonconsequentialists like Nozick are committed to denying resources
being morally distinct from standard doings of harm or standard
allowings of harm.
25
I understand by a justifiable infringement of a right a situation in which one
might permissibly proceed with a behavior that a right ordinary forbids, but in which
some moral residue of the right remains. In some cases, as in taking the stores someone
has left in an empty cabin in an emergency, the residue of the right might entail compensation be paid to the right-holder. In Falling People, the residue of the right might
tell in that respectperhaps the survivor should try to make good to some degree to
the surviving dependents of the other person. Perhaps the right tells in a more subtle
way, via agent-regret not appropriate in a denial of resources case.
26
Nozicks view can serve as a stand in for certain interpretations of Locke, as well.
Nozick, Anarchy, State, and Utopia, pp. 17482.
310
timothy hall
Many people will respond to this example by claiming that what Jones
did seems impermissible, and seems impermissible, furthermore,
despite his right to the resource involved, his patch. The prospective
victims of this behavior would have been able to use violence to prevent Jones from removing the patch, most people will agree. The case
thus seems like it is morally on a par with a standard doing of harm.
In response, one might deny that Jones has a right to the patch. Even
if Joness placing of the patch is consistent with his retaining legal ownership over it, legal ownership is not the last word on the moral rights
to the resources involved. One can argue that the vital importance of
the patch in place implies that others protected by it have a right that it
remain. One can argue that they come to have such a right through the
moral equivalent of eminent domain appropriation of legal property
a case made more plausible in this case by Joness initial provision of
the patch.
Evidence for this view can be found in considering a case in which it
was clearly understood by Jones and by the protected parties that Jones
retained a right to the patch. Suppose, for example, that Jones provided
the patch only on the condition that, should he need it for an emergency himself, he would be permitted to remove the patch at a moments
notice. In such a case, Jones removing the patch does not seem on a par
with a standard killing. It should not be thought that this non-parity is
owing to the agreement alone. It does not seem to be true that this contract is on a par with a contract that Jones be permitted to shoot other
parties if that is necessary for him. That latter contract would be much
more controversial among non-consequentialists, perhaps because of
the belief that the rights to life of the other parties are inalienable. There
will not be such a principled objection to other parties recognizing
Joness right to a patch, even when the patch provides vital assistance.
On the other hand, one might argue that, owing to the longstanding
presence of the patch, Jones creates a threat, rather than removing protection against one. Intuitions about this suggestion will vary. The metaphysics of threats is a notoriously difficult subject. A case like this
might be a hard case for a view on offer. The fact that there are hard
cases for the determination of a threat, however, is not itself an objection to the claim that the presence of an independent threat is morally
significant. That there is, in any event, a metaphysical distinction
between creating and not creating threats seems no less plausible than
the much-attacked claim that there is a distinction between doing and
allowing harm that animates the received view.
A related complaint is that there are cases that appear to fall under
my characterization of denials of resource, but seem (also) to be paradigmatic cases of doing harm. For example, if one removes a manhole
cover from the patch of road over which Smith will walk, one might
claim, one harms Smith when he falls down the hole. For purposes of
moral explanation, Smiths right against being harmed is in question
here, not his right to the use of the manhole.
I do not see it as an embarrassment, however, that denials of resources
are much more prevalent behaviors than one would expect from considering only the withdrawal of life support or other unusual behaviors. The strong intuition that removing a manhole is on a par with
striking someone with a club or otherwise causing some threat to harm
the person may be a function of the strength with which we think a
person is entitled to use resources like manholes in public areas like
streets to protect himself from an injurious fall. Note that our intuitions would be different if we were considering two shipwrecked men
who contend over one life preserver, and we were also sure that one of
the two men had a right to the life preserver. There, as the discussion in
section III made clear, we would not view the rightful owners denying
the life preserver as on a par with initiating a threat against the other
person. Were we to imagine, then, some extremely unusual situation in
which the manhole were owned by one person and not another, and in
which there was a conflict of interest like the life preserver example, it
is not at all clear that removing the manhole would be on a par with a
standard doing of harm.
4. One might also object that both arguments above do not address
denials of resources in which the resource denied is the agent herself.
For example, consider again Thomsons Violinist. In this case, the denial
seems permissible, and it seems permissible in large part because of
the right Jones has to his body. This is not a right that most nonconsequentialists will recognize as distributed, however, that is
bestowed in accord with principles of just distribution of resources
within a society. It is a right Jones has in virtue of being the person
occupying his body.
That is the same right to which it would seem appropriate to appeal
in cases of standard allowings of harm. If, for example, Jones were
called by the hospital and informed of the violinists need, rather than
being kidnapped, Jones would be permitted to refuse to connect himself, and would be so permitted because the right he has to his own
body entails a permission to refrain from making bodily sacrifices on
312
timothy hall
314
timothy hall
30
One might worry that taking an unowned resource from someone who already
possesses it is on a moral par with a standard doing of harm, not because of view (1),
first come, first served, but because the denial itself has a default status on a par with a
standard doing of harm. One reason to reject this view is precisely because different
assumptions as to rights to the resource will affect the permissibility of denial behaviors in ways that standard doings of harm are not affected.
he has a right and to which the victim has no right allows harm; an
agent who denies a resource to which he has no right and to which the
victim has a right does harm. Consider a denial-doing of harm. On this
view, the denial is a doing of harm because the agent lacks a right to the
resource denied and the victim has a right. On such a view, the fact that
the denial is a doing of harm might itself be a morally significant fact.
One can readily accept that a doing of harm that is determined as a
doing of harm by a lack of a right to a resource could itself be salient in
the explanation of the moral status of the behavior.
Nevertheless, as I argued in section II, there would not be a close
similarity between the sort of doing and allowing that occur in denial
cases and the sort that occur in standard cases. So, generality of moral
explanation across all doings and allowings in terms of doing and
allowing harm will be lost.
The preceding discussion has taken place at some remove from practical moral problems. Nevertheless, the distinctness of denying
resources bears in significant ways on practical moral questions. I will
give two examples.
The first is the withdrawal of life support in medical contexts. One
much-discussed issue is the agency involved in the withdrawal of
life support, in particular whether it is active or passive, doing or allowing. On the view I have laid out, at least some moral questions that
were thought to be questions of the significance of agency are questions
of just claims to property. For example, on my view, a doctor will be
permitted, absent reasons of charity, to remove life support when she
has a claim to this support. Establishing such a claim will involve
consideration of a range of factorssuch as her relationship to the hospital in which the equipment resides, her contractual obligations to the
patient, the role of financially interested third parties like insurers or
the statethat play a much more prominent role than would be true of
determining whether a standard act of killing violated a patients right
to life.
Among other things, this result makes the path clearer for the involuntary withdrawal of life support. It is easier, other things equal, to
show that a person in need lacks a claim on a resource because of more
pressing demands of others than it is to show that that person lacks his
right of life because of those demands.
Another issue of practical significance is abortion. On a prominent
line of argument, abortion is analogous to the withdrawal of aid. The
example Thomsons Violinist above is, of course, famously advanced to
316
timothy hall
argue that even if a fetus had a right to life many abortions would be
permissible.
The moral distinctness of denying resources casts doubt on any
direct inference from the permissibility of withdrawal behavior in an
example like Thomsons Violinist to the permissibility of initiating
threats, such as the mechanical or chemical attacks of most abortions.
While withdrawal behaviors can be justified by an appeal to the rights
the benefactor has to the resource withdrawnher bodyjustifying
the initiation of threats against the fetus requires overcoming the fetus
assumed right to life. The burden involved in the latter might be significantly greater than in the former.
318
s. matthew liao
which the wrongness of killing beings who are above a certain threshold of respect does not vary according to their strengths of timerelative interests, other things being equal.7 McMahan proposes that to
be above this threshold of respect, one must have a certain minimum
level of psychological capacities such as autonomy, which presupposes
self-consciousness and some degree of rationality.8 But he does not say
whether having this minimum level of psychological capacities is
equivalent to having a certain minimum level or strength of timerelative interests. The assumption that having a minimum level of psychological capacities is equivalent to having a minimum level of timerelative interests implies that an indi-vidual with a minimum level of
psychological capacities will always have stronger time-relative interests than a being who has less than the minimum level of psychological
capacities. This means, for example, that an elderly human person who
has the minimum level of psychological capacities but who only has a
few minutes to live will still have stronger time-relative interests than a
normal healthy animal, e.g. a dog, who has less than the minimum
level of psychological capacities. Independent of our discussion, this
assumption and its implication seem to me to be correct because
although the dogs future life will contain vastly more good, the elderly
human person who will die in a few minutes still has a very strong
time-relative interest in avoiding death in a few minutes, that is, he has
a very strong time-relative interest in living for many more years. It is
true that his interest in continuing to live will be frustrated but it is still
an interest that he has and it seems that it would be stronger than any
interest the dog might have.
In any case, this assumption is important for our purpose because if
the two thresholdsthe minimum level of psychological capacities
and the minimum level of time-relative interestswere not equivalent,
then the Time-Relative Interest Account of the wrongness of killing
would not be providing the justification for the permissibility of abortion. Abortion would be permissible because embryos and foetuses
have less than the minimum level of psychological capacities, and not
on account of the strengths of their time-relative interests.
If the two thresholds were equivalent, then the Time-Relative Interest
of the wrongness of killing could provide a justification for the permissibility of abortion, because the embryos and foetuses having less than
McMahan, Ethics of Killing, p. 276.
McMahan, Ethics of Killing, p. 261.
7
8
320
s. matthew liao
And he writes that if the possession of certain psychological capacities is what relevantly distinguishes us from animals and so is the basis
of respect, the developed fetus must fall outside the scope of the morality of respect, for it clearly does not possess the capacities that distinguish persons from animals.16
Given this, if the killing of animals like dogs is not governed by the
morality of respect, the killing of embryos and foetuses should also not
be. If so, the Time-Relative Interest Account seems to entail that killing
embryos and foetuses via abortion is not as wrong and may be more
permissible.
In this paper, I first argue that if the Time-Relative Interest Account
permits abortion, then it would also permit infanticide. Given the
impermissibility of infanticide, this is a strong reason to reject the
Time-Relative Interest Account. I next consider and reject the proposal
that the Time-Relative Interest Account can at least explain the permissibility of early abortion, even if it may not be able to explain the
322
s. matthew liao
17
Account has this implication, then this is a serious problem for this
account.
McMahans second strategy is to argue that there are differences in
the strength of time-relative interests between a developed foetus and a
newborn that might account for the distinction between abortion and
infanticide. It might be worth noting that I do not think that McMahan
believes in this strategy. He is inclined, I think, towards the first strategy. As he says at one point, I believe, therefore, that common sense
beliefs about the morality of infanticide have to be revised. Because the
newborn differs only slightly in nature and status from the developed
foetus, there is no basis for a radical moral distinction between abortion and infanticide.18 Nevertheless, it is instructive to consider what
McMahan says here or what someone else who also holds this view
could say to see if the Time-Relative Interest Account can avoid this
implication.
McMahan offers three suggestions. First, he points out that a newborn infant is exposed at birth to a large amount of external stimuli,
which impels its mind to operate at a higher level, thereby accelerating
its psychological development.19 As a result, the infants time-relative
interest in continuing to live will be stronger than when it was a
foetus.
However, that a newborn infant is exposed at birth to a large amount
of external stimuli may just be a contingent fact. For, we can imagine
that after birth, an infant is placed and raised in a dark room in
which external stimuli are minimized. Infanticide would not be more
acceptable just because the newborn infant is now less stimulated
psychologically.
But let us suppose that being born inevitably causes the newborn
infants mind to be additionally stimulated. The issue here is whether
the newborn will achieve a level of time-relative interests that is higher
than that of normal animals such as dogs. It should be obvious that
immediately after birth, it is doubtful that a newborn infant would
achieve a level of time-relative interests comparable to that of a normal
animal, let alone a higher level of time-relative interests, even taken
into account the external stimuli that it may receive. If this is right, then
even granting that there may be a slight difference between the infant
and the foetus in their strength of time-relative interests, this would
McMahan, Ethics of Killing, p. 342.
McMahan, Ethics of Killing, p. 343.
18
19
324
s. matthew liao
hardly explain how the Time-Relative Interest Account need not entail
the permissibility of infanticide.
McMahans second suggestion is that at birth a newborn infant
begins actively to form special relations with the parents and others in
ways in which it is not possible for a foetus. According to McMahan,
the special relations may magnify the reason these people have not to
frustrate its time-relative interest in continuing to live.20
First, it is worth mentioning that developing foetuses can also
actively form some special relations with the parents and others. At late
stages of a pregnancy, a foetus may kick back when its foot is touched
and may respond to music and other kinds of sounds from parents and
others. Secondly, when someone is contemplating infanticide, it is
unlikely that they would try to form special bonds with the newborn
infant. If so, in such a case, it is unlikely that there would be much
stronger special relations to the infant after it was born than before it
was born. This means that while special relations may explain why
some people do not kill their newborns, for people who are thinking
about killing their newborn, it does not seem to provide any reason
why they should not do so. Finally, it is hard to see how special relations
at an early stage of infancy could drastically change a newborns timerelative interests. If it does not, then it remains the case that the newborn does not have significantly stronger time-relative interests than
the foetus. If so, the Time-Relative Interest Account still seems to entail
the permissibility of infanticide.
McMahans third suggestion is that the reasons favouring killing a
foetus are often stronger than in the case of infanticide. He points to
Judith Jarvis Thomsons argument that the foetus inside the pregnant
womans body is dependent on the continued use of her body for survival, which can be quite burdensome and invasive to the pregnant
woman.21 Newborns on the other hand exist independently and the
sacrifices they may require from others for their survival may be of a
fundamentally different kind and possibly less burdensome. According
to McMahan, this means that possible justifications for infanticide may
be more limited and substantially weaker than possible justifications
for abortion. If so, according to McMahan, this may be the main reason
why infanticide is generally more objectionable than abortion.
McMahan, Ethics of Killing, p. 343.
J.J. Thomson, A Defense of Abortion, Philosophy and Public Affairs 1 (1971),
pp. 4766, at pp. 4849; McMahan, Ethics of Killing, p. 344.
20
21
This strategy also does not work because we are not just interested in
any distinction between abortion and infanticide. Rather, we are interested in a distinction that would enable one to say that on the TimeRelative Interest Account, while abortion is permissible, infanticide is
not. While McMahans point above offers an explanation why infanticide may be generally more objectionable than abortion, it does not
change the fact that on the Time-Relative Interest Account, infanticide
is still permissible; that is, it is not subject to deontological constraints.
Again, the permissibility of infanticide on the Time-Relative Interest
Account derives from the fact that newborn babies have weaker timerelative interests than normal dogs. Given that it is permissible to kill
dogs because they have less than the minimum level of time-relative
interests, it should be even more permissible to kill infants given that
they have even weaker time-relative interests. The fact that there may
be additional reasons for favouring abortion does not change the fact
that from the perspective of time-relative interests, an early foetus and
an infant are intrinsically similar. So, if it is permissible from the perspective of time-relative interests to have an abortion, then from the
same perspective infanticide should also be permissible.
Hence, McMahan has not defused the charge that the Time-Relative
Interest Account entails that infanticide is permissible. Given the serious implausibility of this implication, one might also question its
implication about the permissibility of abortion.
The Time-Relative Interest Account and Early Abortion
At this point, some might draw a distinction between early and late
abortion, and some might argue that, irrespective of the Time-Relative
Interest Accounts implication for late abortion, it can at least explain
the permissibility of early abortion.22 One way to define early abortion
is as an abortion prior to the fetal brains acquiring the capacity to support consciousness, which occurs some time between the twentieth
and the twenty-eighth week of gestation.23 Since exactly when this
event takes place is indeterminate given our present understanding of
the nature of consciousness, McMahan proposes that we adopt a conservative estimate and stipulate that an early abortion takes place prior
See, e.g., DeGrazia, Identity, Killing, and the Boundaries of Our Existence.
McMahan, Ethics of Killing, p. 267.
22
23
326
s. matthew liao
which twinning is still possible, would be permissible, using the identity argument.
Elsewhere I have argued that the Embodied Mind Account is false,
that we should accept the Organism View, and that on the most plausible Organism View, we exist before the possibility of twinning ends,
namely, at the moment of fertilization.31 I shall not repeat those arguments here, and I shall simply say that on this particular Organism
View, an early abortion would still be killing one of us.
The second, more serious problem with the identity argument is that
if it were valid, it seems that this argument would be sufficient to establish the permissibility of early abortion. If so, it would render the TimeRelative Interest Account unnecessary. Indeed, as I said earlier,
McMahan himself does not defend early abortion using the TimeRelative Interest Account. This is a problem if the aim is to show that
the Time-Relative Interest Account, and not some other theory, can
explain the permissibility of early abortion.
Is there then a way to use the Time-Relative Interest Account to
demonstrate the permissibility of early abortion without appealing to
the identity argument and while avoiding the implication of infanticide? Here is a possibility: It might be pointed out that how exclusive or
inclusive the Time-Relative Interest Account is depends on where one
sets the minimum level of time-relative interests required to be within
the morality of respect. McMahan, as we have seen, sets it at the level at
which animals such as dogs would not be part of the morality of respect.
But one could lower the requirement in order to avoid the implication
that infanticide is permissible. Admittedly, once the minimum level is
lowered so that infanticide is excluded, this revised Time-Relative
Interest Account would no longer imply that late abortion is permissible, since the difference between a newborn and a late term foetus is, as
we have seen, not significant. However, it might nevertheless enable
one to argue that early abortion is permissible, for the following reason:
Even after drastically lowering the level required to be within the
morality of respect, early foetusesdefined as foetuses that have not
yet developed the capacity for consciousnesswill still not meet it,
even if they are one of us, so that an appeal to the identity argument is
not necessary. This is because early foetuses utterly lack any psychological connection with their later selves, and therefore utterly lack any
See S. Matthew Liao, The Organism View Defended, The Monist 89.3 (2006).
31
328
s. matthew liao
35
330
s. matthew liao
38
If someone finds this argument unpersuasive, there is still the first argument
above.
39
For a good discussion of these two cases along the lines I am suggesting, see
McMahan, Ethics of Killing, pp. 44350.
40
See also McMahan, Ethics of Killing, pp. 42650.
41
I defend this in Liao, The Organism View Defended.
332
s. matthew liao
the early/late foetus case I have been discussing.42 Also, given this
assumption, the not uncontroversial view that we die when our higher
brain, i.e. our cerebrum, dies would be ruled out.)43 The important
question for us here is that given that the PVS and the deep coma
patient are both one of us, but given that the PVS patient utterly lacks
any time-relative interest, while the deep coma patient has just a bit of
time-relative interests left, does this difference justify a very different
treatment of the two such that, for example, it would be permissible to
kill the PVS patient while it would not be permissible to kill the deep
coma patient?
My conjecture is that there should be no difference morally in the
way one treats the two cases. Indeed, some people believe that it would
be wrong to kill the deep coma patient, and that it might also be wrong
to allow the deep coma patient to die, unless the patient has previously
expressed an autonomous preference to be allowed to die in such circumstances.44 If they believe this, and if they believe that both the PVS
and the deep coma patient are one of us, then they should also believe
that it is wrong to kill the PVS patient, and that it might also be wrong
to allow the PVS patient to die, unless the patient has previously
expressed an autonomous preference to be allowed to die in such
circumstances.
Others such as Peter Singer and James Rachels have argued that the
fact that the PVS patient and the deep coma patient are alive does not
in itself tell us whether it is wrong to take their lives.45 Since they believe
that it is permissible to kill the PVS patient in certain circumstances,
they should also believe that it is permissible to kill a deep coma patient
in similar circumstances. In fact, Singer has argued that there is no
It should be noted that the difference between the PVS and the deep coma patient
may be somewhat greater than the difference between the early fetus and the late fetus.
This is because while both the PVS patient and the early fetus utterly lack any timerelative interests, the cerebrum of the deep coma patient is typically developed while
the cerebrum of the late fetus is typically not as developed. This suggests that the difference between the early and late fetus is even less than the difference between the PVS
and the deep coma patient.
43
J. Korein, Ontogenesis of the Brain in the Human Organism: Definitions of Life
and Death of the Human Being and Person, Advances in Bioethics 2 (1997), pp. 174.
Defenders of the higher-brain criterion include M.B. Green and D. Wikler, Brain
Death and Personal Identity, Philosophy and Public Affairs 9 (1980), pp. 10533; R.M.
Veatch, Whole-brain, Neocortical, and Higher Brain Related Concepts, in R. Zaner
(ed.), Death: Beyond Whole-Brain Criteria (Dordrecht: Kluwer, 1988), pp. 17186.
44
McMahan, Ethics of Killing, p. 443.
45
Singer, Rethinking Life and Death, p. 207; Rachels, The End of Life, pp. 4243.
42
46
47
336
s. matthew liao
Many philosophers believe that when we assess who can be a rightholder we must meet what might be called the Species Neutrality
Requirement.4 The Species Neutrality Requirement says that an adequate account of rightholding must provide some criterion for rightholding that in principle does not exclude any species and where the
criterion can be assessed through some objective, empirical method.
The reason for the latter condition is that there does not seem to be an
a priori way of knowing who the rightholders are. For example, if we
meet some alien beings and we want to know whether they are rightholders, it seems that we would not be able to know a priori whether
they are rightholders. If so, the only way to find out is to investigate
empirically what attributes they have and to consider if these attributes
are relevant enough to make them rightholders.
However, supposing that the Species Neutrality Requirement is correct, there does not seem to be a relevant empirical attribute that would
apply to all human beings. The most plausible attributes such as actual
sentience and actual agency do not apply to all human beings. For
example, some human beings such as anencephalic children and comatose persons lack actual sentience, and many human beings including
newborn infants lack actual agency. These human beings would not be
rightholders on these accounts.
Given this, those who wish to defend the claim that all human beings
are rightholders typically reject the Species Neutrality Requirement.
For example, some argue that all human beings are rightholders because
they have intrinsic worth or because they have dignity.5 Neither notion
is an attribute that one can empirically identify and assess. Others
assert that it is just self-evident that all human beings are rightholders,
and that it is not necessary to find out the particular attributes that
make them so.6 Although justification for any moral principle must end
beings as well as some other animals are rightholders. Such a position would not necessarily be speciesist. In fact, the account I shall develop may be an example of such a
position. As far as I am aware, such a position has not been advanced in the literature.
4
J. Feinberg and B. B. Levenbook, Abortion, in T. Regan (ed.), Matters of Life and
Death (New York: McGraw-Hill, 1993), pp. 195234.
5
G. Vlastos, Justice and Equality, in Jeremy Waldron (ed.), Theories of Rights
(Oxford: Oxford University Press, 1984), pp. 4176; R. Dworkin, Lifes Dominion (New
York: Alfred A.Knopf, 1993).
6
H. J. McCloskey, Respect for Human Moral Rights Versus Maximizing Good,
in R. G. Frey (ed.), Utility and Rights (Oxford: Basil Blackwell, 1985), pp. 12136;
R. Nozick, About Mammals and People, New York Times Book Review, 27 November,
II, pp. 2930; Williams, The Human Prejudice.
337
at some point, if being human is a sufficient but not necessary condition for rightholding, it seems problematic that this approach is not
able to help us determine whether the beings with which we presently
live and the beings we might encounter in the future are rightholders.
If being human is a necessary condition for rightholding, it seems that
this approach would have the seemingly counterintuitive implication
that all non-human beings that we presently know as well as those we
might meet in the future cannot be rightholders.
Still others have argued that all human beings have rights, not in
virtue of the actual attributes they possess, but in virtue of belonging to
the kind of beings that typically have the relevant attributes for rightholding. For example, John Finnis says that to be a person is to belong
to a kind of being characterized by rational (self-conscious, intelligent)
nature.7 Scanlon says that the class of beings whom it is possible to
wrong will include at least all those beings who are of a kind that is
normally capable of judgment-sensitive attitudes.8 Indeed, Scanlon
believes that severely mentally handicapped human beings can be
wronged, even though they themselves do not and will not have the
capacity to understand or weigh justifications.9 For conveniences sake,
we can call this account the species norm account.10
The species norm account also fails the Species Neutrality Require
ment. To be sure, some members of a given species would have the
7
J. Finnis, A Philosophical Case against Euthanasia, in J. Keown (ed.), Euthanasia
Examined: Ethical, Clinical, and Legal Perspectives (Cambridge: Cambridge University
Press, 1995), p. 48. Readers will notice that Finnis is talking about persons, whereas
I have been discussing rightholders. In this paper, I understand the terms rightholding and personhood interchangeably. Some people who equate rightholding with
personhood also hold a particular conception of personhood, namely, all and only
those who are actual moral agents are persons. Consequently, they believe that all and
only those who are actual moral agents are rightholders. I do not hold this view. Others
equate personhood with actual moral agency, but not with rightholding. As many of
them also believe that some non-moral agents, e.g. animals, can also be rightholders,
these people argue that personhood is only a sufficient but not a necessary condition
for rightholding. Since my equating rightholding with personhood is a stipulation,
nothing substantive follows from it. If animals are rightholders, then, as I understand
these terms, they would be persons, even if they are not actual moral agents.
8
Scanlon, What We Owe to Each Other, p. 186.
9
Scanlon, What We Owe To Each Other, p. 185.
10
McMahan used the term species norm account to describe something different,namely, a view about the nature of fortune and misfortune. This term is useful,
though, for capturing the ideas advanced by writers such as Finnis and Scanlon. See
J. McMahan, The Ethics of Killing: Problems at the Margins of Life (Oxford: Oxford
University Press, 2002), pp. 14649, 209217.
338
s. matthew liao
339
and that at the same time appears to allow all human beings to be
rightholders. I call this the genetic basis for moral agency account of
rightholding.
Before we begin, it is worth noting that every plausible account of
human moral status comes with a certain theoretical baggage. That
is, each account will face difficult issues particular to it that it must
address. For example, on a sentience account of moral status, according to which beings that have the same amount of sentience have the
same moral status, it seems that one would have to address the issue
of how to treat a human being and a non-human animal, when both
have the same amount of sentience.13 Similarly, an account of moral
status, according to which all human beings are rightholders, will
imply that fetuses and embryos are rightholders since they are human
beings, and this will require discussions about issues such as abortion and embryonic stem cell research. Typically, plausible accounts
of moral status will have responses to these issues. For example, suppose one believes that all human beings are rightholders and that
abortion is permissible. One can follow Judith Jarvis Thomson and
hold the view that even if fetuses were rightholders, abortion would
still be permissible.14 It is beyond the scope of this paper to discuss all
the theoretical issues that can arise from the belief that all human
beings are rightholders. My aim here is therefore to set out clearly an
account that supports this belief, but not to defend this account
against all possible objections. Nevertheless, I shall explain how the
genetic basis for moral agency account is different from a potentiality
account and how it is preferable to an actual moral agency account of
human moral status.
2. The Genetic Basis for Moral Agency Account of Rightholding
My proposal is as follows: all human beings are rightholders because
they all have the genetic basis for moral agency; and it seems that having this genetic basis is sufficient for one to be a rightholder.
See, e.g., Singer, Practical Ethics.
J.J. Thomson, A Defense of Abortion, Philosophy and Public Affairs 1 (1971),
pp. 4766 (4849). In S. Matthew Liao, Rescuing Human Embryonic Stem Cell
Research: The Blastocyst Transfer Method, American Journal of Bioethics 5.6 (2005),
pp. 816, I argue that even if embryos were rightholders, there are still permissible
ways of pursuing embryonic stem cell research.
13
14
340
s. matthew liao
16
341
genes that make up an attribute and genes that undermine the development of an attribute, a distinction that could help us further to narrow
the set of genes that is necessary and sufficient for the genetic basis for
moral agency.
It should be mentioned that some people are hostile to the idea of a
genetic basis for human behavioral traits. A reason for this hostility is
due in part to the fact that, historically, there have been racist and sexist
attempts to show that human beings of a certain race or gender have
the genetic basis for higher intelligence than human beings of another
race or gender; and these attempts typically fail to consider seriously
the role that non-genetic factors play in the development of intelligence. Hence, by association, these people might also be hostile to the
idea of a genetic basis for moral agency.
However, the genetic basis for moral agency, as I understand it, gives
rise to a capacity and not a behavioral trait. This means that the fact
that human beings have the genetic basis for moral agency does not
mean that they will act morally. Also, nothing I have said precludes the
idea that non-genetic factors are also necessary for an adequate development of moral agency. Indeed, the idea of a genetic basis for moral
agency is compatible with the idea that much of the complexity needed
for the development of moral agency is located in the developmental
environment, a developmental resource that has been specifically
adapted to afford the development of moral agency (in a niche
construction-type way).
Here it is worth making two further points regarding the genetic
basis for moral agency. First, the idea of a genetic basis does not mean
that there is only one way this genetic basis can be sequenced or realized. In fact, this genetic basis could be multiply realizable. That is, it
could be the case that in certain environments, certain genes, A, B, C,
might be the genetic basis for moral agency for a particular being,
while in certain other environments, genes D,E,F would be the genetic
basis for moral agency in the same being.17
Secondly, to have the genetic basis for a certain attribute, the genes
that make up that attribute must be activated and be coordinating
with each other in an appropriate way. A being does not have the
genetic basis for a certain attribute if a being just possesses somewhere in its genome the genes that could make up the attribute, but
I thank David Wasserman for this point.
17
342
s. matthew liao
these genes are either not activated or are scrambled in such a way
that they do not coordinate with each other in an appropriate way.
To support the point about coordination, consider the following:
Suppose there is a book containing many random words, which if put
together in the right way, would result in a Shakespeare book. That
book would not be a Shakespeare book just because it contains the
correct words; those words must be organized in the right way.
How do we know that all human beings have the genetic basis for
moral agency? We know that all normal functioning human beings and
those at the beginning of life (e.g. infants, young children, fetuses) who
will develop normally have this genetic basis, since they exercise moral
agency or will exercise it; and after conception, the genetic codes of a
human being do not change very much, if at all. We also know that
most comatose human beings have this genetic basis, because they
have exercised moral agency. Most of the causes of anencephaly are
not genetic but are caused by environmental factors such as folic acid
deficiency.18 So we can assume that many of those with anencephaly
also have this genetic basis.19 Moreover, those with mild mental retardation, such as children with Down syndrome, typically exhibit some
moral agency. This suggests that they also have the genetic basis for
moral agency.20
Finally, to see how those severely defective human beings whose
conditions are the result of genetic defects rather than environmental
factors would also have this genetic basis, it is useful to distinguish
18
Other environmental factors include undiagnosed diabetes; hypervitaminosis A;
high temperatures of 102 degrees or higher for more than five hours; anticonvulsant
medication, especially valporic acid (valporate); or environmental/chemical exposure.
See, e.g., Center for Disease Control and Prevention Effectiveness in Disease and
Injury Prevention: Use of Folic Acid for Prevention of Spina Bifida and Other Neural
Tube Defects, 19831991, Morbidity and Mortality Weekly Report 40 (1991), 513516;
MRC Vitamin Study Research Group, Prevention of Neural Tube Defects: Results of
the Medical Research Council Vitamin Study, Lancet 338 (1991), pp. 13137.
19
At the moment, it is an open question whether anencephaly even has a genetic
cause. Scientists know that some genetic disorders such as Waardenburg syndrome
seem to correlate with a higher incidence of anencephaly. But there is no reason at the
moment to believe that anencephaly is the result of a defect of the genes that make up
moral agency.
20
About 34 percent of the mentally retarded population is severely retarded, and
only 12 percent of the mentally retarded population is classified as profoundly
retarded. See, e.g., American Psychiatric Association, Diagnostic and Statistical Manual
of Mental Disorders (Washington, DC: American Psychiatric Press, Inc., 4th edn,
1994); M.L. Batshaw and B.K. Shapiro, Mental Retardation, in M.L. Batshaw (ed.),
Children with Disabilities (Baltimore: Paul H. Brookes, 4th edn, 1997).
343
344
s. matthew liao
It might be necessary to concede that there is a theoretical possibilitythat ahuman being could lack the genetic basis for moral agency,
even if all present cases are not ones in which human beings lack this
basis.23 In particular, owingto advances in genetic engineering, someone might be able to create such an individual artificially. Of course, it
could be questioned whether such an individual would be a human
being at all. This raises the difficult issue of what makes a being human,
and whether the genetic basis for moral agency is an essential
human property. At present, I do not know of a good answer to this
question. Consequently, it should be admitted at least for now that it is
theoretically possible that a human being could lack the genetic basis for
moral agency. Still, we can conclude that for practical purposes, virtually
all living human beings we are likely to encounter will have the genetic
basis for moral agency.
The claim that the genetic basis for moral agency is sufficient for
rightholding is attractive for several reasons. First, as this account practically supports the widely held intuition that all human beings are
rightholders, for many, this would be a reason in favor of it.
Secondly, the genetic basis for moral agency is an identifiable, actual,
physical attribute. This means that this account meets one of the primary conditions of the Species Neutrality Requirement. It also means
that this account avoids speciesism. Indeed, if we were to learn that
chimpanzees or some other animals have the genetic basis for moral
agency, then they would be rightholders.
Thirdly, this account captures what is intuitively appealing (at least
to some) about the species norm account, namely, it too is motivated
by the thought that the kind of being that is typically characterized by
moral agency should be a rightholder. However, the genetic basis for
the moral agency account offers a more adequate interpretation of this
claim, because on this account, one actually possesses an identifiable,
actual, physical attribute and is not just a member of a group that possesses this attribute. In fact, this account can handle the kind of objection posed by McMahan against the species norm account. If a
Superchimp has a rational nature as a result of gene therapy, this
account can accommodate this case by locating this nature in the
345
genetic makeup of the Superchimp. Or, if Superchimps come to outnumber the normal chimpanzees, this account can explain how the
two groups can be treated differently, if, for example, the Superchimps
have the genetic basis for rational nature while the normal chimpanzees do not.
Finally, since the genetic basis for moral agency is only a sufficient
condition for rightholding, it avoids the intuitive cost of denying the
status of rightholding to those non-human animals or other beings
who may plausibly qualify as rightholders but who may not have the
genetic basis for moral agency. For example, even if a dog does not have
the genetic basis for moral agency, the dog could still be a rightholder
on other grounds. Or, suppose it were possible to genetically engineer
a being to lack just one gene for the genetic basis for moral agency and
still to be human; although this human being would not have the
genetic basis for moral agency, this human being could still be a rightholder on other grounds.
Here it is worth mentioning that it is possible that some beings,
e.g. some alien being or some super artificial intelligent being, could be
made up of non-genetic, that is, non-carbon-based, isomorphic material, and still possess something functionally similar to the genetic
basis for moral agency. In my view, they would also be rightholders,
given that they have the physical basis for the development of moral
agency even though they do not have the genetic basis for moral agency.
Hence, a more precise name for this account should be the physical
basis for the development of moral agency account of rightholding.
However, since most of the living beings we know are genetic carbonbased life forms, to keep things simple, I shall continue to refer to this
account as the genetic basis for moral agency account.
The intuitive appeal of the genetic basis for moral agency account of
rightholding should be easy enough to grasp. Some people might of
course not share these intuitions and they might demand that one produces an independent argument as to why one should believe this
account. This demand may not be entirely fair though, since, as far as
I am aware, those who advance other criteria for rightholding typically
offer no non-circular, independent arguments for why their preferred
criterion is relevant for rightholding. To give just one example, suppose
one holds the view that if X has actual sentience, then X is a rightholder. It might be asked, why is this so? Asserting that pain is bad
does not seem to be providing an independent argument for this
account. Arguably, this is just a circular way of restating that actual
346
s. matthew liao
sentience is relevant. In any case, since my aim here to set out clearly
the genetic basis for the moral agency account and not to argue that it
is the correct account of moral status, I shall leave this matter here.
What I shall try to do in the remaining sections of the paper is to consider some possible objections to this account and to respond to these
objections on its behalf.
3. Just a Potentiality Account?
To start, some might wonder if the genetic basis for moral agency
account is just a potentiality account in disguise. That is, it might be
thought that having the genetic basis for moral agency is just having
the potential for moral agency. As such, so the argument goes, the
genetic basis for the moral agency account contributes nothing new to
the debate and inherits all the problems that have been attributed to the
potentiality account.
First, even if the genetic basis for the moral agency account were just
a potentiality account, this may not be a problem, because the arguments against the potentiality account are not conclusive. I do not want
to defend the potentiality account here, but let me present two oftrehearsed arguments against the potential account. One goes as
follows:
1. According to the potentiality account, if X is a potential F, then X
has the same rights and interests as an actual F, where F could be a
rightholder, human being, person, and so on.
2.However, potential Fs typically do not have the same rights and
interests as actual Fs. For example, a potential president does not
have the same rights and privileges as an actual president.
3.Therefore, the potentiality account is mistaken.
Clearly, proponents of the potentiality account would deny premise
1. No such proponent would hold the view that a potential rightholder has the same rights and interests as an actual rightholder.
A more plausible interpretation of the potentiality account is something like If X has the potential forV, where V denotes attributes such
as moral agency, sentience, and so on, then X is an F, where F could be
a rightholder, human being, person, and so on. For example, A.I.
Melden has proposed that if X has the potential for moral agency, then
347
348
s. matthew liao
349
But actual moral agency cannot be the sole ground for rightholding.
The reason is that if rightholding served any function at all, one would
be the following:
If and when the rightholders interest is in conflict with the same kind of
interest, that is, with the comparable interest, of a non-rightholder, the
rightholders interest should prevail.
A corollary of this is that if one were to give the interest of a nonrightholder priority over the comparable interest of a rightholder, then
one would be acting wrongly.
For example, normal adult human beings are typically regarded as
rightholders, whereas normal adult turtles are not. (Those who
believe that all animals have rights may be unhappy with this
example. If so, I suggest that they substitute the turtle with whatever
they would regard as a non-rightholder. The general point would
remain valid.) Suppose this is correct, and suppose rightholding has
the function I suggested, this means that if and when the interest of a
normal adult human being conflicts with the comparable interest of
a normal adult turtle, one ought to give the interest of the human
being more weight. Hence, if a normal adult turtle and a normal adult
human being both require rescue (suppose both are crossing the
street and are in danger of being hit by oncoming traffic), and suppose that one can only save one of them, then one ought to save the
human being, because the turtle is not a rightholder while the human
being is. If one did not do this, then one would be acting wrongly.
This does not mean that one does not have any duties at all regarding
non-rightholders. For example, if a turtle requires rescue and it
would cost one little effort to save it, it seems that one would have a
duty to save the turtle. It also does not mean that any interest of a
rightholder would be more important than any interest of a nonrightholder. If the turtle requires rescue and it would cost one little
effort to save it, and suppose that I, a rightholder, am having a cup of
tea at the moment, it seems that I may have a duty to save the turtle
even if my tea may become cold.
If actual moral agency is the sole ground for rightholding, though,
adult human beings would be rightholders while human infants would
not be, as the latter do not have actual moral agency. If rightholding has
the function I suggested, it would mean that an adult human beings
interest should be given priority over the comparable interest of an
infant. And, if one were to give an infants interest priority over the
350
s. matthew liao
30
Some might suggest that perhaps infants are given preference because they are
smaller. However, this cannot be the explanation, since if someone were to have a small
turtle, and assuming that turtles are not rightholders, we would not give the turtle
preference over an adult male just because it is smaller.
31
See, e.g., J. Rachels, The End of Life: Euthanasia and Morality (Oxford: Oxford
University Press, 1986), p. 6; Marquis, Why Abortion is Immoral, p. 189.
351
Others might argue that the problem is that the status of rightholding does not have the function I discussed above. They might point out
that sometimes non-rightholders interests are given priority over the
comparable interest of arightholder. For example, suppose to rescue
an adult human being who onlyhas a short period to live, one must
blow up a gigantic planet that contains many wonderful life forms,
none of which are rightholders. One might decide against blowing up
the planet. If so, this seems as if non-rightholders interests can sometimes be given priority over the comparable interest of a rightholder.
However, the argument that the status of rightholding has the function I discussed above need not claim that the interest of a rightholder
can never be outweighed by any amount of non-rightholders interests.
In fact, the rule implies that the comparison should be one to one, that
is, one rightholders interest versus a comparable interest of a nonrightholder; not one to a larger number. On a one-to-one basis, I cannot think of an example where it would be permissible to give preference
to a non-rightholders interest over a comparable interest of a rightholder. At the same time, there are many instances where it seems permissible to save an infants life over that of an adult human being, even
if the infant does not have more to lose.
Still others might suggest that perhaps we save the infants not
because the infants are rightholders but because in doing so we would
be promoting the interest of rightholders, namely, the interest of the
parents. This argument may have some plausibility if the choice is
between saving an infant and the infants parent, and the parent has
given indications that he would want the infant rather than himself to
be saved. But even when the choice is between saving an infant and an
unrelated adult human being, it often still seems permissible to give
preference to the infant. If this is right, this is one reason why actual
moral agency cannot be the sole ground for rightholding.
5. Actual Agency Matters More
Still other people might accept that the genetic basis for moral agency
matters somewhat for moral status, but they might insist that actual
moral agency always matters more, and therefore, only those with
actual moral agency should be rightholders. For example, suppose an
IVF clinic is burning and one could either save an embryo, which has
the genetic basis for moral agency, or a five-year-old child, who has
352
s. matthew liao
actual moral agency.32 It seems that one should save the five-year-old.
If so, this seems to show that actual moral agency matters more.
First, this case does not show that actual moral agency always
matters more. Imagine a variant of this case in which another IVF
clinic is burning, and there is another embryo and another five-yearold child. This time, though, the five-year-old child has inhaled too
much smoke, has temporarily passed out, and will die soon (in a few
days time) and the embryo happens to belong to you. You have been
desperately trying to have a child for years and the doctors tell you that
this time there is a very good chance you will succeed, but that this is
your last chance to have your own biological child. Is it so intuitively
clear that you should save the five-year-old? My intuition is not as clear
in this case. Indeed, I think you may be permitted to save your last
embryo. Note that suppose the alternative is between saving the fiveyear-old and saving, for example, your beloved Picasso painting.
Clearly, you should save the five-year-old. This suggests that the reason
why you may be permitted to save the embryo in this case is not just
because the embryo is your embryo, since the Picasso is also your
Picasso. If this is right, there may be certain circumstances in which
saving someone with the genetic basis for moral agency matters more
than saving someone with actual moral agency.
In any case, it is not necessary to deny that often an actual moral
agent will matter more than a being with just the genetic basis for moral
agency. But this would not show that the two beings have different
moral status. Indeed, when a stranger and ones partner are both
drowning, other things being equal, ones partner typically matters
more than the stranger; but this would not show that the two have different moral status.
Because an embryo typically has a greater future potential than
a grown child, some people might worry instead that the genetic basis
for moral agency account implies that we should always save the
embryo rather than the child. This implication need not follow. First,
just as there could be agent-relative reasons for choosing ones
embryo,there could also be agent-relative reasons for choosing ones
grown child, even if the embryo indeed has a greater future potential.
Secondly, there could also be agent-neutral considerations why one
32
See, e.g., S. Matthew Liao, The Embryo Rescue Case, Theoretical Medicine and
Bioethics 27.2 (2006), pp. 14147.
353
would choose the grown child over the embryo. In particular, it could
be argued that the embryo will have little or no time-relative interests
while the grown child may have very strong time-relative interests.33
To have time-relative interests is to be able to stand in some psychological relations to ones future and past selves. The strength of ones
present time-relative interests depends on how strongly one is psychologically connected to those future and past selves. For example, an
infant will typically have weaker time-relative interests in continuing
to live than a grown adult, since an infant has little or no awareness of
his or her future self. Similarly, the embryo will have little or no timerelative interests, since the embryo will not have the required capacities
to be able to stand in some psychological relations to the embryos
future and past selves. The grown child, on the other hand, may have
very strong time-relative interests, since the child could by then have
fully developed p
sychological capacities. If so, this could provide an
agent-neutral reason to choose the grown child over the embryo,
because although the grown child may have less future potential, the
grown child may have stronger time-relative interests. Still, the fact
that there is a difference in time-relative interests between the embryo
and the grown child need not be taken to mean that there is also a
difference in moral status.34
6. Anencephalic Infants and Embryos Cannot Have Interests
Some people might ask, in what sense can anencephalic infants and
embryos have interests that can ground rights? Following Feinberg,
many people hold that to have an interest is to be interested in something, which requires that one has the capacity to desire and want
things and to be self-aware.35 Given that anencephalic infants and
embryos lack brains, they lack the capacity to desire and want things.
Consequently, on this view of interest, they cannot have interests that
can ground rights.
33
The concept of time-relative interest comes from McMahan, The Ethics of Killing,
pp. 80, 17074, 18388. See also S. Matthew Liao, Time-Relative Interests and
Abortion, Journal of Moral Philosophy 4.2 (2007), pp. 24256.
34
I develop this point in greater detail in Liao, The Embryo Rescue Case.
35
J. Feinberg, The Rights of Animals and Unborn Generations, in Rights, Justice,
and Bounds of Liberty (Princeton, Princeton University Press, 1980), pp. 15984.
354
s. matthew liao
355
356
s. matthew liao
injected these genes into a cabbage, would the cabbage be a rightholder?40 To answer this question, we should inquire whether these
genes are integrated into the cabbage or not. Ifthey are, then the cabbage may no longer be a cabbage, since its nature may have changed.
In such a case, the resulting being may be a rightholder, given that it
has the genetic basis for moral agency. A real-life parallel may be transplanting human genes into a rabbit egg and creating an embryo with
human characteristics.41 Arguably, such an entity may be a rightholder.
On the other hand, suppose that the genes for the genetic basis for
moral agency are not integrated into the cabbage, but instead these
genes just sit inside the cabbage. In such a case, the cabbage may not
have the kind of moral status that a cabbage with integrated genetic
basis for moral agency may have.
8.Conclusion
In this paper, I proposed that the genetic basis for a moral agency
account of rightholding can allow all human beings to be rightholders
without being speciesist. Although I did not try to argue that this
account is the correct account of human moral status, I did argue that
this account is different from a potentiality account and that it is preferable to an actual moral agency account of human moral status. While
it was beyond the scope of this paper to consider all issues arising from
this account, I have laid out some of its particular strengths. By being
able to incorporate the intuition that an appropriate account of rightholding should be species neutral and by being able to explain how all
human beings are rightholders, the genetic basis for the moral agency
account appears to provide a real and coherent alternative to present
accounts of rightholding and therefore deserves further study and to be
part of the debate about the foundations of human moral status.
358
martin peterson
However, according to some authors, a lottery in which each individual has an equal chance of being rescued is unjustified in many
cases, e.g. in a choice between saving either a million people or ten
people.4 A fifty percent risk of not saving a million people is simply
morally unacceptable. In response to this objection, Broome has proposed that the number problem could be resolved by setting up a
weighted lottery, in which the probability that the first group (of m people) is saved is m/m+n, and the probability that the second group (ofn
people) is saved is n/m + n.5
That said, arranging a lottery seems to be entirely out of the question
in a choice between saving a million or ten people, no matter the probabilities. In such an extreme situation, it can be argued that the positive
moral value gained by arranging a lottery is entirely outweighed by the
number of people saved when saving the greatest number. The present
article aims at exploring this claim further. The hypothesis is that
arranging a lottery contributes towards making an act fairer, while
other features, such as non-optimal consequences, contribute towards
making the act worse in other morally relevant respects. Doing the
right thing ultimately amounts to balancing different values against
each other. In what follows, this view will be referred to as the mixed
solution.
The mixed solution has been briefly touched upon by John Broome
and Iwao Hirose, but it has never been extensively researched.6 Contrary
to Broome and Hirose, I argue that the mixed solution can be defended
without a ssuming the possibility of interpersonal value comparisons.
That is, the mixed solution can be accepted without claiming that one
ought to save two lives rather than one because two counts for more
than one. This might look like a rather technical and modest point, but
closer examination reveals that it is not. A p
rominent reason for taking
the number problem seriously is the widespread belief that moral value
cannot be aggregated across different people, not even in principle.
Anyone wishing to combine the claims of different people into a unified, stronger moral claim has to explain how this can be done in a
non-arbitrary way. We are all individuals, and this fact is surely of some
moral importance.
See e.g. Broome, Kamm on Fairness, and Hirose, Aggregation and Numbers.
Broome, Kamm on Fairness.
6
See e.g. Broome, Kamm on Fairness, and Hirose, Aggregation and Numbers.
4
5
359
360
martin peterson
Alternative 1:
Alternative 2:
Alternative 3:
ABC
(saved, dead, dead)
(dead, saved, dead)
(dead, saved, saved)
361
1 and 2 are equally good, and Pareto implies that Alternative 3 is better
than Alternative 2. Therefore, given that the relation better than is
transitive, it follows that Alternative 3 is better than Alternative 1, i.e.,
we should save the greatest number.9
Unfortunately, Hiroses argument only works as long as all alternatives are non-probabilistic. Even though it is true that Alternative 3 is
indeed better than Alternative 1, nothing more can be concluded. In
particular, it does not follow that Alternative 3 is better than any of the
two major alternative positions in the debate over the number problem, the Equal-Chances view and the Proportional-Chances view,
mentioned in section 1. By letting (50% saved; 50% (saved, saved))
denote a lottery in which the probability is 50% that thefirst individual
is saved and 50% that the two others are saved, the Equal-Chances view
and the Proportional-Chances view can be described as in Table 2.
Evidently, Hiroses two principles, Symmetry and Pareto, are not sufficient for deciding which alternative to choose. Since Pareto is only
applicable to deterministic cases, it cannot be used for adjudicating
among probabilistic alternatives. Therefore, Hiroses argument fails to
show that the Greatest Number view (Alternative 3 above) is morally
better than the Equal-Chances view and the Proportional-Chances
view. Consequently, since the latter views are among the most wellestablished ones in the literature, the reasoning suggested by Hirose is
not sufficient for giving a comprehensive analysis of the number problem. It is simply not true that Symmetry and Pareto imply that consequentialists can justify the case for saving the greater number without
combining the goods of separate people.10
Table 2
AB
C
(50% saved; 50% (saved, saved))
(33% saved; 66% (saved, saved))
9
This conclusions has been questioned in the literature; see e.g. Thom Brooks,
Saving the Greatest Number, Logique et Analyse 17778 (2002), pp. 5559.
10
Hirose, Aggregation and Numbers, p. 68.
362
martin peterson
3. The Mixed Solution
Toss a coin
Save the greatest number
1000.5
1001
0
-u 1000
11
The most extensive discussion of this view can be found in Hirose, Aggregation
and Numbers. See also Broome, Kamm on Fairness.
12
Hirose, Aggregation and Numbers.
363
Aggregated unfairness
Toss a coin
Save the greatest number
1.5
2
0
-u 1
364
martin peterson
G(p) =1
pi p j .
2n2 p i =1
j =1
365
I shall refrain from discussing whether or not this is the case. My point
is simply that Hiroses analysis relies on an assumption that Burnham
has shown to be doubtful: a single, fixed value of u cannot be used for
analysing cases involving very different numbers of people.
John Broomes view is to a large extent similar to Hiroses. Broome
argues that, fairness is not everything. Fairness requires tossing a coin,
but just as Ithink the fairness of saving no one is outweighed by the
badness of the result, so I think the fairness of tossing a coin is outweighed by the expected badness of the result.16 I take this to mean that
Broome also accepts that unfairness can be aggregated into a whole.
Or, put slightly differently, that Broome agrees that it makes sense to
multiply u by 1000 when a thousand people are left behind, and so on
and so forth.
That said, the idea that fairness can be aggregated into a whole is
presumably a bit questionable. How can there be such a thing as an
aggregated amount of unfairness? Why should we believe that unfairness can be aggregated interpersonally, i.e. that it makes sense to say
that persons A and B together are treated more or less unfairly than a
single person C? Arguably, the available evidence about fairness, which
mainly consists of vague moral intuitions, only supports the idea that a
single person can be treated more or less unfairly in relation to other
possible ways of treating that particular person. To say that person A
and B are facing more unfairness than C is a question-begging claim;
anyone wishing to make that claim first has to tell us more about the
concept of fairness. Of course, a mere reference to the Gini-concept
would not do in the present context. Firstly, it is far from clear that it
makes sense (in a substantial way, not just in a mere technical way) to
compare the Gini-index for populations of different sizes. Secondly, a
more fundamental problem is whether there is any reasonable way at
all in which fairness can be interpersonally measured. The Giniformula in itself does not answer that question. It is a mere mathematical abstraction, which may or may not correspond to a phenomenon in
the world.
Surely, if 1000 people are treated unfairly, that is morally worse than
if one person is treated unfairly, all else being equal. However, this conclusion does not presuppose that there exists some collective, interpersonal entity corresponding to an aggregated amount of unfairness.
Broome, Kamm on Fairness, pp. 95657.
16
366
martin peterson
It is sufficient to point out that 1000 people are being treated badly; the
reason need not be that the Gini-index of the corresponding distribution of probabilities is morally unfavourable.
4. An Alternative Account of the Mixed Solution
This section proposes an alternative account of the mixed solution,
which avoids at least two of the problems associated with the BroomeHirose analysis. Firstly, the new account does not assume that unfairness can be aggregated interpersonally. As argued above, there is no
reason to believe that there is such thing as an accumulated sum of
unfairness for each group of people treated badly. Secondly, the new
account does not assume that moral value can be compared interpersonally; this point is analogous to the first one.
The main feature of the new account is that fairness is construed as a
moral value that is good for the individual. There are no collective
moral values. All moral values, including fairness, can at most be aggregated on a personal level. Consider the following, general formulation
of this idea:
Individualism: All moral values, including fairness, are individual values
which cannot be interpersonally aggregated.
367
about whether the winner suffers more or less unfairness than a group
of other athletes do together. Such claims would invariably require
interpersonal aggregation of individual moral values.
Hirose defines the concept of aggregation as follows:17 a vector of
normatively relevant features (a1, a2,, ai)is at least as F as (a'1, a'2,,
a'j)if and only if f (a1, a2,, ai) f (a'1, a'2,, a'j)where f( )is a strictly
increasing function that is not bound above. Suppose that F is fairness.
Then, the principle of Individualism should be taken to mean that there
is no function f( ) that orders different distributions according to
degrees of fairness. However, it might still be possible to rank the set of
vectors according to some other feature F', such as choice-worthiness.
Individualism is a moral principle on the same level of abstraction as
the two principles suggested by Hirose, i.e. Symmetry and Pareto. This
means that all three principles are fundamental moral principles, from
which everyday moral rules are to be derived. Taken together, these
principles can explain the moral difference between the choice of saving 1000 or 1001 people, r espectively the choice of saving either one or
two people. First consider the choice between saving 1000 people or
1001. In Table 5, one could either save a group, A, of 1000 people,
oranother group B of 1001 people. By applying Symmetry and Pareto
to the latter alternative, it follows that this alternative is morally better
than saving the smaller group A. However, the third alternative, tossing
a fair coin, will be morally better than the second alternative in case
u > 0.5. Let us, for the sake of the argument, suppose that this condition
is fulfilled. It then follows that tossing a fair coin is the morally right
Table 5: Save 1000 or 1001?
A1
A2
A999
A1000
B1
B2
B1000
B1001
Save A
(1000)
Save B
(1001)
Toss a
coin
1u
1u
0u
0u
0u
0u
1u
1u
0.5
0.5
0.5
0.5
17
368
martin peterson
A1
B1
B2
Save one
Save two
Toss a coin
0 u'
1 u'
0.5
0 u'
1 u'
0.5
1 u'
0 u'
0.5
thing to do in the choice between saving A or B. Note that this conclusion has been reached without assuming that moral values, such as
fairness or utility, can be interpersonally aggregated.
Now consider the choice between saving either one or two people;
see Table 6. Since each person treated unfairly in this example faces less
unfairness than in the previous example, it may safely be assumed that
u' < 0.5. In this example, the moral principles considered so far are
insufficient for deciding what to do. This is because no alternative is
Pareto-better than the others. All that can be concluded is that saving
two people is morally better than saving one. In order to conclude that
saving two is better than tossing a coin, one has to assume that u is low
and that moral values are not individual, i.e., that they can be aggregated interpersonally by, e.g., adding each individuals value to that of
others. As argued above, this is a questionable assumption. Before one
decides whether to save one, or two, or to toss a fair coin, some additional moral assumptions are required.
However, there is an easy way out of this predicament. The key idea
is to make a sharp distinction between evaluative and normative claims.
In Table 6, each number represents each individuals moral value under
different alternative acts. The numbers can thus be seen as a summary
of a set of evaluative statements. This set of evaluative statements is, of
course, consistent with several different normative (action-guiding)
principles. For instance, it could be argued that if an alternative is not
Pareto-better than a set of other alternatives, then all those alternatives
are morally permissible. In the example illustrated in Table 4, this liberal normative principle would imply that saving two p
eople and tossing a fair coin are permissible alternatives, whereas saving one person
is not. That is, an act is morally permissible as long as it is not defeated
by another act, in terms of Pareto-betterness. Typically, this means
that a rather large subset of a set of alternative acts will be morally
permissible.
369
To sum up, the main argument of this article looks as simple as it can
r easonably be. Both fairness and the total number of people saved matter, but neither value can be aggregated interpersonally. Therefore, the
liberal normative principle outlined here ought to be adopted when
facing the number problem.18
18
I would like to thank Iwao Hirose and Barbro Bjorkman for helpful comments on
an earlier draft. My work on this paper has been supported in part by a grant from the
Swedish Emergency Management Agency.
* I owe particular thanks to Brad Hooker, Jeff McMahan, Daniel Wikler, and Melissa
Lane, who have provided helpful criticisms and encouragement. Wong Wai-Hung and
Luke Mulhall have helped polish the English of this article.
372
william sin
1
Bernard Williams, Moral Luck (Cambridge: Cambridge University Press, 1981),
p. 18. See also James Griffin, Value Judgement (Oxford: Oxford University Press, 1996),
pp. 9091; Samuel Scheffler, Human Morality (New York: Oxford University Press,
1992), p. 102; Barbara Herman, Morality and Everyday Life, Proceedings and Addresses
of the American Philosophical Association 74 (2000), pp. 2945.
2
See e.g. Peter Singer, Famine, Affluence and Morality, Philosophy & Public Affairs
1 (1972), pp. 22943; Peter Unger, Living High and Letting Die: Our Illusion of Innocence
(New York: Oxford University Press, 1996); Shelley Kagan, The Limits of Morality
(Oxford: Oxford University Press, 1989).
373
be obviously wrong for her not to do so.3 However, there may be some
Minimalists who hold that morality may demand agents to make some
small, but less-than-moderate, amount of sacrifice for the needy. The
closer Moral Minimalism gets to the position of the Moral Moderate,
the harder Minimalism is to reject.
As a response to famine relief, Minimalism deservedly receives the
least attention from philosophers in the debate. But note that the
Minimalist View still dominates many peoples intuition about the case
of famine relief. People in the street will not condemn someone, who
has an average income in an affluent country, for choosing not to make
any donations to save the needy in poor countries. In the minds of
these people, even if it takes only trivial sacrifices to save the needys
lives, it will still be a charity, instead of a moral requirement, for agents
to make the donations.4
Extremism, Minimalism and Moderate Morality divide the conceptual space under which a theorist may take sides in the case of famine
relief.5 The soundness of these positions is dependent on a number of
considerations. One such consideration is the Principle of Easy Rescue.
That is, if an agent can prevent something very bad from happening, or
alleviate someones dire plight, by making a trivial to moderate sacrifice, then it will be wrong not to do so.6 In the case of famine relief,
itseems that agents have some obligation to help the needy. However,
note that complying with the Principle of Easy Rescue in such a
3
This is the Principle of Easy Rescue, which we will introduce shortly. We will consider its counterarguments in section two of the paper.
4
Singer, Famine, Affluence and Morality, p. 235.
5
There are different ways to assess the weight of a loss. We may assess the percentage of a persons total wealth or some absolute level of sacrifice. For someone as rich as
Bill Gates, an amount as much as one billion dollars is only a tiny percentage of his
wealth, yet a great sum in absolute terms. In this paper, while I use the idea of percentage to understand the concepts such as extreme sacrifice and moderate sacrifice, I have
in mind people with an average income in the affluent world. For a sharper understanding, we may need a progressive standard, or even a standard which is defined
with reference to an agents level of well-being after she has made the donation. See
Liam Murphy, Moral Demands in Nonideal Theory (New York: Oxford University
Press, 2000), pp. 2021.
6
T. M. Scanlon, What We Owe to Each Other (Cambridge, MA: Harvard University
Press, 1998), p. 224. Also see Singers Famine, Affluence and Morality, pp. 22943. On
another occasion, Singer expresses his argument in a form that is similar to the
Principle of Easy Rescue: [N]ot to relieve the suffering of the worlds poor, when we
can do so at little cost to ourselves, is a serious moral failing. Peter Singer, A Response,
in D. Jamieson (ed.), Singer and His Critics (Oxford: Blackwell, 1999), p. 302.
374
william sin
375
Extremism
376
william sin
that agents should sacrifice in the case of famine relief. (As this amount
may be great, moderate, or even small, even some Minimalists may
support, in principle, a kind of Threshold View.10) Second, agents are
morally free to decide when they will stop giving once they have made
a certain sum of aggregate sacrifice to the needy. Of course, since holders of the Threshold View suggest different thresholds, each of them
has to explain why it is the threshold she suggests that is the right
threshold. But this is not the most important issue here. I will now turn
to the issue concerning iterative effects that give rise to burdensome
demands. In the following discussion, I will not differentiate between
different Threshold Views.
When trivial moral demands iterate and become burdensome to
agents, a typical response is to emphasize the distinction between the
Iterative View and the Aggregate View. The Iterative View focuses on
the marginal cost that agents bear if they are to save an additional victim. As long as each of the victims can be saved at trivial costs, the
Principle of Easy Rescue applies. Then, agents compliance with the
Principle may generate great demands in aggregate.
Critics of the Iterative View may, however, hold that we do not have
to understand agents obligation in terms of their capability to save the
next victim. We may understand agents obligation in light of a cumulative approach. That is, if an agent has already contributed a great deal
to rescuing victims in poor countries, then the total sacrifices will
exceed a trivial (or even moderate) amount and she is not required to
donate further. So, on the Aggregate View, the Principle of Easy Rescue
does not apply to those agents who have accumulated a great sum of
total contribution, even though they can save the next victim by making an additional sacrifice that is trivial on absolute terms.11
Generally speaking, the Aggregate View is employed by Moderates
to reject Extremism. However, note that, it is possible for one to uphold
the Aggregate View without specifying which aggregate sum agents
should sacrifice. As holders of the Aggregate View believe in the existence of a threshold dividing the sphere of obligation and that of supererogation, we may now take the Aggregate View and the Threshold
10
This will be an exceptional case of Minimalism. With regard to the mainstream
Minimalists, their position will involve a denial of the Threshold View.
11
On the distinction between the Iterative View and the Aggregate View, see also
Garrett Cullitys The Moral Demands of Affluence (Oxford: Oxford University Press,
2004), pp. 8284.
377
12
In describing the various interpretations of the Aggregate View, Cullity has noted
in passing the oddness of this implication. See Moral Demands of Affluence, p. 84.
378
william sin
them; humanitarian agencies are willing to save lives via private funding, etc.
The Inexhaustible Nature of Positive Duties is especially hard to
reject when agents can save the next victim at a very small cost to
themselves. Suppose Joe has an average income in an affluent country
and he has aggregately contributed 5000 to Oxfam, which is indeed a
non-trivial sum. Now, if Joe can save the next victim only by giving
another 500, he may have reason not to help. However, if Joe can save
an additional victim by sacrificing one more penny, it seems wrong if he
does not do it, and this is so regardless of how many starving children
Joe has saved through his prior contribution to humanitarian agencies.13 In this case, the great aggregate sacrifice that Joe has made does
not provide him with any special prerogative against the claims of the
needy.
Consider this case. Suppose that Joe has discovered a magic button.
Every time he presses the button, he saves one life from starvation.
Suppose further that this button is portable and it will be effective only
if Joe does not press it again within every minute. So Joe can carry the
button around when he is doing other things, too.
In this case, no matter how many times Joe has pressed the button
before, the level of his aggregate sacrifice still provides no reason for him
to stop pressing the button for good. Or, no matter whether Jack has
donated 30 per cent or 80 per cent of his income to relief organizations
to save lives, he still has the same reason to press the button to save one
more life. What it shows, I think, is that when the Principle of Easy
Rescue is coupled with the Inexhaustible Nature of Positive Duties,
Super Extremism follows. When agents can save victims lives at no cost
to themselves, they may have to do it indefinitely. In the case of Joe he
may have to keep pressing the button until he physically cannot do so.
We are thus faced with a trilemma. First, the Principle of Easy Rescue
is compelling. It is hard to imagine that agents may not be responsible
for saving someones life if they can do it at trivial costs to themselves.
Second, positive duty seems to have an inexhaustible nature. It is not
something that can be fully completed. Agents will remain connected
with those who are in need insofar as they are in a position to provide
help. Third, it is hard to abandon the non-extreme nature of morality.
In the situation of the magic button, Joe faces a super-extreme demand.
Cf. Unger, Living High and Letting Die, pp. 5961.
13
379
He does not have the prerogative to stop helping others even after he
has contributed a great deal. This goes way beyond what people generally think morality may permissibly impose on agents.
2. The Arguments
Super Extremism is hard to accept. However, compared with the other
alternatives that are open to us, it seems that Super Extremism may be
the least implausible. My view can be expressed in this argument (SE):
1. We cannot reject the Principle of Easy Rescue.
2. We cannot reject the Inexhaustible Nature of Positive Duties.
3. The Synthesis: Taken together, the Principle of Easy Rescue and the
Inexhaustible Nature of Positive Duties imply Super Extremism.
Therefore, we cannot reject Super Extremism.
380
william sin
magic button once a week, or even just once a month. These agents have
not terminated their mission of saving lives, but their selected scheme
of contribution does allow sufficient room for them to live their own
lives. Since such a reaction to the case of famine relief does not involve
great sacrifices, the Synthesis in argument SE can be rejected.
This response from holders of the Aggregate View contends the
point that agents may follow the Inexhaustibility Principle without
having to meet burdensome demands. Their case might be supported
by the fact that agents in general comply with certain other persistent
demands too, such as the duty not to lie, not to steal and the duty not
to perform murder, etc., To fulfill these demands, agents may also need
to make an indefinite amount of sacrifices, big or small, from time to
time, but agents compliance in this aspect will not amount to SuperExtremism at all. However, for two reasons, the comparison is misleading. First, the persistent demands of the negative duties we
mentioned above have been accommodated into agents lives. Most
agents have accepted them as the background circumstance under
which they develop their life plans. So the demands and agents life
plans will not systematically conflict with one another. But with regard
to the case of magic button we discuss here, the demand emerges only
later on in agents lives. It is hard for agents to find room to accommodate the demands within the established order of their lives.
The second reason why the comparison is misleading is that in the
case of the magic button, if agents are allowed to comply with the
Inexhaustibility Principle only periodically, these agents can be acting
in contradiction to the prescription of the principle itself. Suppose Joe
chooses to press the magic button once a week. On the Aggregate View,
if Joe has just pressed the button, he has thereby obtained free time of
a whole week to deal with his own business without having to save
additional people he might save at trivial costs to himself. Most of us do
not think, and would not be convinced, that there can be such moral
free time (with respect to saving lives). Such a scheme of contribution
by periodic compliance does not differ in substance from the original
proposal of the holders of the Aggregate View; that is, agents may make
a great aggregate sacrifice and then become exempted from all future
obligations to help needy, including that of easy rescue. In both cases,
we allow agents to exit, temporally or permanently, from the obligation
to honour their positive duties. This is a permission that lets victims die
too easily.
381
What could be the problem with the Principle of Easy Rescue? Perhaps
we could doubt whether human actions should always be guided by
moral principles. For instance, features of the situation can vary insofar
as an agent has saved a lot of lives. There may be increasingly strong pro
tanto reasons for the agent to care for her own well-being once she has,
say, pressed the magic button a great number of times. This may justify
a different response when the agent comes upon the next victim.
My objection is that relevant features of the situation may not, however, change drastically even if agents have performed a great number
of easy rescues. As long as it takes very little effort to press the button,
and the performance of this action saves some starving victims life, the
reason for the agent to continue pressing the button overwhelms any
reason she might have for leaving the button for any substantial period
of time.14 (She can permissibly take breaks as long as these will enhance
her ability to keep on pressing the button in the future.) So, Moderates
cannot reasonably discard the Principle of Easy Rescue.
Here is another objection to the Principle of Easy Rescue: It is not
clear that even if agents can save peoples lives or alleviate someones
dire plight by easy rescue, they should always do it. For, if the good that
is generated is not great and persistent enough, it may not be worthwhile for agents to devote their efforts to the mission time and again.
Imagine a case in which an agent is left on a deserted island with a
dying patient, whose life can only be maintained by the operation of a
respirator. And it happens that this agent can make the respirator work
by peddling a bicycle every hour for ten to fifteen minutes to recharge
14
Brad Hooker, Intuitions and Moral Theorizing, in P. Stratton-Lake (ed.), Ethical
Intuitionism (Oxford: Clarendon Press, 2002), pp. 16183.
382
william sin
383
386
alison stone
388
alison stone
and that feminists have most regularly addressed and opposed methodological essentialism.4
Heyess typology helpfully clarifies the range of possible varieties of
essentialism, but, because she introduces precise distinctions into the
essentialism which feminists have generally discussed and criticized as
a relatively unified phenomenon, her account obscures how feminist
debates around essentialism have actually developed. Despite the variety of strands within these debates, retrospectively they can be seen to
be engaged with an essentialism that has a relatively unitary meaning,
deriving from the traditional philosophical understanding of essentialism. This relatively unitary sense of essentialism gives feminist debates
a coherent history, within which different contributions can be recognized to interweave with and build upon one another. To support this
assertion, I shall briefly reconstruct this history, starting from the philosophical sense of essentialism which forms the point of departure for
feminist explorations.
Philosophically, essentialism is the belief that things have essential
properties, properties that are necessary to those things being what
they are. Recontextualized within feminism, essentialism becomes the
view that there are properties essential to women, in that any woman
must necessarily have those properties to be a woman at all. So defined,
essentialism entails a closely related view, universalism: that there are
some properties shared by, or common to, all womensince without
those properties they could not be women in the first place. Essential
properties, then, are also universal. Essentialism as generally debated
in feminist circles embraces this composite view: that there are properties essential to women and which all women (therefore) share.5
4
Heyes, Line Drawings, p. 37. Charlotte Witt argues that feminist critiques of essentialism have four similar targets: (1) the metaphysical belief that gender and sex
are core attributes of the self; (2) biological determinism; (3) the belief that the word
feminine has a fixed meaning; (4) the practice of making false generalizations about
women; see Witt, Anti-Essentialism in Feminist Theory, Philosophical Topics 23.2
(1995), pp. 32144.
5
Naomi Schor stresses that essentialism and universalism differ in This Essentialism
Which Is Not One: Coming to Grips with Irigaray, in N. Schor and E. Weed (eds.), The
Essential Difference (Bloomington, IN: Indiana University Press, 1994), p. 24. One
might, for instance, argue that there are certain characteristics which all women share,
but which are accidental rather than essential (that is, these features could be changed
without women thereby ceasing to be women)such as, perhaps, the feature of being
disempowered relative to men. Yet, though universal features need not be essential,
essential properties are necessarily universal, hence essentialism and universalism are
generally assimilated in feminist discussion.
6
For a classic socialist feminist critique of radical feminism, see Lynne Segal, Is the
Future Female? (London: Virago, 1987).
390
alison stone
392
alison stone
394
alison stone
15
Denise Riley, Am I That Name?: Feminism and the Category of Women in History
(London: Macmillan, 1982), p. 112.
396
alison stone
19
20
398
alison stone
Sartre puts it, neutral sense.21 Yet the reason why Sartre continues to
identify series as groups in this broad sense is because he regards series
as self-alienated versions of strict, unified groups. For Sartre, series are
self-alienated groups in that their unity is located outside them, in the
objective artefacts or structures by which their members are constrained and organized.22 Thus, from Sartres perspective, series only
count as groups at all to the extent that they remain unified, albeit in a
relatively indirect, exteriorized way.
Youngs Sartrean concept of the series proves inadequate to the task
which she wishes it to play: that of enabling her to reconceive women
as a specifically non-unified type of social group. Despite this failing,
the importance of Youngs argument lies in her more basic insight that
reconceiving women as a non-unified type of social group could surmount the problems generated by feminist critiques of essentialism.
I therefore propose to develop Youngs insight, by jettisoning the idea
that women constitute a series, and instead rethinking women as having a genealogythat is, as constituting a group that is internally
diverse, and yet remains a group in virtue of having a complex history
composed of multiple, overlapping threads of interpretation.
Women as Having a Genealogy
In this section, I will argue that an appropriation of the concept
of genealogy can provide a way to reinstate the idea that women comprise a distinct social group even in the absence of any common properties that constitute them all as women. Several prominent feminist
thinkers have already drawn on the concept of genealogy: for example,
in Gender Trouble, Judith Butler proposes to outline a genealogical
understanding of what it means to be a woman.23 Similarly, Moira
Gatens praises the project of charting a genealogy of the category
woman or women. On this approach women itself is under
stoodto have a history, a genealogy, a line of descenta genealogical
approach asks: how has woman/women functioned as a discur
sive category throughout history?.24 These approving references to
21
Sartre, Critique of Dialectical Reason. I. Theory of Practical Ensembles (trans.
A. SheridanSmith; London: New Left Books, 1976), p. 256.
22
Sartre, Critique of Dialectical Reason, pp. 25859.
23
Butler, Gender Trouble, p. 5.
24
Gatens, Imaginary Bodies, p. 76.
400
alison stone
29
See Raymond Geuss, Nietzsche and Morality, in Raymond Geuss (ed.), Morality,
Culture and History: Essays on German Philosophy (Cambridge: Cambridge University
Press, 1999), pp. 16769.
30
Foucault, Nietzsche, Genealogy, History, p. 346.
31
Nietzsche, On the Genealogy of Morality, p. 55.
32
Nietzsche, On the Genealogy of Morality, p. 57.
33
As Geuss explains, Nietzsches conflictual model implies that no reinterpretation
will ever encounterjust a tabula rasa, but a set of actively structured forces, practices
etc. which will be capable of active resistance (Geuss, Nietzsche and Genealogy, in
Morality, Culture, and History, p. 13). This resistance ensures that new interpretations
will not in general be so fully successful that nothingremains of the pre-existing
meanings (p. 11).
34
Foucault, Nietzsche, Genealogy, History, pp. 34748.
402
alison stone
404
alison stone
406
alison stone
Bibliography
The Journal of Moral Philosophy originally published the essays in this
collection. This bibliography lists the full contents of all articles published since the JMP was launched in April 2004. The JMP acceptance
rate is about 7%.
Allen, Robert Francis, Robust Alternatives and Responsibility, Journal of Moral
Philosophy 1(1) (2004), pp. 2129.
Alm, David, Deontological Restrictions and the Good/Bad Asymmetry, Journal of
Moral Philosophy 6(4) (2009), pp. 46481.
Anderson, Scott A., Of Theories of Coercion, Two Axes, and the Importance of the
Coercer, Journal of Moral Philosophy 5(3) (2008), pp. 394422.
Andreou, Chrisoula, Standards, Advice, and Practical Reason, Journal of Moral
Philosophy 3(1) (2006), pp. 5767.
Bedi, Sonu, Expressive Exclusion: A Defense, Journal of Moral Philosophy 7(4) (2010),
pp. 42740.
Bennett, Christopher, Forgiveness and the Claims of Retribution, Journal of Moral
Philosophy 1(1) (2004), pp. 89101.
Bennett, Christopher, State Denunciation of Crime, Journal of Moral Philosophy 3(3)
(2006), pp. 288304.
Besser-Jones, Lorraine, Personal Integrity, Morality and Psychological Well-Being:
Justifying the Demands of Morality, Journal of Moral Philosophy 5(3) (2008),
pp. 36183.
Besson, Samantha, Democracy, Law, and Authority, Journal of Moral Philosophy 2(1)
(2005), pp. 8999.
Bittner, Rdiger, A Horse in the Basement: Nietzschean Reflections on Political
Philosophy, Journal of Moral Philosophy 7(2) (2010), pp. 32133.
Brake, Elizabeth, Rawls and Feminism: What Should Feminists Make of Liberal
Neutrality? Journal of Moral Philosophy 1(3) (2004), pp. 293309.
Reprinted in Thom Brooks and Fabian Freyenhagen (eds), The Legacy of John
Rawls. New York: Continuum, 2005.
Brock, Gillian, The Difference Principle, Equality of Opportunity, and Cosmopolitan
Justice, Journal of Moral Philosophy 2(3) (2005), pp. 33351.
Broome, John, Does Rationality Consist in Responding Correctly to Reasons? Journal
of Moral Philosophy 4(3) (2007), pp. 34974.
Brown, Stephen, Naturalized Virtue Ethics and the Epistemological Gap, Journal of
Moral Philosophy 1(2) (2004), pp. 197209.
Butler, Brian E., Rorty, the First Amendment and Antirealism: Is Reliance upon Truth
Viewpoint-Based Speech Regulation? Journal of Moral Philosophy 1(1) (2004),
pp. 6988.
Carens, Joseph, The Integration of Immigrants, Journal of Moral Philosophy 2(1)
(2005), pp. 2946.
Carter, Alan, The Evolution of Rawlss Justification of Political Compliance: Part 1 of
The Problem of Political Compliance in Rawlss Theories of Justice, Journal of Moral
Philosophy 3(1) (2006), pp. 721.
Carter, Alan, Political Liberalism and Political Compliance: Part 2 of The Problem of
Political Compliance in Rawlss Theories of Justice, Journal of Moral Philosophy 3(2)
(2006), pp. 13557.
410
bibliography
Clark, Michael, Retribution and Organic Unities, Journal of Moral Philosophy 3(3)
(2006), pp. 35158.
Choi, Yoon, Revisiting Kants Ethics: Two Challenges to the Status Quo, Journal of
Moral Philosophy 5(1) (2008), pp. 13749.
Cholbi, Michael, A Kantian Defense of Prudential Suicide, Journal of Moral Philosophy
7(4) (2010), pp. 489515.
Christie, Tim W., Natural Separateness: Why Parfits Reductionist Account of Persons
Fails to Support Consequentialism, Journal of Moral Philosophy 6(2) (2009),
pp. 17895.
Coyle, Sean, The Ideality of Law, Journal of Moral Philosophy 6(4) (2009),
pp. 52134.
Crisp, Roger, Ethics Without Reasons? Journal of Moral Philosophy 4(1) (2007),
pp. 4049.
Cruft, Rowan, On the Non-instrumental Value of Basic Rights, Journal of Moral
Philosophy 7(4) (2010), pp. 44161.
DAgostino, Fred, The Legacies of John Rawls, Journal of Moral Philosophy 1(3) (2004),
pp. 34965.
Reprinted in Thom Brooks and Fabian Freyenhagen (eds), The Legacy of John
Rawls. New York: Continuum, 2005.
Dancy, Jonathan, Defending the Right, Journal of Moral Philosophy 4(1) (2007),
pp. 8598.
DeGrazia, David, Moral Vegetarianism from a Very Broad Basis, Journal of Moral
Philosophy 6(2) (2009), pp. 14365.
de Muijnk, Wim, Thinking about Normativity: Ralph Wedgwood on Ought, Journal
of Moral Philosophy 7(1) (2009), pp. 13344.
Deonna, Julien A., The Structure of Empathy, Journal of Moral Philosophy 4(1) (2007),
pp. 99116.
Dietsch, Peter, Distributive Lessons from Division of Labour, Journal of Moral
Philosophy 5(1) (2008), pp. 96117.
Enoch, David and Ehud Guttel, Cognitive Biases and Moral Luck, Journal of Moral
Philosophy 7(2) (2010), pp. 37286.
Eylon, Yuval, Just Threats, Journal of Moral Philosophy 6(1) (2009), pp. 94108.
Flikschuh, Katrin, Duty, Nature, Right: Kants Response to Mendelssohn in Theory and
Practice III, Journal of Moral Philosophy 4(2) (2007), pp. 22341.
Friedman, Alex, Intransitive Ethics, Journal of Moral Philosophy 6(3) (2009),
pp. 27797.
Fuller, Lisa F., Poverty Relief, Global Institutions, and the Problem of Compliance,
Journal of Moral Philosophy 2(3) (2005), pp. 28597.
Reprinted in Thom Brooks (ed.), The Global Justice Reader. Oxford: Blackwell,
2008.
Goddu, G. C., More on Blameworthiness and Alternative Possibilities, Journal of
Moral Philosophy 3(1) (2006), pp. 6975.
Grau, Christopher, Moral Status, Speciesism, and Liaos Genetic Account, Journal of
Moral Philosophy 7(2) (2009), pp. 38796.
Green, Michael, Social Justice, Voluntarism, and Liberal Nationalism, Journal of Moral
Philosophy 2(3) (2005), pp. 26583.
Hall, Timothy, Doing Harm, Allowing Harm, and Denying Resources, Journal of
Moral Philosophy 5(1) (2008), pp. 5076.
Harcourt, Edward, Crisps Ethics Without Reasons?: A Note on Invariance, Journal
of Moral Philosophy 4(1) (2007), pp. 5054.
Hayward, Tim, Thomas Pogges Global Resources Dividend: A Critique and an
Alternative, Journal of Moral Philosophy 2(3) (2005), pp. 31732.
Hendrix, Burke A., Authenticity and Cultural Rights, Journal of Moral Philosophy 5(2)
(2008), pp. 181203.
bibliography411
Hills, Alison, Practical Reason, Value and Action, Journal of Moral Philosophy 4(3)
(2007), pp. 37592.
Hirose, Iwao, Aggregation and Non-Utilitarian Moral Theories, Journal of Moral
Philosophy 4(2) (2007), pp. 27384.
Hofmeyr, Benda, The Power Not to Be (What We Are): The Politics and Ethics of Selfcreation in Foucault, Journal of Moral Philosophy 3(2) (2006), pp. 21530.
Holroyd, Jules, Substantively Constrained Choice and Deference, Journal of Moral
Philosophy 7(2) (2010), pp. 18099.
Hookway, Christopher, Ethics and the Pragmatist Enlightenment, Journal of Moral
Philosophy 3(2) (2006), pp. 23136.
Hutton, Eric, Han Feizis Criticism of Confucianism and Its Implications for Virtue
Ethics, Journal of Moral Philosophy 5(3) (2008), pp. 42353.
James, Aaron, Rights and Circularity in Scanlons Contractualism, Journal of Moral
Philosophy 1(3) (2004), pp. 36774.
Reprinted in Thom Brooks and Fabian Freyenhagen (eds), The Legacy of John
Rawls. New York: Continuum, 2005.
Kaufman, Alexander, Rawlss Practical Conception of Justice: Opinion, Tradition and
Objectivity in Political Liberalism, Journal of Moral Philosophy 3(1) (2006), pp. 2343.
Kirchin, Simon, Moral Particularism: An Introduction, Journal of Moral Philosophy
4(1) (2007), pp. 815.
Kirchin, Simon, Particularism and Default Valency, Journal of Moral Philosophy 4(1)
(2007), pp. 1632.
Knight, Carl, Egalitarian Justice and Valuational Judgment, Journal of Moral Philosophy
6(4) (2009), pp. 48298.
Knight, Kelvin, MacIntyres Progress, Journal of Moral Philosophy 6(1) (2009),
pp. 11526.
Kristjnsson, Kristjn, A Utilitarian Justification of Desert in Distributive Justice,
Journal of Moral Philosophy 2(2) (2005), pp. 14770.
Laden, Anthony Simon, Taking the Distinction between Persons Seriously, Journal of
Moral Philosophy 1(3) (2004), pp. 27792.
Reprinted in Thom Brooks and Fabian Freyenhagen (eds), The Legacy of John
Rawls. New York: Continuum, 2005.
Landemore, Hlne, Politics and the Economist-King: Is Rational Choice Theory the
Science of Choice? Journal of Moral Philosophy 1(2) (2004), pp. 17796.
Landrum, Ty, Persons as Objects of Love, Journal of Moral Philosophy 6(4) (2009),
pp. 41739.
Lang, Gerald, Luck Egalitarianism, Permissible Inequalities, and Moral Hazard,
Journal of Moral Philosophy 6(3) (2009), pp. 31738.
Lefkowitz, David, Partiality and Weighing Harm to Non-Combatants, Journal of
Moral Philosophy 6(3) (2009), pp. 298316.
Lenard, Patti Tamara, Motivating Cosmopolitanism? A Skeptical View, Journal of
Moral Philosophy 7(2) (2010), pp. 34671.
Lenman, James, How to Live, What to Do: A Critical Study of Allan Gibbard, Thinking
How to Live, Journal of Moral Philosophy 3(3) (2006), pp. 35969.
Liao, S. Matthew, Time-Relative Interests and Abortion, Journal of Moral Philosophy
4(2) (2007), pp. 24256.
Liao, S. Matthew, The Basis of Human Moral Status, Journal of Moral Philosophy 7(2)
(2010), pp. 15979.
Lippert-Rasmussen, Kasper, Publicity and Egalitarian Justice, Journal of Moral
Philosophy 5(1) (2008), pp. 3049.
Lippke, Richard L., Imprisonable Offenses, Journal of Moral Philosophy 3(3) (2006),
pp. 26587.
Lippke, Richard L., Punishing the Guilty, Not Punishing the Innocent, Journal of
Moral Philosophy 7(4) (2010), pp. 46288.
412
bibliography
Ludwig, Bernd, Kant, Garve, and the Motives of Moral Action, Journal of Moral
Philosophy 4(2) (2007), pp. 18393.
Lynch, Sterling, The Fact of Diversity and Reasonable Pluralism, Journal of Moral
Philosophy 6(1) (2009), pp. 7093.
Mahoney, Jon, Public Reason and the Moral Foundation of Liberalism, Journal of
Moral Philosophy 1(3) (2004), pp. 31131.
Reprinted in Thom Brooks and Fabian Freyenhagen (eds), The Legacy of John
Rawls. New York: Continuum, 2005.
Marino, Patricia, Expressivism, Deflationism and Correspondence, Journal of Moral
Philosophy 2(2) (2005), pp. 17191.
Marino, Patricia, Moral Rationalism and the Normative Status of Desiderative
Coherence, Journal of Moral Philosophy 7(2) (2010), pp. 22752.
Martin, Rex, Twp Concepts of Rule Utilitarianism, Journal of Moral Philosophy 5(2)
(2008), pp. 22755.
Martin, Wayne, Hegel and the Philosophy of Food, Journal of Moral Philosophy 7(2)
(2009), pp. 27990.
Matravers, Matt, Whos Still Standing? A Comment on Antony Duff s Preconditions
of Criminal Liability, Journal of Moral Philosophy 3(3) (2006), pp. 32030.
McKeever, Sean and Michael Ridge, Turning on Default Reasons, Journal of Moral
Philosophy 4(1) (2007), pp. 5576.
Merritt, Maria W., Aristotelian Virtue and the Interpersonal Aspect of Ethical
Character, Journal of Moral Philosophy 6(1) (2009), pp. 2349.
Mookherjee, Monica, Feminism and MulticulturalismPutting Okin and Shachar in
Question, Journal of Moral Philosophy 2(2) (2005), pp. 23741.
Moyar, Dean, Unstable Autonomy: Conscience and Judgment in Kants Moral
Philosophy, Journal of Moral Philosophy 5(3) (2008), pp. 32760.
Mulnix, M. J., Harm, Rights, and Liberty: Towards a Non-Normative Reading of Mills
Liberty Principle, Journal of Moral Philosophy 6(2) (2009), pp. 196217.
Naticchia, Chris, The Law of Peoples: The Old and the New, Journal of Moral Philosophy
2(3) (2005), pp. 35369.
Reprinted in Thom Brooks and Fabian Freyenhagen (eds), The Legacy of John
Rawls. New York: Continuum, 2005.
Nolan, Daniel, Consequentialism and Side Constraints, Journal of Moral Philosophy
6(1) (2009), pp. 522.
Norman, Richard, Particularism and Reasons: A Reply to Kirchin, Journal of Moral
Philosophy 4(1) (2007), pp. 3339.
Nussbaum, Martha C., Radical Evil in the Lockean State: The Neglect of the Political
Emotions, Journal of Moral Philosophy 3(2) (2006), pp. 15978.
ONeill, Martin, The Facts of Inequality, Journal of Moral Philosophy 7(2) (2009),
pp. 397409.
ONeill, Onora, Experts, Practitioners, and Practical Judgement, Journal of Moral
Philosophy 4(2) (2007), pp. 15466.
ONeill, Onora, Normativity and Practical Judgement, Journal of Moral Philosophy
4(3) (2007), pp. 393405.
verland, Gerhard, Self-defence among Innocent People, Journal of Moral Philosophy
2(2) (2005), pp. 12746.
verland, Gerhard, Poverty and the Moral Significance of Contribution, Journal of
Moral Philosophy 2(3) (2005), pp. 299315.
verland, Gerhard, Conditional Threats, Journal of Moral Philosophy 7(2) (2010),
pp. 33445.
Papdaki, Lina, What is Objectification? Journal of Moral Philosophy 7(1) (2010),
pp. 1636.
Peterson, Martin, The Mixed Solution to the Number Problem, Journal of Moral
Philosophy 6(2) (2009), pp. 16677.
bibliography413
Pink, Thomas, Normativity and Reason, Journal of Moral Philosophy 4(3) (2007),
pp. 40631.
Primoratz, Igor, Patriotism and Morality: Mapping the Terrain, Journal of Moral
Philosophy 5(2) (2008), pp. 20426.
Redondo, Mara Cristina, Legal Reasons: Between Universalism and Particularism,
Journal of Moral Philosophy 2(1) (2005), pp. 4768.
Reitan, Eric, Defining Terrorism for Public Policy Purposes: The Group-Target
Definition, Journal of Moral Philosophy 7(2) (2010), pp. 25378.
Richardson, Henry S., Our Call: The Constitutive Importance of the Peoples Judgment,
Journal of Moral Philosophy 5(1) (2008), pp. 329.
Ridge, Michael, Anti-Reductionism and Supervenience, Journal of Moral Philosophy
4(3) (2007), pp. 33048.
Riley, Evan, Libertarian Self-Defeat, Journal of Moral Philosophy 7(2) (2010),
pp. 20026.
Riley, Patrick, Kant against Hobbes in Theory and Practice, Journal of Moral Philosophy
4(2) (2007), pp. 194206.
Riley, Jonathan, Ethical Pluralism and Common Decency, Journal of Moral Philosophy
1(2) (2004), pp. 21121.
Ronzoni, Miriam, Constructivism and Practical Reason: On Intersubjectivity,
Abstraction, and Judgment, Journal of Moral Philosophy 7(1) (2010), pp. 74104.
Rosebury, Brian, Reply to Silcox on Moral Luck, Journal of Moral Philosophy 6(1)
(2009), pp. 10913.
Sable, Andrew, Virtue for Pluralists, Journal of Moral Philosophy 2(2) (2005),
pp. 20735.
Scarre, Geoffrey, Corrective Justice and Reputation, Journal of Moral Philosophy 3(3)
(2006), pp. 30519.
Scarre, Geoffrey, The Banality of Good? Journal of Moral Philosophy 6(4) (2009),
pp. 499519.
Segall, Shlomi, How Devolution Upsets Distributive Justice, Journal of Moral
Philosophy 4(2) (2007), pp. 25772.
Seglow, Jonathan, Associative Duties and Global Justice, Journal of Moral Philosophy
7(1) (2010), pp. 5473.
Shafer-Landau, Russ, Moral and Theological Realism: The Explanatory Argument,
Journal of Moral Philosophy 4(3) (2007), pp. 31129.
Sheinman, Hanoch, Raz on the Social Dependence of Values, Journal of Moral
Philosophy 3(1) (2006), pp. 7787.
Silcox, Mark, Virtue Epistemology and Moral Luck, Journal of Moral Philosophy 3(2)
(2006), pp. 17992.
Silcox, Mark, Reply to Rosebury, Journal of Moral Philosophy 6(2) (2009),
pp. 24548.
Simpson, Matthew, A Paradox of Sovereignty in Rousseaus Social Contract, Journal of
Moral Philosophy 3(1) (2006), pp. 4556.
Sin, William, Trivial Sacrifices, Great Demands, Journal of Moral Philosophy 7(1)
(2010), pp. 315.
Singleton, Jane, Neither Generalism nor Particularism: Ethical Correctness is Located
in General Ethical Theories, Journal of Moral Philosophy 1(2) (2004), pp. 15575.
Slomp, Gabriella, Kant against Hobbes: Reasoning and Rhetoric, Journal of Moral
Philosophy 4(2) (2007), pp. 20722.
Smith, M. B. E., Does Humanity Share a Common Moral Faculty? Journal of Moral
Philosophy 7(1) (2010), pp. 3753.
Smith, Stephen R., Keeping Our Distance in Compassion-Based Social Relations,
Journal of Moral Philosophy 2(1) (2005), pp. 6987.
Sobel, David, The Limits of the Explanatory Power of Developmentalism, Journal of
Moral Philosophy 7(4) (2010), pp. 51727.
414
bibliography
Spector, Jessica, The Grounds of Moral Agency: Lockes Account of Personal Identity,
Journal of Moral Philosophy 5(2) (2008), pp. 25681.
Stark, Susan, A Change of Heart: Moral Emotions, Transformation, and Moral Virtue,
Journal of Moral Philosophy 1(1) (2004), pp. 3150.
Stern, Robert, The Autonomy of Morality and the Morality of Autonomy, Journal of
Moral Philosophy 6(3) (2009), pp. 395415.
Stone, Alison, Essentialism and Anti-Essentialism in Feminist Philosophy, Journal of
Moral Philosophy 1(2) (2004), pp. 13553.
Talisse, Robert B., Does Value Pluralism Entail Liberalism? Journal of Moral Philosophy
7(2) (2010), pp. 30320.
Taylor, Robert S., Self-Realization and the Priority of Fair Equality of Opportunity,
Journal of Moral Philosophy 1(3) (2004), pp. 33347.
Reprinted in Thom Brooks and Fabian Freyenhagen (eds), The Legacy of John
Rawls. New York: Continuum, 2005.
Telfer, Elizabeth, Animals Do It Too!: The Franklin Defence of Meat-Eating, Journal
of Moral Philosophy 1(1) (2004), pp. 5167.
Thomas, Alan, Practical Reasoning and Normative Relevance: A Reply to McKeever
and Ridge, Journal of Moral Philosophy 4(1) (2007), pp. 7784.
Tideman, Nicolaus, Secession as a Human Right, Journal of Moral Philosophy 1(1)
(2004), pp. 919.
Timmermann, Jens, Good but Not Required?Assessing the Demands of Kantian
Ethics, Journal of Moral Philosophy 2(1) (2005), pp. 927.
Timmermann, Jens, Simplicity and Authority: Reflections on Theory and Practice in
Kants Moral Philosophy, Journal of Moral Philosophy 4(2) (2007), pp. 16782.
Tropman, Elizabeth, Renewing Moral Intuitionism, Journal of Moral Philosophy 6(4)
(2009), pp. 44063.
Tyler, Colin, Brian Barry and Writings on Social Justice from the Left, Journal of Moral
Philosophy 5(2) (2008), pp. 30112.
Upton, Candace, Virtue Ethics, Character, and Normative Receptivity, Journal of
Moral Philosophy 8(1) (2008), pp. 7795.
van Zyl, Liezl, Agent-based Virtue Ethics and the Problem of Action Guidance, Journal
of Moral Philosophy 6(1) (2009), pp. 5069.
Vargas, Manuel, Taking the Highway on Skepticism, Luck, and the Value of
Responsibility, Journal of Moral Philosophy 6(2) (2009), pp. 24965.
Ward, Lee, Locke on Punishment, Property and Moral Knowledge, Journal of Moral
Philosophy 6(2) (2009), pp. 21844.
Webber, Jonathan, Virtue, Character and Situation, Journal of Moral Philosophy 3(2)
(2006), pp. 193213.
Weirich, Paul, Utility Maximization Generalized, Journal of Moral Philosophy 5(2)
(2008), pp. 28299.
Wenar, Leif, The Unity of Rawlss Work, Journal of Moral Philosophy 1(3) (2004),
pp. 26575.
Reprinted in Thom Brooks and Fabian Freyenhagen (eds), The Legacy of John
Rawls. New York: Continuum, 2005.
Westphal, Kenneth R., From Convention to Ethical Life: Humes Theory of Justice
in Post-Kantian Perspective, Journal of Moral Philosophy 7(1) (2010), pp. 10532.
White, Heath, Fitting Attitudes, Wrong Kinds of Reasons, and Mind-Independent
Goodness, Journal of Moral Philosophy 6(3) (2009), pp. 33964.
Wiland, Eric, On Indirectly Self-defeating Moral Theories, Journal of Moral Philosophy
5(3) (2008), pp. 38493.
Williams, Reginald, Morality and Privilege, Journal of Moral Philosophy 5(1) (2008),
pp. 11835.
Wolff, Jonathan, Equality: The Recent History of an Idea, Journal of Moral Philosophy
4(1) (2007), pp. 12536.
bibliography415
Zaibert, Leo, The Fitting, the Deserving, and the Beautiful, Journal of Moral Philosophy
3(3) (2006), pp. 33150.
Zaibert, Leo, The Paradox of Forgiveness, Journal of Moral Philosophy 6(3) (2009),
pp. 36593.
Ziegler, Rafael, Tracing Global Inequality in Eco-space: A Comment on Tim Haywards
Proposal, Journal of Moral Philosophy 4(1) (2007), pp. 11724.
index
abortion 4, 156, 292, 299, 31528, 330,
333, 336, 339, 3478, 350, 353
Andreou, Chrisoula 1
animal/animals 11, 223, 745, 82, 216,
319, 321, 323, 32730, 3367, 339,
3445, 349
Anscombe, G. E. M. 80, 86
anti-essentialism see essentialism
Aristotle 18, 88, 175, 183, 207, 210,
2138, 222, 234, 261, 2737
Aristotelian 4, 64, 209, 271
Ashford, Elizabeth 374
attitudinal reason see reason/reasons
Brink, David 132, 242
Brooks, Thom 45, 287, 361
Broome, John 1, 64, 317, 330, 335,
3578, 362, 3656
buck-passing122
character 4, 116, 190, 201, 204, 20713,
21528, 2302, 247, 25964, 26873,
2768, 2813, 387, 38997
common sense 2, 989, 238, 323
Confucius 34, 1758, 1802,
1989, 201
consequentialism 66, 175, 232, 239, 248,
289, 304, 307, 375
Crisp, Roger 2, 10710
Dancy, Jonathan 2, 95107, 1656, 205
death 4, 141, 194, 288, 292, 298, 300,
3179, 326, 328, 3323, 336, 350
DeGrazia, David 318, 320, 326, 328
Doris, John 210, 235, 25971, 273, 277
essentialism 5, 38598, 404, 4067
anti-essentialism 385, 392, 395, 405
experimental psychology 4, 210
evil 4, 134, 215, 226, 2334, 257,
399, 401
Feizi, Han 34, 17380, 182, 184205
feminism/feminist 5, 38598, 4057
Freyenhagen, Fabian 1, 56
friendship 18, 69, 207, 214, 217,
220, 254
418 index
Mackie, J. L. 127, 153, 161
Majors, Brad 125, 132, 147, 1623
Marcus, Ruth Barcan 91, 113
McDowell, John 88, 183, 205, 2746
McKeever, Sean 2, 1112, 147
McMahan, Jeff 4, 28795, 309, 314,
31729, 3313, 335, 3378, 344,
353, 371
Mencius 1768, 182, 1856, 188, 1924,
1967, 201
Merritt, Maria 4, 209, 2601, 268,
270, 277
metaethics 1, 3, 57, 126, 147
Milgram, Stanley 2656
Miller, Christian B. 209, 262, 2689
moral facts 3, 12738, 1404,
154, 209
moral non-naturalism 147
moral realism 1, 3, 1256, 1302, 134,
1368, 1534, 156, 159
moral status 4, 289, 293, 298, 301,
303, 3135, 326, 333, 339, 346,
3513, 356
Rabinowicz, Wlodek 25
rationality 1, 256, 2832, 345, 3745,
4752, 545, 59, 81, 105, 145, 210,
224, 2402, 276, 319, 340
reason/reasons
attitudinal reason 302, 347
practical reason 50, 54, 57, 5968, 73,
76, 99100
Ridge, Michael 23, 1112
Ross, W. D. 98, 167
Nagel, Thomas 31
non-Western philosophy 3
Norman, Richard 2, 1167, 121
norms 2, 7785, 87, 8992, 125,
209, 2112, 226, 229, 231,
397, 402
normativity 2, 59, 812, 11920
Nozick, Robert 305, 3089, 336
number problem 5, 3579,
3612, 369
ONeill, Onora 2, 92
Otsuka, Michael 287, 307, 357
particularism 1, 95, 979, 101, 106,
1112, 116
Peterson, Martin 5
Plato 11, 165, 210
practical judgement 2, 78, 85, 8792
practical reason see reason/reasons
punishment 26, 178, 189, 193, 400
vegetarian/vegetarianism11
Velleman, David 5860, 6667
virtue ethics 1, 34, 1745, 1823, 201,
20511, 219, 2325, 2602, 264, 271
Webber, Jonathan 4, 208
Wedgwood, Ralph 29, 125
Wiggins, David 88, 276
Williams, Bernard 13, 1920, 613, 88,
100, 183, 189, 1912, 195, 1978,
2013, 213, 276, 3356, 372
Xunzi 1734, 1778, 184, 188, 194,
199201
Young, Iris Marion 386, 392, 395
Zimmerman, David 132, 1345