Philip Kitcher The Ethical Project 2011

The Ethical Project
The Ethical Project
Philip Kitcher
Harvard University Press

Cambridge, Massachusetts • London, England
2011
Copyright © 2011 by the President and Fellows of Harvard College
All rights reserved
Printed in the United States of America
Library of Congress Cataloging-in-Publication Data
Kitcher, Philip, 1947-

The ethical project / Philip Kitcher.
p. cm.
Includes bibliographical references (p. ) and index.
ISBN 978-0-674-06144-6 (hardcover : alk. paper)
1. Ethics, Evolutionary. I. Title.
BJ1311.K53 2011
171'.7—dc22 2010051684
For A.G.P.K.
Contents
Introduction 1
§1. The Shape of Things to Come 1
§2. Methodological Preliminaries 9
I. An Analytical History
1. The Springs of Sympathy 17
§3. Psychological Altruism: Basics 17
§4. The Varieties of Altruistic Reactions 25
§5. Some Dimensions of Altruism 31
§6. Maternal Concern 35
§7. Broader Forms of Altruism? 42
§8. Possibilities of Evolutionary Explanation 47
§9. The Coalition Game 57
2. Normative Guidance 67
§10. The Limits of Altruism 67
§11. Following Orders 74
§12. Punishment 87
§13. Conscience 92
§14. Social Embedding 96
viii Contents
3. Experiments of Living 104

§15. From There to Here 104
§16. Cultural Competition 107
§17. The Unseen Enforcer 111
§18. Some Dots to Be Connected 115
§19. Divisions of Labor 122
§20. Roles, Rules, and Institutions 125
§21. Altruism Expanded 131
4. One Thing after Another? 138
§22. Mere Change? 138
§23. Three Ancient Examples 140
§24. Second-Sex Citizens 145
§25. Repudiating Chattel Slavery 153
§26. The Withering of Vice 162
§27. The Divine Commander 166
II. A Metaethical Perspective

5. Troubles with Truth 173
§28. Taking Stock 173
§29. Prima Facie Problems 178
§30. Truth, Realism and Constructivism 186
§31. The Sources of the Troubles 193
6. Possibilities of Progress 209
§32. The Centrality of Ethical Progress 209
§33. Generalizations from History 213
§34. Problems, Functions and Progress 218
§35. Modes of Refinement 229
§36. Functional Generation 237
§37. Local and Global Progress 242
§38. Ethical Truth Revisited 245
§39. Residual Concerns 249
7. Naturalistic Fallacies? 253
§40. Hume’s Challenge 253
§41. Authority Undermined? 263
Contents ix
§42. Troublesome Characters 269

§43. Settling Disputes 280
III. A Normative Stance

8. Progress, Equality, and the Good 285
§44. Two Visions of Normative Ethics 285
§45. Dynamic Consequentialism 288
§46. Failures and Successes 294
§47. From the Local Community to the Human
Population 302
§48. Equality and the Good Life 311
§49. Population Size 318
§50. Aspects of the Good Life 324
9. Method in Ethics 330
§51. Varieties of Ethical Change 330
§52. Method and the Good 338
§53. Mutual Engagement 342
§54. Ethical Debate 349
§55. Dissent and the Limits of Tolerance 355
§56. The Challenger Revisited 360
10. Renewing the Project 370
§57. Philosophical Midwifery 370
§58. Scarce Resources 374
§59. Habits and Their Limits 381
§60. Confl icting Roles 385
§61. Ethically Insulated Spheres 391
§62. Maintaining Equality 396
§63. The Challenges of Technology 401
Conclusion 409
§64. Summing Up 409
Acknowledgments 411
Index 417
Moral conceptions and processes grow naturally
out of the very conditions of human life.
—John Dewey
Introduction
§1. The Shape of Things to Come

Ethics pervades every human society and almost every human life. Peo-
ple deliberate about what they should do on specific occasions, about
what is worthwhile, about the kinds of lives they should aspire to lead.
In subtle ways, their everyday actions presuppose habits of conduct,
roles and institutions current in their societies, endorsed sometimes
after serious reflection, often accepted without much thought. With the
exception of those afflicted with psychological disruptions that pro-
foundly limit their cognitive capacities or that cut them off from their
fellows, we are all embedded in the ethical project.
Yet for ordinary people, and for philosophers too, the status of our
ethical judgments and practices is hard to fathom. What exactly do we
mean when we praise someone for a correct decision? How could an
evaluation like that be grounded? Bertrand Russell famously described
mathematics as a subject in which “. . . we never know what we are talking
about, nor whether what we are saying is true.”1 The impressive ability of
1. “Recent Work on the Principles of Mathematics,” International Monthly, 4, 1901, 84.

I shall not venture into the controversy about how seriously Russell intended his
characterization.
1
2 the ethica l project
mathematicians to reach agreement on conclusions that endure fosters

confidence in their power to acquire knowledge, despite the mysteries
swirling around the content and grounds of their judgments. By con-
trast, the persistence of ethical debate reinforces a sense of unease about
the status of ethical practice, often leading those engaged in public con-
troversies to shy away from “value judgments,” as if any hope of reaching
consensus were absurd.
One popular view of the ethical life persists. Many people believe, with
Dostoyevsky’s Ivan Karamazov, that if ethical precepts were not grounded
in God’s commands, anything would be permitted. From Plato on,
however, the philosophical tradition has frequently—and cogently—
questioned the idea of a religious foundation for ethics. Supposing ethics
to be grounded in the divine will remains popular because alternatives,
including the philosophical alternatives, appear so elusive and uncon-
vincing. Could ethical correctness really consist in representing some
independent realm of values? Could ethical judgments really express
particular privileged emotions? Could they really be arrived at by fath-
oming the “moral law within” or by apprehending the deliverances of
practical reason?
More than a century ago, Darwin outlined a novel way of thinking
about the living world: his fundamental insight was to regard the organ-
isms around us as products of history. We can liberate ourselves from
mysteries about many of our current practices by emulating Darwin:
think of them, too, as historical products.2 The aim of this book is to
pursue this program in the case of ethics. Ethics emerges as a human
phenomenon, permanently unfinished. We, collectively, made it up, and
have developed, refined, and distorted it, generation by generation. Eth-
ics should be understood as a project—the ethical project—in which we
have been engaged for most of our history as a species.
2. Darwin himself made some first efforts to inaugurate this program in the early chap-
ters of The Descent of Man (John Murray, 1872). Thomas Kuhn proposed a similar under-
standing of the natural sciences as historical products (The Structure of Scientific Revolu-
tions [Chicago: University of Chicago Press, 1962 and 1970]). I have followed their lead,
both for the sciences and for mathematics (in The Advancement of Science [New York: Oxford
University Press, 1993] and in The Nature of Mathematical Knowledge [New York: Oxford
University Press, 1983]).
Introduction 3
The position to be elaborated—pragmatic naturalism, to give it a

name—envisages the ethical project as begun by our remote ancestors,
in response to the difficulties of their social life. They invented ethics.
Successive generations have amended the ethical legacy transmitted
to them, sometimes, but by no means always, improving it. Doubtless,
many traditions have died out, but some have continued into the pres-
ent, forming the bases of the ways in which people today regulate their
conduct. In principle, but not in practice, it would be possible to con-
struct an evolutionary tree, drawing a diagram like the single figure Dar-
win inserted into the Origin, with the important difference that the
connecting lines would represent cultural rather than biological descent
(so the picture might show fusion, as well as separation of lineages). 3
As the name suggests, pragmatic naturalism has affi nities with both
pragmatism and naturalism. In focusing on ethical practice and its his-
tory, it attempts to honor John Dewey’s call for philosophy to be recon-
nected with human life.4 Further, it articulates a Deweyan picture of
ethics as growing out of the human social situation; its conception
of ethical correctness is guided by William James’s approach to truth.5
The naturalism consists in refusing to introduce mysterious entities—
“spooks”—to explain the origin, evolution, and progress of ethical prac-
tice. Naturalists intend that no more things be dreamt of in their philoso-
phies than there are in heaven and earth.6 They start from the inventory
of the world allowed by the totality of bodies of well-grounded knowl-
edge (the gamut of scholarly endeavors running from anthropology and
art history to zoology), and, aware of the certain incompleteness of the
list, allow only such novel entities as can be justified through accepted
3. Charles Darwin, On the Origin of Species (facsimile of the fi rst edition), edited by
Ernst Mayr (Cambridge, MA: Harvard University Press, 1964), between 116 and 117.
4. See The Quest for Certainty, vol. 4 in John Dewey: The Later Works (Carbondale:
Southern Illinois University Press, 1988), 204; and Reconstruction in Philosophy (Boston:
Beacon Press, 1957), 147.
5. See Dewey and James Tufts, Ethics, 2nd ed. (New York: Henry Holt, 1932), 307– 9;
William James, Pragmatism (Cambridge, MA: Harvard University Press, 1978), Lecture
VI.
6. The inversion of Hamlet stems from Nelson Goodman: Fact, Fiction, and Forecast
(Indianapolis, IN: Bobbs-Merrill, 1956).
methods of rigorous inquiry. Appeals to divine will, to a realm of values,

to faculties of ethical perception and “pure practical reason,” have to go.
Pragmatic naturalism engages with the religious entanglement of
ethics more extensively than is usual in secular philosophical
discussion—for the pragmatist reason that the entanglement pervades
almost all versions of ethical life. Yet, in accordance with its naturalist
scruples, it cannot maintain the image favored by those who would
ground ethics in the divine will. As we shall see (§27), there are powerful
reasons to suppose, even if there were any deity, ethics could not be fi xed
by its (his? her?) tastes. More fundamentally, pragmatic naturalism
maintains that, when religion is understood as a historically evolving
practice, it is overwhelmingly probable that all the conceptions of a tran-
scendent being ever proposed in any of the world’s religions are false.
For the conceptions introduced in the various religions are massively
inconsistent with one another. Each supernaturalist view rests on epis-
temically similar grounds—typically there was some revelation, long ago,
that has been carefully transmitted across the generations to the devout of
today—yielding a condition of complete symmetry. Under these circum-
stances, no believer has any basis for thinking only he and his group are
privileged to know the truth about the transcendent realm, while others
live in primitive delusions. Further, serious inquiries into the ways in
which canonical scriptures are constructed, into the evolution of reli-
gions, into the recruitment of converts, into the phenomena of religious
experience, demonstrate how radically unreliable are the processes that
have yielded the current corpora of belief. Nor can one isolate some core
doctrine, shared by all religions, something capable of being viewed as a
shared insight. If there are beings of a hitherto unrecognized sort, ap-
proximating some idea of the “transcendent,” we have every reason to
think we have absolutely no clues, or categories, for describing them.7
Religious entanglement in ethical practice is no accident. As we shall
see (§17), appealing to gods as “guardians of morality” can bring social
benefits. Nevertheless, that appeal has distorted the ethical project. Un-
doing the distortions is not simply a matter of eradicating religion, hack-
7. The considerations bluntly advanced here are elaborated at much greater length in
the last chapter of my Living with Darwin (New York: Oxford University Press, 2007).
Introduction 5
ing out the places where false belief has intruded. A secular renewal of
the ethical project requires constructive work, positive steps going
beyond brusque denial.8
Given these clarifications, I can now explain the structure of the fol-
lowing chapters. Part I, Chapters 1–4, elaborates an “analytical history,”
aimed at providing insight into the evolution of our ethical practice. It
provides a basis on which Part II (Chapters 5–7) can explore questions
about ethics: given this account of the origins and unfolding of ethics,
can we make sense of ethical truth or ethical knowledge? The history of
Part I and the metaethical account of Part II are then extended, in Part
III (Chapters 8–10), into a normative stance, an attempt to suggest how
we might best go on from where we are.
It is worth supplementing this bald characterization with a little more
detail. A “history of ethical practice” might take many forms, and the
one I offer may initially appear strange. Since I suppose our species to
have been engaged in the ethical project for tens of thousands of years, it
would be hopeless to offer a narrative showing how particular aspects of
ethical life have gradually emerged. Until the invention of writing (five
thousand years ago), the clues are fragmentary, far scantier than the fos-
sil record, whose poverty provoked Darwin’s lament.9 Primatology, an-
thropology, and archeology enable us to offer a plausible account of the
conditions under which our preethical ancestors lived, but many sub-
sequent steps are beyond our evidential grasp.
The analytical history starts by attempting to understand relevant
psychological capacities of the preethical ancestors, and, on that basis,
to portray the initial stages of the ethical project. Hominid social life was
akin to the contemporary lives of our closest evolutionary relatives: our
precursors lived in small groups, mixed by age and sex. For that, they
needed a capacity for psychological altruism. Yet the limitations of their
altruistic dispositions made living together tense and difficult. The first
ethicists overcame some of the problems by agreeing on rules for conduct,
8. See the closing pages of Living with Darwin, as well as “Challenges for Secularism,”
in The Joy of Secularism, ed. George Levine (Princeton, NJ: Princeton University Press,
2011).
9. Darwin, Origin of Species, chap. 9, esp. 310–11.
rules remedying a few of the recurrent altruism failures that had plagued
their group life. Very probably, they began with precepts about sharing
scant resources and not initiating violence.
Because the character of early ethical practices is so much simpler
than the forms of ethical life visible once written documents are avail-
able, it is important to show (Chapter 3), how a series of gradual steps
might have taken the ethical project from its relatively crude initial phase
to the complex articulation of rules and stories found in the first written
documents. Thereafter, it is possible to trace, although not with the
completeness one might hope for, how actual changes in ethical practice
have occurred. Chapter 4 considers a few examples from history (rather
than prehistory) with the aim of supporting two main theses. First, it is
hard to resist the recognition of occasional progress in the evolution of
ethics: perhaps ethical progress is rare, but there are transitions (like the
repudiation of slavery) in which it seems to occur. Second, even when
the records kept by people who participated in apparently progressive
ethical change are most extensive, moments of ethical discovery are elu-
sive: there are no analogs of episodes of scientific insight.
The history of Part I offers hypotheses about how the ethical project
actually began, and how, in recent history, it has actually gone. It also
addresses concern about the vast difference between the early stages of
the project and the rich practices found at the dawn of history by show-
ing how it would have been possible for the bare beginnings to evolve, by
gradual steps, into the complex systems discernible in the earliest texts.
Because the differences in these two modes of explanation need to be
clearly appreciated, the next section will address some methodological
preliminaries.
How can any history, however carefully focused and articulated, bear
on philosophical questions about ethics? One possibility, already illus-
trated by the example of religion, is that a historical account might un-
dermine current practice. Seeing where our approaches have come from
could breed skepticism and disillusionment. In those episodes of ethical
change most susceptible to analysis, the participants do not appear to
apprehend some previously unrecognized value, or to reason their ways
to some novel moral principle. Historical detail, to the extent it can be
provided, is inhospitable to philosophical theories about ethical truth
Introduction 7
(Chapter 5). Yet the history of Part I also reveals examples of ethical
progress. The metaethical perspective of Part II centers on trying to
reconcile these points.
A “mere change” view of ethical evolution, in which the history is
simply one damned thing after another, conflicts with the pull to charac-
terize some transitions as advances. Chapter 6 resolves the conflict by
seeking an account of ethical progress, one that abandons the idea of
progress as accumulation of (prior, independent) truth. If this appears a
strange idea, we should recall that, in some areas of human practice, prog-
ress does not consist in the increase of truth. Technological progress is
often a matter of discharging certain functions more efficiently or more
fully. Moreover, in line with the history of Part I, the initial ventures in
the ethical project are readily conceived as introducing a new—social—
technology, aimed at remedying disruptive altruism failures.
Amelioration of altruism failure was the initial function of ethical
practice. Yet the obvious differences between the pioneering ventures
and the complex codes present at the dawn of recorded history show
clearly that other functions have emerged. That is the way with technol-
ogy in general. People begin with a problem and achieve partial suc-
cesses in solving it. The successes generate new problems to be solved.
Chapter 6 attempts to anchor the concept of ethical progress in the dis-
charging of functions, originating with the problem of remedying fail-
ures of altruism, and understanding later functions as generated from
the solutions previously obtained.
It thereby paves the way for a concept of ethical truth. Ethical truths
are those acquired in progressive transitions and retained through an
indefinite sequence of progressive transitions. Pragmatic naturalism pro-
poses that some ethical statements—typically, vague generalizations, com-
mending honesty and disavowing violence, for example—are true. They
owe their truth to the role they play in ethical progress: “truth happens
to an idea.”10
To declare that our ancestors invented ethics is to deny that they dis-
covered it or that it was revealed to them. Pragmatic naturalism rejects
the idea of a special moment (long ago on Mount Sinai, perhaps) when
10. James, Pragmatism, 97.

people received authoritative information about how they should live,

and also abandons surrogate philosophical theories about external con-
straints discovered by special faculties. Yet to declare that ethics is a
human invention is not to imply it was fashioned arbitrarily. The ethical
project began in response to central human desires and needs, arising
from our special type of social existence.
There is an obvious concern, one probably already exercising any
reader suspicious about misadventures in naturalism. Why should the
ethical project, even at its most “progressive,” have force on those people
who appear late in its evolution? Critics have often charged that natural-
ism commits a fallacy, and those criticisms need to be addressed. Chap-
ter 7 considers a number of versions of the accusation, attempting to
show that pragmatic naturalism has resources equal to those of any non-
naturalistic rival. Yet one important version persists. Because the ethical
project generates new functions, not necessarily in harmony with one
another, it appears to leave open the possibility of different ways of con-
tinuing. To settle worries about radical disagreement, to finish articulat-
ing the metaethical perspective, a normative stance is needed. That is
the work of Part III.
Parts I and II portray the ethical project as an enterprise in which
people work out how to live together. It began without presupposing any
sources of truth ethical deliberators sought to fathom—whether those
sources lay in a divine will or in any of the philosophical substitutes.
Our ancestors needed to fashion their shared life by conversing on equal
terms. Pragmatic naturalism denies ethical expertise. The role philoso-
phy plays in ethics can be one only of midwifery: to suggest a direction for
renewed conversation and some rules for mutual exchange. Chapters 8
and 9 offer a package of proposals, an egalitarian conception of the good
at which we should now aim, together with a method of deliberating
under conditions of mutual engagement.
Convincing proposals come with some form of support. Pragmatic natu-
ralism’s proposals are motivated by conceiving the current human situation
as analogous to that initially prompting the ethical project. As it was in the
beginning, so too now—for the conflicts to which our ancestors’ lives were
subject are mirrored in contemporary hostilities across the human popula-
tion. According to this vision, the original function of ethics—to remedy
Introduction 9
altruism failures—remains primary. Challenging the enduring impor-

tance of this function fails to achieve an important form of coherence, one
pragmatic naturalism attains. So (§56) the deepest challenge—the most
important accusation against naturalistic ethics—is turned back.
Finally, Chapter 10 attempts further philosophical midwifery, by sug-
gesting some specific places at which current ethical practices might be
amended. It offers ideas for continuing the essential conversation.
Readers familiar with contemporary philosophical discussions about
ethics will recognize the large differences between standard approaches
and the material I plan to cover. Although my questions connect with
those posed constantly in the history of ethics, they are often framed
orthogonally to the preferred philosophical formulations of recent de-
cades (at least in the Anglo-Saxon world). At many points, my treatment
has been influenced by the writings of my peers (and betters), but any
serious attempt to expose connections (especially one attending to the
nuances of intricate positions) would require expanding my discussions
by hundreds of pages. Pragmatic naturalism aims to steer ethical prac-
tice and ethical theory in new directions, and I apologize to those who
would like to see it anchored in the “existing literature.” I must also ask
for patience, if a response to an important worry or objection is post-
poned. Not everything can be done at once, and the needed resources
have to be assembled before they can be applied.
§2. Methodological Preliminaries

Pragmatic naturalism differs from previous attempts to link ethics to our
evolutionary past. It does not propose to identify ethical properties in
evolutionary terms, say, by equating what is good with what is adap-
tive.11 Nor does it suppose ethical practice is already present, at least in
embryo, in our evolutionary cousins or our hominid progenitors.12 The
11. This idea, defended by some sociobiologists, is criticized in the fi nal chapter of my
Vaulting Ambition: Sociobiology and the Quest for Human Nature (Cambridge, MA: The
MIT Press, 1985).
12. For a forthright attempt to link human ethical practice to the altruistic tendencies of
other primate species, see Frans de Waal, Primates and Philosophers (Princeton, NJ:
Princeton University Press, 2006).
ethical project is not simply the unfolding of previously existent altruis-

tic tendencies—it is more than just a population acquiring capacities for
“nice behavior.” Ethical practice involves conversation, with others and
with yourself, juxtaposing desires you recognize as part of you and other
desires you would prefer to move you to action. Neither does it posit a
special evolutionary advance, in which our ancestors acquired a “moral
instinct,” conceived along the lines of our innate capacity for language.13
Views of this type are in danger of confusing ethics with “nice behav-
ior,” and, as we shall see (§14), they underplay the influence of the social
environments in which ethical practice occurs.14
Given these differences, some familiar methodological concerns about
naturalistic ethics do not bedevil pragmatic naturalism. One major worry
does arise, however. Because it is important, and because appreciating
its significance could easily provoke objections to discussions in Part I,
this section endeavors to forestall it.
Darwin’s success in applying his historical method to the living world
rested on the immense body of evidence amassed in the Origin—even
hostile reviewers praised him for “his facts.” A familiar criticism of later
attempts to apply evolutionary ideas to human behavior and to human
social life charges Darwin’s imitators with failing to live up to the stan-
dards set by the master. Evolutionary explanations, it is suggested, run
the risk of becoming exercises in storytelling—just-so stories without
Kipling’s wit.15 A possible hypothesis for the evolution of some trait is
proposed, and, without seriously considering alternatives, discriminat-
13. A view well articulated and defended by Marc Hauser: Moral Minds (New York:
Ecco, 2006).
14. Finally, pragmatic naturalism also rejects the thought that significant advances can
be made in understanding ethical issues by undertaking psychological, or neurological,
experiments in which subjects are asked to respond to abstract philosophical scenarios. It
is unclear what capacities are being fathomed in posing the questions: for questions and
concerns that would arise in everyday life are artificially excluded. Moreover, pragmatic
naturalism looks for an alternative to current ethical theorizing, rather than for an experi-
mental extension building on available options.
15. The locus classicus for this accusation is Stephen Jay Gould and Richard Lewontin,
“The Spandrels of San Marco and the Panglossian Paradigm: A Critique of the Adapta-
tionist Programme,” Proceedings of the Royal Society, B, 205, 1979, 581– 98. In Vaulting
Ambition, I argued that storytelling vitiates much work in human sociobiology.
Introduction 11
ing them in light of the evidence, it is adopted as describing the actual

course of past events.
Reconstructing the actual history of the ethical project, from its begin-
nings to the present, is plainly beyond the evidence available—and proba-
bly beyond the evidence anyone could ever hope to obtain. On the account
sketched in §1, and developed in Part I, many important changes occurred
in the Paleolithic and early Neolithic, in a period lacking any written doc-
uments. Visions of human life at that time can be based on just a few tan-
talizing clues: deposits showing increasing group size, tools found at a far
remove from the closest source of material, burial sites, figurines, and
cave art. These data are too sparse to screen out rival hypotheses about
the sequence of events leading from the beginnings of the ethical project
to the complex form in which it appears in the first written texts we have.
Yet some hypotheses about the actual history can be defended. Ar-
cheological evidence provides grounds for thinking that, until about fif-
teen thousand years ago, human beings lived in groups of roughly the
size of contemporary bands of chimpanzees, and, like the societies of our
evolutionary cousins, these groups were mixed by age and sex. Combin-
ing this conclusion with anthropological studies of living people who are
closest to the circumstances of our ancestors, researchers have provided
a picture of the social context of our forebears, the context in which they
began the ethical project. We know what the problems faced by chim-
panzees are, and how contemporary bands of hunter-gatherers overcome
them. The first ventures in ethical practice probably involved group
discussions, on terms of rough equality, directed toward issues of shar-
ing and intragroup aggression.
Those initial efforts likely occurred tens of thousands of years ago,
after human beings acquired full linguistic abilities. At a conservative
estimate, we can set the age of the ethical project as fifty thousand years.
I conclude that roughly the first forty thousand years of the project were
directed toward the needs of small groups, whose members worked out
their social lives on terms of approximate equality. Their original rules
were crude and simple. Out of the social life they permitted came a se-
quence of dramatic changes, generating societies of greater size, and
eventually the hierarchical cities of Mesopotamia and Egypt. That part
of the story, I shall argue, can be defended as actual history.
How did our ethical ancestors move from those simple beginnings to
the complex forms of ethical life recognizable in the world of ancient his-
tory? No answer can claim to tell the actual story. Here pragmatic natu-
ralism must face the difficulty of discriminating a preferred story from
potential rivals. It is, however, committed to supposing some sequence
of transitions led—without revelations, without discoveries of the struc-
ture of values or the moral law, without any “spooks”—to an endpoint
enormously richer and more complex than the original practices. To an-
swer skeptics claiming that “real ethics” requires resources naturalists
cannot allow, some narrative needs to be given. It cannot be advertised as
a “how actually” explanation; instead it is a “how possibly” explanation.
Explanations come in many varieties, but for the purposes of this
book, these two types will suffice. A historical “how actually” explana-
tion aims to tell the truth about a sequence of events: if it is properly sup-
ported, rival options have to be eliminated by the evidence. A historical
“how possibly” explanation, by contrast, aims only to tell a story, con-
sistent with the evidence and with background constraints: its status is
not impugned by pointing out that there are other options (the more, the
merrier). A “how possibly” explanation is important because we some-
times wonder whether a chain of occurrences could have occurred, or
whether the occurrence of the sequence is permitted by a particular
theory. Opponents wonder, for example, if the processes countenanced
by Darwin and his successors allow for the evolution of the cell. Answer-
ing their doubts requires showing how Darwinian processes might have
produced the cell. It would be marvelous, of course, to be able to say
how the history actually went, but, given the temporal remoteness of the
events and the limitations of our evidence, modesty is required. In the
context of rebutting the skeptical challenge, modesty—settling for “how
possibly”—is enough.
Pragmatic naturalism can advance probable hypotheses about the
original state in which the ethical project began, and about the character
of the evolution of the project during recorded history. With respect
to the transformations that occurred between the early phases and the
practices of the ancient world, all that can be claimed is that these could
have happened without supposing processes or causes of kinds prag-
matic naturalism rejects. The history of Part I is self-conscious about the
Introduction 13
distinction. I have attempted to be clear about the kinds of explanations

at which particular discussions aim—and hence the standards they are
expected to satisfy. In consequence, the metaethical perspective of
Part II can sometimes build on claims about how a process unfolded
(delivered by “how actually” explanations), and sometimes only on the-
ses that par ticu lar initial states evolved into par ticu lar fi nal states
(defended against skeptical challenge by “how possibly” explanations).
Explicitness may be pedantic—but pedantry is probably better than
arousing suspicion that proper standards have been violated.
The aim is to use history—in the ways, and to the extent, we can recon-
struct it—to liberate discussions of ethics from the confining pictures that
prompt a sense of mystery. Let us now turn to the work of reconstruction.
pa rt on e
An Analytical History
chapter 1
The Springs of Sympathy
§3. Psychological Altruism: Basics

At some point in our evolutionary past, before the hominid line split off
from the branch that leads to contemporary chimpanzees and bonobos
(possibly quite a long time before), our ancestors acquired an ability to
live together in small groups mixed in terms of sex and age. That
achievement required a capacity for altruism. It also prepared the
way for unprecedented forms of cooperation, and ultimately for the
enunciation of socially shared norms and the beginnings of ethical
practice. Altruism is not the whole story about ethics, but it is an im-
portant part of it.1
My analytical history of the ethical project thus begins with a hypoth-
esis about the social groups in which the project originated and about the
1. There is a long tradition, stemming from Hume, Adam Smith, and Schopenhauer,
that places a capacity for sympathy at the center of ethics. In recent years, that tradition has
been renewed by phi losophers (Simon Blackburn, Ruling Passions [New York: Oxford
University Press, 1999]) and by primatologists (Frans de Waal, Primates and Philosophers
[Princeton, NJ: Princeton University Press, 2007]). Although the approach I shall defend
overlaps with some of the themes of this tradition, it does not ascribe sympathy (or altru-
ism) so dominant a role. For explicit comparisons, see my discussion of de Waal, “Ethics
and Evolution: How to Get Here from There,” in Primates and Philosophers.
17
psychological capacities of the members of those groups. Fossil evidence,

together with the remains found at hominid and early human sites, re-
veals that our ancestors lived in bands akin to those in which chimpan-
zees and bonobos live today: the members were young and old, male
and female; the band size was (roughly) 30–70.2 This chapter argues
that, to live in this way, hominids and human beings had to have a ca-
pacity for altruism, one contemporary people almost certainly retain.
To understand the historical unfolding of ethics we shall need to recog-
nize the intricacies of the notion, as well as the varieties and limitations
of hominid/human altruism. The next sections supply the necessary
preliminaries.
It is important to distinguish three types of altruism. An organism A is
biologically altruistic toward a beneficiary B just in case A acts in ways that
decrease its own reproductive success and increase the reproductive suc-
cess of B. For a century after Darwin, there was a deep puzzle about how
biological altruism is possible. During the past fifty years, however, that
puzzle has been solved. Biologically altruistic actions directed toward kin
can promote the spread of the underlying genes. Moreover, when organ-
isms interact with one another repeatedly, biological altruism exhibited
on some occasions can gain dividends from future reciprocation.3
2. Different anthropologists use different methods for estimating hominid group size,
some favoring direct comparisons with social groups in other species (either evolution-
ary relatives or primates with a similar ecology), others taking extant hunter-gatherer
bands as models or seeking correlations with mea sur able anatomical features (e.g., skull
size) and extrapolating from the results on hominid skulls (viewed as providing clues to
the relative increase in neocortex size). See Robin Dunbar, Grooming, Gossip, and the
Origins of Language (London: Faber, 1996); Steven Mithen Pre- History of the Mind
(London: Thames and Hudson, 1996); Christoph Boehm, Hierarchy in the Forest (Cam-
bridge, MA: Harvard University Press, 1999); Clive Gamble, The Palaeolithic Societies
of Europe (Oxford, UK: Cambridge University Press, 1999); and Peter MacNeilage, The Ori-
gin of Speech (Oxford University Press, 2008). Although I am inclined to accept a relatively
small value (30–70), my conclusions would not be greatly affected were this increased to,
say, 80–140.
3. The original papers are W. D. Hamilton, “Genetical Evolution of Social Behavior,” I,
II, Journal of Theoretical Biology 7 (1964): 1–52; Robert Trivers, “The Evolution of Recip-
rocal Altruism,” Quarterly Review of Biology 46 (1971): 35–57; Robert Axelrod and William
Hamilton, “The Evolution of Cooperation,” Science 211 (1981): 1390– 96. Lucid and acces-
sible summaries are available in Richard Dawkins, The Selfi sh Gene, 2nd ed. (New York:
The Springs of Sympathy 19
Biological altruism requires no perceptive or cognitive abilities. Even

plants can have traits that make them biologically altruistic, for their
propensities to form roots or to set seeds can limit individual reproductive
success and facilitate the reproduction of neighbors. For animals capable
of recognizing the wishes of those around them, however, we can develop a
useful behavioral analog of the notion of biological altruism.4 An animal A
is behaviorally altruistic toward a beneficiary B just in case A acts in ways
that detract from its fulfillment of its own current desires and that promote
the perceived wishes of B.5 Behavioral altruists do what they take the ani-
mals around them to want. They may act in this way not out of any particu-
lar concern for those other animals, but because they think that some of
their own wishes will ultimately be well served by doing as they do. Behav-
ioral altruism may be practiced by Machiavellian egoists (and, as we shall
eventually see—§11—it can also be practiced by individuals who fall into a
category intermediate between egoism and psychological altruism).
Neither biological altruism nor behavioral altruism is of much help in
understanding the origins of the ethical project. For our purposes, the
significant notion is that of psychological altruism. Psychological altru-
ism has everything to do with the intentions of the agent and nothing to
do with the spread of genes, or even the successful satisfaction of the
wishes of others. Assuming for the moment that there have been human
beings who are psychological altruists, the vast majority of them have
not known much about heredity, and even those who have were rarely
concerned with spreading genes. They acted to promote what they took
to be the wishes, or the interests, of other people.6 Sometimes they
succeeded. Yet, even when they did not, their serious efforts to do so
qualified them as psychological altruists.
Oxford University Press, 1993); and Robert Axelrod, The Evolution of Cooperation (New
York: Basic Books, 1984). I shall be exploring these important ideas in §8.
4. For discussions about behavioral altruism, I am indebted to Christine Clavien.
5. There are complications that I glide over here and that will be addressed more
thoroughly in treating the third type of altruism, the one pertinent to the examination of
ethics. After the presentation of that third notion, it will be easier to see how to charac-
terize behavioral altruism more exactly.
6. As the specification of psychological altruism will show, the account begins with
wishes. Interests come later (§21).
Many people believe psychological altruism does not exist, even that
it is impossible. Often they are moved by a very simple line of reasoning:
when a person acts in a way that could be appraised as altruistic, he or
she acts intentionally; to act intentionally is to identify an outcome one
wants and to attempt to realize that outcome; hence, any potential altru-
ist is trying to get what he or she wants; but to strive for what you want is
egoistic; consequently, the potential altruist turns out to be an egoist
after all. The key to rebutting this argument is to distinguish different
kinds of wants and goals. Some of our desires are directed toward our-
selves and our own well-being; other desires may be directed toward the
welfare of other people. Desires of the former type are the hallmark of
egoism, but those of the latter sort are altruistic. So altruists are inten-
tional agents whose effective desires are other-directed.7
I shall develop this approach to psychological altruism further, by
giving a more detailed account of the character of other-directed desires,
and thereby bringing into the open some of the complexities of the con-
cept of altruism. In focusing on desires, I ignore for the moment the fact
that there are other psychological attitudes—hopes, aspirations, and
particularly emotions—that can be properly characterized as altruis-
tic. Attention to these other types of states will occupy us in the next
section. Because of the connection of desires with intentions and ac-
tions, altruistic desires have a certain priority. They are thus the topic
of the basic account.
The other-directed desires central to the defense of the possibility of
altruism are desires that respond to the altruistic agent’s recognition of the
impact of his or her actions on the situations of others. To be an altruist
is to have a particular kind of relational structure in your psychological
life—when you come to see that what you do will affect other people, the
wants you have, the emotions you feel, the intentions you form, change
from what they would have been in the absence of that recognition. Be-
cause you see the consequences for others of what you envisage doing,
the psychological attitudes you adopt are different. You are moved by the
7. This line of response surfaces in the eighteenth century in the famous series of ser-
mons given by Joseph Butler at the Rolls Chapel. Many subsequent writers have followed
Butler’s lead—as shall I.
perceived impact on someone else. If your response leads you to act al-
truistically, that is because your desires have been affected.8
So far, that is still abstract and vague. I shall motivate the underlying
idea with a simple and stylized example and then offer a more precise
definition.
Imagine that you are hungry and that you enter a room in which
some food is spread out on a table. Suppose further that there is nobody
in the vicinity who might also be hungry and want all or part of the food.
Under these circumstances, you want to eat the food; indeed, you want
all of it. If the circumstances were slightly different, however, if there
were another hungry person in the room or believed to be in the neigh-
borhood, your desire would be different: now you would prefer the out-
come where you share the food with the other person. Here your desire
responds to your perception of the needs and wants of someone else, so
that you adjust what you might otherwise have wanted to align your de-
sire with the wants you take the other person to have.
This is a start, but it is not sufficient to make you an altruist. For you
might have formed the new want when you see that someone else will be
affected by what you do, because you saw profitable future opportunities
for accommodating this other person. Maybe you envisage a series of oc-
casions on which you and your fellow will find yourselves hungry in
food-containing rooms. You see the advantages of not fighting and of not
simply having all the food go to the first person who enters. You resolve
to share, then, because a future of cooperation will be better from your
point of view. For real altruism, the adjustment of desires must not be
produced by this kind of self-interested calculation.
I offer a definition of “A acts psychologically altruistically towards B in
C”—where A is the agent, B is the beneficiary, and C is the context (or
set of circumstances). The first notion we need is that of two situations
differing from each other in the recognizable consequences for others
(people or nonhuman animals). Let us say, then, that two contexts C and
8. You might be affected by another person’s predicament, and form an altruistic emo-
tion, but that might not generate a desire that issues in action. The most basic type of altru-
ism that is of ethical concern is a response to someone else that eventually expresses itself
in conduct.
C* are counterparts, just in case they differ only in that, in one (C*, say)
the actions available to A have no perceived consequences for B, whereas
in the other (C) those actions do have perceived consequences for B. C*
will then be the solitary counterpart of C, and C will be the social coun-
terpart of C*. If A forms different desires in C* from those A forms in C,
the set of desires present in C* will be A’s solitary desires (relative to the
counterparts C and C*). Given these preliminary specifications:
A acts psychologically altruistically with respect to B in C just in

case
(1) A acts on the basis of a desire that is different from the desire
that would have moved A to action in C*, the solitary
counterpart of C.
(2) The desire that moves A to action in C is more closely
aligned with the wants A attributes to B in C than the desire
that would have moved A to action in C*.
(3) The desire that moves A to action in C results from A’s
perception of B’s wants in C.
(4) The desire that moves A to action in C is not caused by A’s
expectation that the action resulting from it would promote
A’s solitary desires (with respect to C and C*).
Condition 1 tells us that A modifies his or her desires from the way they
would otherwise have been, when there is an impact—more accurately,
when there is a perceived impact9—on the wants of B. Condition 2 adds
the idea that the desire, and the behavior it directs, is more in harmony
with the wants attributed to B than it would have been if B were unaf-
fected by what was done. (It is possible to modify your desires in response
to the perceived wishes of another, but to do so in a way that diverges
from their perceived wants—that is spite.) Condition 3 explains that the
increased harmony comes about because of the perception of B’s wants;
it is not, say, some caprice on A’s part that a different desire comes into
play here. Finally, condition 4 denies that the modification is to be un-
9. I shall consider cases in which agents have mistaken beliefs later. For the time being,
I suppose that the parties get things at least roughly right.
derstood in terms of A’s attempt to promote some desire that would have
been present in situations where there was no thought of helping or hurt-
ing B; this distinguishes A from the food sharer who hopes for returns on
future occasions when B is in the position of disposing of the goods.
Condition 4 requires that genuine psychological altruists be different
from Machiavellian calculators who aim to satisfy the wants they would
have in solitary situations (I shall sometimes refer to condition 4 as the
“anti-Machiavelli” condition).
Given this account of psychological altruism, it is now possible to
characterize behavioral altruism more carefully. Behavioral altruists are
people who look like psychological altruists. That is, they perform the
actions people with psychologically altruistic desires would have been
led to perform. In ascribing behavioral altruism, however, we do not sup-
pose any particular psychological explanation of the actions. Perhaps
they are indeed the products of psychologically altruistic desires, or per-
haps the actions are produced by quite different desires having nothing
to do with the satisfaction of the beneficiary—a desire for status, or for
feeling oneself in accordance with some socially approved pattern of con-
duct, or even a self-interested calculation. (We shall explore some
possibilities of behavioral altruism later; see §§7, 11.)
The stylized food example allows the introduction of an obvious con-
cept, one that will be important in future discussions, and that further
articulates the account of altruistic desires. The altruistic modification
of solitary desires can be more or less intense. I have spoken—somewhat
vaguely—of the altruist as aligning his or her wants with those attributed
to the beneficiary.10 That alignment is often a matter of degree, for
example, when there is a continuum of possibilities intermediate between
complete egoism (retaining one’s solitary desires in the social counter-
part) and complete subordination of one’s solitary wishes to those one
perceives the other to have (where one comes to want exactly what
one perceives the other as desiring). In sharing food, this is easily
10. For a more precise and formal discussion of many aspects of altruism, see my essays
“The Evolution of Human Altruism,” Journal of Philosophy 90 (1993): 497–516, and “Vari-
eties of Altruism,” Economics and Philosophy (2010): 121–148. As I shall note at various
places, there are several aspects of the account of altruism provided in this chapter that can
be treated mathematically, and these articles make a start on that.
expressed in terms of the mode of division: egoists give nothing, self-

abnegating altruists give everything, and in between lie a host of inter-
mediate altruists. One obvious style of altruism is golden- rule altruism,
distinguished by its equal weighing of the solitary desires and those at-
tributed to the beneficiary.
Inspired by the food example, we can undertake a simple way of rep-
resenting the intensity of psychological altruism, one that will be useful
in some (but by no means in all) instances. Suppose that people’s desires
can be represented by (real) numbers that correspond to how much they
value a given outcome. If one result, eating all the food, say, is worth 10
to me, and another, eating half the food, is worth 7, then I prefer eating
everything to eating half, but I also prefer an assured outcome in which I
receive half to the state of being awarded all or nothing dependent on the
flip of a fair coin. (For, in the latter case, my expected return is measured
by 5—half of 10 plus half of 0—which is less than 7.)
When you are in the picture, I also take into account the values you
attribute to various outcomes. My social desire could be represented
as a weighted average of the values represented in my solitary desires
and those I take to measure your solitary desires. Thus, the numbers as-
signed in my social desires would be given by the simple equation:
vSoc = wEgovSol + wAltvBen
where vSoc measures my social desires, vSol my solitary desires, vBen the
measurements of desire I attribute to the beneficiary (you), and wEgo and
wAlt the weights given to my solitary desires and my attributions of desire
values to you (so that wEgo + wAlt = 1). The intensity of my altruism is rep-
resented by the size of wAlt —and hence inversely by the size of wEgo ; if
wEgo = 1 (wAlt = 0), then I am, at least with respect to you on this occasion,
a psychological egoist; if wAlt = 1, then I am a self-abnegating altruist; if
wAlt = 0.5 (= wEgo), then I am a golden-rule altruist.
We should not assume that all types of altruistic alignment with the
wishes of others can be conceived in this very simple way. Cases of shar-
ing show that a simple approach sometimes works, and the simple ex-
pression of social wants as weighted averages will be useful in explaining
and illustrating some of the ideas of later sections.
§4. The Varieties of Altruistic Reactions

As already recognized, altruism is not always about the modification
of desire, though we are often reasonably suspicious about alleged ex-
amples of altruism that do not change desires in ways leading to action:
it is not enough simply to “feel another’s pain.” We can be moved to share
the hopes of others, to modify our own long-term intentions and aspira-
tions to accommodate what we see them as striving for, and, most impor-
tant, we can feel different emotions because of our awareness either of
what they feel or of the situations in which they find themselves. For
some kinds of psychological states, hopes and long-term intentions, for
example, accounts of altruistic versions of these states can be generated
straightforwardly in parallel to the treatment of the previous section.
Emotions, however, deserve special consideration, both because they are
frequently components of the psychological attitudes with which we shall
be concerned, and because they involve types of reactions more broadly
shared among animals than the psychological states on which I have so
far concentrated.
Altruistic emotional responses to others might be—and probably of-
ten are—mediated by perception and cognition. We see that another
person is suffering—or jubilant, for altruistic emotions are not always
dark—and our own emotional state changes to align itself more closely
with that attributed to the other. Or, in a different mode of altruistic re-
sponse, we understand the situation in which another person is placed,
and our emotional state changes to take on, to some extent, the feeling(s)
we would have if placed in that state.11 When people, or other animals,
have dispositions to modify their emotional states in light of their under-
standing of the feelings or the predicaments of others, we can treat
emotional altruism just as §3 analyzed altruistic desire. The emotional
altruist feels one thing in the solitary counterpart and feels differently in
the social counterpart; the emotion in the social counterpart is more
closely aligned with that attributed to the other (or more closely aligned
11. The distinction between these two modes of altruistic emotional response—forms of
“sympathy”—was already clearly recognized by Adam Smith in Theory of Moral Senti-
ments, Knud Haakonssen, ed. (New York: Cambridge University Press, 2002).
with the emotion the altruist supposes he or she would feel if placed in
the other’s shoes), and the alignment comes about because of the recog-
nition of the other’s feelings (or of the other’s situation); finally, it is not
caused by any background solitary emotion or solitary desire. Now,
whereas in the understanding of altruistic desire this last condition re-
sponds to a genuine worry—for we readily think people can form osten-
sibly other-directed desires on the basis of selfish calculations (I can
want to share with you because I think it will be good for me in the long
run)—the anti-Machiavelli condition seems odd and gratuitous in the
emotional case. It is natural to think, and it may even be true, that self-
directed psychological states simply have no power to generate emotions
toward others, that our emotional life is not under that sort of control.
Emotional responses, one may suppose, are caused by processes more
direct and automatic than the perceptions and cognitions figuring in my
analyses. Consequently, an account of emotional altruism parallel to the
analysis of altruistic desire will be at least incomplete, and perhaps even
radically misguided.
This is a serious challenge. To meet it, we shall have to consider, if only
briefly, the character of emotions. Without taking sides in unresolved
controversies, I shall argue that some kinds of emotional response can
be understood along the lines just sketched, while others cannot. An ac-
count of more basic altruistic emotional reactions, or “affective states,”
as I shall call them, provides a valuable supplement to the approach to
psychological altruism begun in the previous section.
Emotions involve changes in our physiology, and some students of
emotion have identified the emotion with the alteration in physiological
state. Others propose that there are important distinctions among emo-
tions that cannot be recognized without supposing those who feel the
emotions to have particular beliefs, desires, and intentions: specific forms
of awareness are required for guilt and shame, for resentment and indig-
nation, and for certain kinds of contentment and anger. A natural way of
responding to the findings of neuroscientists, psychologists, and anthro-
pologists is to suppose that many emotions are complex entities, perhaps
processes in which particular types of physiological conditions are ac-
companied by special kinds of cognitive and volitional states. When some-
one resents the insensitive remarks made by another, he or she undergoes
a physiological response connected to judgments about what has been

said and desires about what will happen next. The causal details of these
connections are matters of speculation, but, even in advance of knowing
them, we can reject an approach to emotions that would leave out either
the physiological or the cognitive/volitional features.12
Yet there may be emotional states, felt by nonhuman animals and by
human infants perhaps, for which the cognitive/volitional component is
negligible, even entirely absent. With respect to our own species, it has
been argued that there are a number of basic emotions, found in all hu-
man societies and typically giving rise to the same facial expressions.13
Although a widespread aspect of human psychology or behavior is often
taken as evidence of some biological (typically genetic) basis that gener-
ates the common feature across all environments, it is worth treading
carefully here. For, trivially, there will be some environments in which
members of our species will not develop so as to exhibit the typical
reaction—neural and psychological development can be disrupted in
many different ways. The interesting questions are whether there are
subtle properties of potential human social environments capable of
prompting something different (so that the widespread fi nding of the
common feature depends on the absence of those subtle properties in the
human societies studied), and whether, if so, the potential environments
are in some way pathological. These questions are not settled, but, for
the sake of the present discussion, I shall allow that human beings who
develop in environments, physical and social, that do not involve dam-
aging disruptions of development, will all share dispositions to basic
affective reactions—that is, they will have capacities for basic affective
states, like disgust and anger and fear, and they will exhibit similar facial
expressions characteristic of the individual affective states.14
12. Here I am much influenced by the thoughtful and ecumenical approach adopted by
Jenefer Robinson in the first three chapters of Deeper than Reason (New York: Oxford
University Press, 2005).
13. See Paul Ekman and Richard Davidson, eds., Nature of Emotion (New York: Oxford
14. Since the role of social environments is central to my approach to our altruistic ten-
dencies and the character of the ethical project, my position would be strengthened if this
concession proves false.
This concession does not entail any of a number of conclusions some-

times drawn from it. First, there is no implication that the affective reac-
tions in people who belong to different societies will be generated by the
same things, events, and states of the world: virtually all people feel dis-
gust, but different groups find very different things disgusting. Second,
to allow basic affective reactions does not retract the earlier judgment that
many emotions have cognitive/volitional components: every emotion
might involve some basic affective state, but a large diversity of emotions
might be distinguished by the cognitions and volitions connected to that
state. Third, and most important for present purposes, we should resist
the thought that “because they are biological,” affective reactions are
based on some mechanism more “immediate,” more “primitive” than
human cognition. It is easy to muddle together two different senses in
which a capacity may be “biological,” one in which its development oc-
curs across all (nonpathological) environments, and one in which the
ways in which it is activated bypass our beliefs and wishes. Conceding
the “biological” status of affective reactions in the first sense does not
commit us to supposing them “biological” in the second.
We can now address the critical issue concerning altruistic emotions.
Even though it is not required that there be affective reactions that do not
depend on the causally prior recognition of the feelings or the predica-
ments of others (on beliefs that they are suffering, for example), it is pos-
sible that there should be mechanisms prompting particular affective re-
sponses, mechanisms not mediated by prior cognition. It is well-known
that infants in the same hospital nursery react to the crying of others: an
initial solo can set off an entire chorus.15 Supposing the unhappiness of
the one spreads to the many because each believes that someone around
is unhappy strains credulity. A more sober account would view the process
as a kind of contagion, effected by a species-typical neural mechanism,
that transfers the misery of one baby to those around. Recent studies of
the activity of so-called mirror neurons (primarily investigated in
macaques) may offer clues about the potential mechanism. Perceptions,
or even sensations, can cause an animal to activate the same neurons as
15. See Martin Hoff man, Empathy and Moral Development (New York: Cambridge
those giving rise to the behavior perceived or causing the sensations: A’s
observation of B’s facial expression produces neuronal firings that tense
the pertinent muscles and that result in A’s imitation of B; perhaps the
sound of another baby crying induces a pattern of neural activity that
mirrors that in the source of the crying and thus causes the originally
contented baby to cry.16 Mechanisms of this sort require a different ap-
proach to altruistic emotions.
Once we have the challenge clearly in view, however, it is not hard to
see how to liberate the account of altruism begun in §3 from its depen-
dence on cognition. The task is to provide a defi nition of “A feels an
altruistic emotion in response to B in C.” As before, we shall suppose
the notions of solitary and social counterparts. The conditions are as
follows:
1. A feels an emotion different from the emotion A would have felt in

C*, the solitary counterpart of C.
2. The emotion A feels is more closely aligned with the emotion A
attributes to B in C than the emotion A would have felt in C*; or it
is more closely aligned with the emotion A would have felt had A
been in B’s position in C; or it is more closely aligned with the
emotion B actually feels in C.
3. If the emotion A feels is more closely aligned with that attributed
to B or if the emotion A feels is more closely aligned with the
emotion A would have felt in B’s position in C, then the emotion A
feels in C results from A’s perception of B’s situation in C; if no
recognition of B’s state plays a causal role and if A’s emotion
aligns itself with that felt by B, then the emotion felt by A is caused
by the operation of some automatic neural mechanism, a mecha-
nism triggered by A’s observation of B (the activation of mirror
neurons might be one such mechanism).
4. The emotion felt by A in C is not caused by A’s expectation that
feeling this emotion would promote A’s solitary desires (with
respect to C and C*).
16. William Damon, The Moral Child (New York: Free Press, 1998); and Hoff man, Em-
pathy and Moral Development.
This account introduces clauses into the second and third conditions
in order to allow the possibility of altruistic emotions produced in
ways that bypass cognition. Although the fourth condition is retained, it
is highly plausible that Machiavellian manipulation of our emotional
lives is beyond our powers, and, if that is indeed so, this requirement is
redundant.
The analysis just given preserves a fundamental feature of my original
characterization of psychological altruism (§3): altruists have a particu-
lar type of relational structure in their psychological lives—when others
are around, the altruist’s desires, hopes, intentions, and emotions are
different from what they would otherwise have been, closer in some way
to those of the others, and the difference is produced by some sort of re-
sponse to those others, not by something enclosed within the self (calcu-
lations of future benefit, for example). What the more complex approach
to altruistic emotions adds is the possibility that the generation of the
response might involve some precognitive mechanism.
It is easy to overinterpret this last point. One might suppose that
affective states are always generated by some mechanism that does not
involve cognition—but, not only do I see no basis for holding so sweep-
ing a generalization, but it also seems belied by the fact that affective re-
actions are often founded in complex and explicit understanding (when
I see pictures of Jewish refugee children being greeted at English ports
by policemen and willing foster parents, I feel a complex mixture of
emotions, surely involving affective states, but these states are clearly
dependent on my conscious understanding of what the photographs
display). The causal relations among affective and cognitive states may
be quite various, and, while we await definitive accounts of them, it is
well to suspend judgment and to be open to many possibilities.
Nor should we suppose that noncognitive mechanisms are inevitably
involved in whatever altruistic responses occur in nonhuman animals.
Although questions about the extent of animal abilities to recognize the
wishes and thoughts of their conspecifics are much debated, there is no
reason to take an advance stand on these issues.17 I shall later defend the
17. For defenses of opposing views, see de Waal, Primates and Philosophers; and Derek
C. Penn and Daniel J. Povinelli, “On the Lack of Evidence that Non-human Animals Pos-
thesis that some of our evolutionary cousins have altruistic desires (in
the sense of §3; see §7) and that similar capacities were shared by our
hominid ancestors.
§5. Some Dimensions of Altruism

One further aspect of psychological altruism needs to be emphasized
before we have all the tools required for probing the hominid preethical
state. On the account of the last sections, there are many varieties of al-
truism. Or, to use a suggestive metaphor, altruism is a multidimensional
notion. For animals capable of psychological altruism, each individual
occupies a particular place in a multidimensional space where brute
(non-Machiavellian) egoism is represented by a single plane, and the
various forms of altruism range over the entire rest of the space.18
An animal’s altruism profile (where he or she is located in altruism
space) is determined by five factors: the intensity of the animal’s re-
sponses to others, the range of those to whom the animal is prepared to
make an altruistic response, the scope of contexts in which the animal is
disposed to respond, the animal’s discernment in appreciating the con-
sequences for others, and the animal’s empathetic skill in identifying the
desires others have or the predicaments in which they find themselves.
Non-Machiavellian egoists never respond to anyone else in any context:
for the dimensions of intensity, range, and scope they score 0, 0, and 0;
their discernment and empathetic skill can be as you please, for these are
never called into play.
Altruists are not like that. They modify their desires and emotions to
align them with the perceived desires and (perceived or actual) emotions
of at least some others in at least some contexts. As §3 already proposed,
their responses may be more or less intense. With respect to altruistic
desires, an altruist may give more or less weight to the perceived desire
of the beneficiary. My treatment of the stylized example in terms of
weighted averaging provides a clear paradigm for intensity—the intensity
sess Anything Remotely Resembling a ‘Theory of Mind,’ ” Philosophical Transactions of

the Royal Society B 362 (2007): 731–44.
18. For more details about this spatial metaphor, see Kitcher, “Varieties of Altruism.”
of altruism is represented by how much of the food you are willing to

relinquish. If
vSoc = wEgovSol + wAltvBen
egoists set wEgo at 1 and wAlt at 0. People for whom wEgo = 1 − ε, where ε is
tiny, are altruists in a very modest sense: they will act to advance the
wishes of others only when the perceived benefits to others are enormous
compared to the forfeits for themselves—they may suffer the scratching of
their finger in order to avoid the destruction of the world, but refuse larger
sacrifices. People for whom wAlt = 1, by contrast, are completely self-
abnegating. They abandon their own solitary desires entirely, taking on
the wishes they attribute to the beneficiary. In between, we find golden-
rule altruists, for whom wAlt = 1/2, who treat the perceived wishes of the
other exactly as they do their own solitary desires.
Even when averaging is not appropriate for representing altruistic de-
sires, there will often be a comparable notion of the degree to which one
has accommodated the perceived wishes of the other. Moreover, with
respect to altruistic emotions there is surely a similar concept. Notori-
ously, we can be relatively unsympathetic, even with those who are dear-
est to us, when we are preoccupied or distracted. At other times, we en-
ter fully into the feelings of friends and loved ones, even of strangers. It is
not obvious how to delineate the notion of intensity in the emotional
case as precisely as the food-sharing example allows, but the varying in-
tensity of altruism in emotional responses is uncontroversial. Notice,
however, that it should not be confused with the intensity of emotion:
intensity depends on the degree of alignment with the other’s feelings
(or with the feeling one would have had in the other’s situation), not
with the force of what one feels.
Most altruists, indeed probably all, lack a fi xed intensity of response,
applying with respect to all potential beneficiaries and all contexts.
There are many people to whom we would rarely make an altruistic re-
sponse: these people effectively fall outside the range of our altruism.
Even with respect to those to whom we are disposed to respond, there
are many contexts in which we do not take their perceived wishes or their
feelings into consideration (or into our own minds). For many, perhaps,
we are prepared to offer limited forms of aid and support; for a few, we
are willing to sacrifice everything. Often our altruistic responses to some
are colored by indifference to others: parents who make sacrifices to help
their children obtain things the children passionately want frequently
do not take into account the wishes of other children (or the altruistic
desires of the parents of the other children).
Someone’s altruism profile typically shows a relatively small number
of people to whom the focal individual responds, frequently with signifi-
cant intensity, across a wide set of contexts. The beneficiaries lie at the
center of the range of altruism for the focal individual, and the scope for
these beneficiaries is wide. As we consider other potential beneficiaries
more distant from the center, the scope narrows (there are fewer contexts
in which the more peripheral people elicit an altruistic reaction) and the
intensity falls off, until we encounter people to whom the focal individ-
ual makes no altruistic response at all. Henceforth, I shall conceive of
the range of A’s altruism in terms of the metaphor of center and periph-
ery: the center is the select set of potential beneficiaries for whom A’s
response is relatively intense across a relatively wide scope of contexts; at
the periphery, the intensity of the response and the scope of contexts
narrow and vanish.
Someone’s character as an altruist is not fi xed simply by the factors so
far considered—intensity, range, and scope—because there are also sig-
nificant cognitive dimensions to altruism. A may make no response in a
particular context through failure to understand the consequences for
B; perhaps A does not differentiate the social from the solitary counter-
part. Often this is an excusable feature of our fallibility, for the impact on
the lives of others may be subtle; we may just not see that following some
habitual practice—buying at the most attractive price, or investing in
promising stocks—has deleterious consequences for people about whose
welfare we care. Evidently, however, acuity with respect to consequences
comes in grades, and we admire those who appreciate the intricate ways
in which others can be affected, while blaming those who “ought to have
seen” the damage they cause.
Similarly, there are degrees to which people are good at gauging the
desires of others. Almost everyone is familiar with the well-intentioned
person who tries to advance the projects of an intended beneficiary but
who is hopelessly misguided about what the beneficiary wants: almost

everyone has had a friend or relative who persists in giving presents no
longer appropriate for the recipient’s age or conditions of life. It would be
hard, I think, to declare that people who attribute the wrong desires to
their beneficiaries, or who overlook consequences for those whom they
intend to benefit, are not acting altruistically when they carry out their
variously misguided plans—their intentions are, after all, directed toward
doing good for others—but their altruism needs to be differentiated from
that of their more acute fellows. Hence I add two cognitive dimensions,
one representing A’s skill in understanding the nature of a social coun-
terpart to a solitary context, and one assessing A’s ability to empathize
with B, to ascribe desires B actually possesses.
A simple reaction to the prospect of human egoism is to propose that
people living in community with one another—or even all people—should
be altruistic; some even take the second commandment of the New Tes-
tament to constitute a complete ethical system. Recognizing the dimen-
sions of altruism undermines that thought. There is no single way to
be an altruist, and, consequently, the commendation of altruism must be
given more specific content. What kind of altruist should we urge some-
one to be? Moreover, is it right to suppose that the best state of the com-
munity (or the entire species) is achieved by having each member (each
person) manifest the same altruism profi le? You might think the ques-
tions have straightforward answers. Along the cognitive dimensions,
accuracy is always preferable: ideally people should be aware of the
potential impact for others and should understand what others want.
For issues of intensity, range, and scope, we ought to aim at golden-rule
altruism with respect to all people across all contexts.
The demand for accuracy on the cognitive dimensions is more plau-
sible but still not uncontroversial. Debate about the second part of the
proposal arises in obvious ways. It might be valuable for people to de-
velop strong ties with some others—the range of human altruism should
have a definite center; from Freud’s worries about the “thinning out” of
our libido in the development of civilization to familiar philosophical
examples about parents who wonder whether they should save the
drowning child who is closer, when their own drowning child is farther
out and harder to rescue, a spectrum of troublesome cases arouses suspi-
cion about completely impartial altruism.19 Moreover, in a world with

finite resources, the desires of others often conflict. If A accurately per-
ceives that both B1 and B2 want some indivisible good, it should not be
automatic that A’s desire should be formed by treating B1 and B2 sym-
metrically. (We may, for example, want A to respond to aspects of the
history of the situation, including what B1 and B2 have previously done.)
None of this is to deny that there may be a level at which we want altru-
ism profiles to respond impartially to others, but merely to insist that the
impartiality we want cannot be adequately captured as golden-rule al-
truism toward all people in all contexts.
Further complexities of the notion of psychological altruism will
occupy us later. For the present, however, we have enough to begin chart-
ing the history of our ethical practices, by understanding how the most
basic forms of psychological altruism could have evolved, and how they
formed an important part of the social environment in which the ethical
project began.
§6. Maternal Concern

Before our human ancestors invented ethics, they had a capacity for psy-
chological altruism. This thesis might be disputed in any of several ways,
but the one of immediate concern recapitulates the skepticism about
altruism mentioned earlier (§3). Armed with the elements of an account
of psychological altruism, the first task is to decide if any such capacity
exists, and if it could plausibly be attributed to contemporary human
beings, our hominid ancestors, and our evolutionary cousins. Let us
begin with the most straightforward case.
Behavior directed toward the survival of young is quite widespread in
the animal kingdom, found, for example, among birds as well as mam-
mals. With respect to some types of animals, the hypothesis that this
behavior is directed by altruistic desires appears extravagant, for it pre-
19. Sigmund Freud, Civilization and Its Discontents (New York: Norton, 1961); Bernard
Williams, “Personhood, Character and Morality” in Moral Luck (Cambridge, UK: Cam-
bridge University Press, 1981); Peter Railton, “Alienation, Consequentialism, and the De-
mands of Morality,” in Facts, Values, and Norms (Cambridge, UK: University Press, 2003).
supposes the propriety of attributing wants and intentions apparently

beyond the cognitive capacities of the pertinent organisms. Nevertheless,
we might view the animals as driven by altruistic emotions (or primitive
versions of them), generated through the operation of automatic neural
mechanisms. Among primates, however, particularly those closest to our
own species, our evolutionary cousins the great apes, there is consider-
able evidence for the ability to have desires and to recognize the desires
of others.20 For the sake of concreteness, we can think of psychologically
altruistic dispositions to care for the young as emerging in apelike ances-
tors of Homo sapiens, but it is eminently possible that they evolved much
further back in our primate (or even mammalian) past.
Even those who share the orthodox primatological views about the
cognitive sophistication of our evolutionary cousins may be skeptical of
any hypothesis that parental care is sometimes directed by altruistic
desires, in the sense I have explicated in §3. They may wonder, for
example, whether any dispositions of this kind could evolve under Dar-
winian natural selection, or whether the apparently altruistic behavior
is really the product of some quite different mechanism. Perhaps the ani-
mals are really calculating how to achieve future benefits, violating con-
dition 4 of my account, the anti-Machiavelli condition. Many primatolo-
gists take the social organization of primate life to reveal “Machiavellian
intelligence,” and evolutionary psychologists often propose that increased
cognitive powers in hominids reflect the need to manipulate others and
to avoid being manipulated oneself.21 Or perhaps the plausible candi-
20. There are many excellent sources for attributing complex cognitive states to nonhu-
man primates. See, for example, Dorothy Cheney and Robert Seyfarth, How Monkeys See the
World (Chicago: University of Chicago Press, 1990), esp. chaps. 3 and 8; Jane Goodall, The
Chimpanzees of Gombe (Cambridge, MA: Harvard University Press, 1986); C. Bachmann
and H. Kummer, “Male Assessment of Female Choice in Hamadryas Baboons,” Behavioral
Ecology and Sociobiology 6 (1980): 315–21; R. Byrne and A. Whiten, eds., Machiavellian
Intelligence (Oxford, UK: Oxford University Press, 1988), particularly the essay by Nicho-
las Humphrey (“The Social Function of Intellect,” 13–21).
21. Many, though not all, of the essays in Byrne and Whiten, Machiavellian Intelligence
(see n. 20), adopt this perspective. For a more pronounced articulation of the theme that
intelligence is a tool for calculating egoists, see James Barkow, Leda Cosmides, and John
Tooby, eds., The Adapted Mind (New York: Oxford University Press, 1992). In “The Social
Function of Intellect,” Nicholas Humphrey offers a broader vision (see esp. p. 23).
dates for altruistic responses to the young are affective and immediate.
That would allow for altruistic emotions, even emotions that direct
behavior, but not necessarily for altruistic desires. To address this latter
concern, I shall begin with an example that involves serious cognition
and planning.
Primates roaming on the savannah sometimes encounter carcasses
that could serve as food. Imagine a female finding a carcass in the ab-
sence of her young. Instead of devouring it on the spot, she quickly sum-
mons her young. It is difficult to think of behavior of this sort as an
action driven by instincts or emotions. Apparently, the mother has to
recognize this as food she can share, and to prefer sharing to devouring
it entire. Perceiving the possibilities for her young, she forms a different
desire from the one she would have formed had they been out of range or
fully mature and dispersed. That desire underlies her efforts to summon
them to the scene before the food spoils or is taken by another animal.
On the face of it, this is an example of altruistic desires in the sense of §3.
One line of concern about attributing altruistic desires is that capaci-
ties for such wishes could not have evolved and been maintained under
natural selection. In settling this worry, we can use the tools supplied to
solve the problem of biological altruism. Suppose that food has decreas-
ing marginal value (in terms of promoting reproductive success), so that,
although eating a whole carcass has a higher effect, on fitness it is consid-
erably less than double the effect of eating just half. Assume that the
mother has a disposition to golden-rule altruism (or some approximation
of it) with respect to her offspring, and that there is just one of her young
in the vicinity. Then it is not hard to show that this disposition can be
favored by kin selection.22
The more difficult challenge asks whether all the conditions for psy-
chological altruism have been met. Perhaps the adjustment of desires to
accommodate the perceived needs of young is based upon “Machiavel-
lian” calculations. What form might these supposedly self-directed
processes take? Begin with a style of skeptical argument rarely made
explicit, but one underlying the conviction that references to psychologi-
cal altruism are exercises in sentimental self-deception. According to
22. For details, see Kitcher, “Evolution of Human Altruism.”

this line of thought, the benefits to off spring, favoring the evolution-
ary success of altruism, undermine its genuineness. In the described
scenario, however, the mother must do something psychologically
sophisticated—she has to recognize this as an occasion for seeking out
her young—rather than simply exhibit some instinctive reaction. What,
then, is the alternative cognitive account that replaces the disposition to
adjust preferences with Machiavellian calculation? It strains credulity to
suppose mothers recognize the evolutionary advantages of sharing: only
a few very select primates could calculate the genetic gains and losses
(and those who do make their judgments in this way are, to say the least,
misguided). So if she calculates it will have to proceed via proxies,
through the attempt to attain selfish goals correlated with increases in
reproductive success. What could those be?
The most plausible answer is that maternal care proceeds from expec-
tations of future reciprocity—the child is expected to grow into a future
ally, maybe eventually a caregiver. Here, the consequences of the present
action would be represented in terms we can imagine being within the
mother’s conceptual repertoire, but we are supposing animal abilities to
abstract from present conditions and to envisage a very different future,
to overlook the weak juvenile and see a future strong ally. Even if we allow
such amazing foresight, problems remain. If dispositions to share with
young evolve under natural selection because of inclusive fitness consid-
erations, then the expectations of future aid ought not to be an accurate
guide to the kinds of behavior selection would favor—the alleged prox-
ies do not match up well with the variables (the gene frequencies) that
are the “ultimate currency” of evolution. From the standpoint of inclu-
sive fitness, mothers should provide some aid when there is very little
chance of reciprocity in the future (simply because, even without recip-
rocation, helping offspring is a good way to spread the genes), and they
should provide extra aid to offspring who can be expected to recipro-
cate. If the hypothetical calculation is to give values that correlate with
inclusive fitness, the perceived gains from reciprocity have to be inflated.
Why should mothers think their care will be remembered, or, if recalled,
it will trigger a disposition to repay? If sharing is based on the expectation
of returns, the young seem bad targets. Other, more mature, members of
the group would appear to be better prospects for future aid.
The best version of skepticism invokes psychological variables corre-

lating closely with the well-being of the young, and thus presumably
with the spread of the pertinent alleles. Determined “Machiavellians”
may concede that the scenario I described—in which mothers bring
young to share food—involves cognitive abilities, but they may view the
calculations that occur as directed toward benefits arising from simpler,
more instinctive, reactions. Start, then, with maternal responses to dis-
tress. Here, it might be alleged, mothers promote their own ease by pre-
venting wails, facial expressions, and upsetting bodily gestures; or, more
positively, mothers find psychological pleasure in observing smiles or
hearing happy gurgles. This, it is conceded, is a form of emotional altru-
ism. Hence, on occasions where offspring are present, maternal behavior
(hugging, caressing, giving food) is directed by the desire either to avoid
an unpleasant state (“the pang,” generated from emotional responses to
unhappy young) or to attain a pleasant one (“the glow,” similarly gener-
ated from emotional reactions to contented young).23 When the young
are not directly present, however, but available to be brought to the car-
cass, the mother recognizes the possibility of attaining the glow (by
bringing them to the scene) or the dangers of experiencing the pang (if
she devours the whole carcass and then encounters hungry offspring).
So she calculates that her own selfish desires can better be satisfied by
sharing. Because the anti-Machiavelli condition is violated, she does not
count as a psychological altruist.
At least two problems confront this skeptical response. The first, and
more obvious, is the highly implausible style of cognition it attributes to
the mother at the scene of the carcass. She is supposed to be capable of
representing to herself not only her absent offspring and their need for
food (as on the interpretation of her as a psychological altruist), but also
the ways potential actions will bring about glows and pangs—she has
to have such thoughts as “If I find the young and share, I shall enjoy the
glow” or “If I devour all the food, I shall suffer a pang when I meet the
23. In Unto Others (Cambridge, MA: Harvard University Press, 1998), Elliott Sober
and David Sloan Wilson rightly regard this kind of skeptical response as the most impor-
tant challenge to the existence of psychological altruism. I think their way of dealing with
it is unnecessarily complex, and offer a simpler treatment. Nonetheless, we are in agree-
ment that the challenge can be met.
offspring.” Even the most liberal cognitive ethologist is likely to wonder

if thoughts like these are within the repertoire of our primate relatives.
Moreover, to deliver the appropriate behavior, the envisaged glow or pang
has to be sufficiently vivid to override the present desire for the available
meat. Only anti-altruist prejudice could inspire the idea that these hypo-
thetical calculations plausibly reconstruct the animals’ psychological
lives.
The story already presupposes one type of altruistic tendency: moth-
ers feel altruistic emotions. That was allowed in describing the situa-
tions from which the skeptical response sets out, the distress felt in the
presence of howling infants, pleasure in smiles and gurgles. By the skep-
tic’s own lights, altruistic emotional responses (in the sense of §4) under-
lie the Machiavellian calculation. Curiously, the skeptical complaint
assumes that these emotional responses engender complicated cognitive
and volitional states (beliefs and desires about glows and pangs) but do
not issue in much simpler desires. The mother’s emotional response to
her needy young produces no desire to feed them, but a longing for
glows or a fear of pangs. Invoking complex Machiavellian calculation
and ignoring the far simpler psychological route leading from emotion to
simple desire again looks like an egoist prejudice, not a serious rival
hypothesis.
These points can be developed further by temporarily leaving our
evolutionary past and focusing on apparent altruism among human par-
ents. Imagine a mother whose child has some serious need, a need diffi-
cult to satisfy—the child must be rescued; the mother has to engage in
an intricate and risky procedure to have any chance of saving the child’s
life. Enough determined mothers pursue similar causes with unusual
energy and persistence, and for them hypotheses about future reciproca-
tion, respect from third parties, or enhanced social status would be jokes
in extremely poor taste. The most difficult form of the skeptical hypoth-
esis proposes that these mothers are driven by internal mechanisms—
particularly by desires to avoid the pang. We find it natural to suppose
that they “couldn’t live with themselves” unless they did everything
possible for the child (interestingly, in the human case, we tend to rec-
ognize the supposed psychological states, the glows and pangs, as inter-
twined with matters of conscience, a point that will be important later).
Hence, the skeptic proposes, mothers do the impressive things they

do because they want to avoid a future of terrible self-reproach and
self-torment.
At least two things cast doubt on the skeptical hypothesis. First, the
fact that the mother envisages the future of self-reproach testifies to the
motivating power of her recognition of the child’s wishes (or, in this in-
stance, more likely the child’s interests—see §21). It is often preposter-
ous to suppose a mother will reproach herself because she is concerned
with attitudes in her society—frequently, those around her would praise
her for doing far less than she does, constantly reassure her that she has
done more than anyone could possibly expect, and so forth. The drive
to pursue every possible avenue comes from within, and it could not be
abated by any amount of well-intentioned commendation and comfort.
If she fails, the mother will suffer, no matter how much she has done and
no matter what others say, and the suffering will stem from her deep de-
sire that the child survive and flourish. So, at least, we might initially
believe. On the skeptical hypothesis, however, that desire must be de-
nied. Instead, the mother must be viewed as being able to feel altruistic
emotions in response to her child. This ability, and the emotions to which
it gives rise, does not express itself in a desire for the child’s well-being.
Instead, the ability leads her to fear a particular type of future state, and
the fear replaces the denied desire as the driver of her conduct. We have
no grounds for accepting this speculative psychology.
A final—fanciful—way to underscore the point: Our world hardly
abounds with clever spirits, willing to offer bargains. Yet the mother
might have a particular disposition to react to temptations. Imagine that
she were visited by a Mephistophelean figure with a straightforward pro-
posal: “I can give you a pill to ensure you will not feel any guilt should
things go badly for your child. The pill will wipe away both the pangs of
conscience—you will reflect on your efforts and feel you did your best—
and any memory of this conversation and the decision to accept the pill.
The downside is truly tiny. The probability of your saving your child if
you don’t take the pill is p; the probability if you do take the pill is p − ε
(where ε is really infinitesimal). Surely the reasonable thing is to accept?”
With respect to many actual mothers, we have no doubt about how they
would respond—by telling Mephisto to get lost. They view their future
psychological comfort as trivial compared with the value of saving the

child—any diminution of the probability of success is a loss for which
future amnesia cannot compensate. Their assessment of relative value
expresses just the desire for the child’s well-being the skeptic attempts to
deny.
Psychological altruism is real, it is exemplified in maternal concern,
and it originally evolved through the most fundamental type of kin selec-
tion. Because it is hard to envisage how psychological altruism could
take hold without directing maternal care—no other social bond is as
pervasive in our evolutionary past, no other recurrent situation is as rel-
evant to reproductive success—it is the most basic and primitive type of
altruism.
§7. Broader Forms of Altruism?

How far does psychological altruism extend? Is it merely something
mothers (or parents) direct toward their young?
For a first example, we can turn to the inverse of the relationship just
examined, to occasions on which offspring help their parents. In her
study of the chimpanzees of Gombe, Jane Goodall relates a moving story
about the behavior of an adult female, Little Bee, who tended to her par-
tially paralyzed mother, Madam Bee.24 On several occasions, Little Bee
and her mother lagged behind the rest of the troop, often arriving at the
nesting site hours later than the others. Mother and daughter took fre-
quent rests, and, when food was needed, Little Bee climbed trees, col-
lecting fruit to share with Madam Bee. Apparently, Little Bee adjusted
her preferences to accommodate the perceived needs of her mother, and
by doing so she exposed herself to risks she might otherwise have
avoided. Reading Goodall’s account, it seems clear that the fi rst three
of my conditions for psychological altruism are satisfied. The crucial
requirement, where skepticism so often arises, is the anti-Machiavelli
condition.
Was Little Bee’s adjustment of her preferences based on calculating
some narrow advantage for herself? It is hard to think what it might be.
24. Goodall, Chimpanzees of Gombe, 357, 386.

There was no realistic possibility of her pronounced efforts on behalf of

her mother being reciprocated by some future benefits conferred by
Madam Bee. Nor could she obtain extra status among the members of
her troop, who were in no position to witness her actions—indeed, be-
cause her time for interacting with others in the group was so drastically
curtailed, her chances for cooperative interactions with them were di-
minished. If her behavior resulted from calculation, aimed at advancing
her own solitary wants, the only possible conclusion is that she miscal-
culated, but the miscalculations would have been so gross as to be quite
at odds with her demonstrated social intelligence. Far more plausible is
the hypothesis that Little Bee was what she seemed to be—a psychologi-
cal altruist.
Similarly for a young male chimpanzee, observed by Frans de Waal:25
Early one morning, de Waal watched two members of the Arnhem chim-
panzee colony enter the outdoor enclosure: Krom, a somewhat retarded
mature female, and Jakie, a healthy young male. It had rained overnight,
and rain had collected in one of the tires hanging from a horizontal pole
attached to the climbing frame. Krom wanted to free that tire, but, unfor-
tunately, it was the innermost of five, and her efforts at removing all five
tires at once proved futile. After she sat down in a corner of the enclo-
sure, Jakie approached the frame. Intelligently, he removed the tires one
at a time, carefully carried the rain-filled tire to Krom, and set it gently
before her. She made no gesture of gratitude.
As with the complex pattern of behavior exhibited by Little Bee, it is
very hard to suppose Jakie’s action stemmed from the operation of some
automatic precognitive mechanism. The whimsical hypothesis that, as
he saw Krom’s efforts, his own mirror neurons fired in ways producing
a readiness for tire-pulling behavior, expressed in imitation of her efforts,
could only beguile us if we ignored the direction of his actions to-
ward the release of the innermost tire and the subsequent careful carry-
ing of that tire to Krom. To explain what he did, we must credit him with
recognizing that Krom wanted the innermost tire—with the water in-
side it.
25. Frans de Waal, Good Natured (Cambridge, MA: Harvard University Press,
1996), 83.
Jakie modified his wishes from what they would have been in Krom’s
absence, and he did so in light of his perception of her desires. He
aligned his wants with hers. Are there grounds for skepticism about his
altruism? If so, they must stem from concerns that the anti-Machiavelli
condition is violated. Perhaps Jakie expected some future reciprocation—
but that would be to impute to him a seriously misguided appraisal of
Krom’s future abilities to reward him (an appraisal quite at odds with his
clear social intelligence; Jakie understands Krom’s place in the troop).
Perhaps he aimed to impress others—but Jakie was surely aware that
the only other primate around was the (socially irrelevant) de Waal. Or
should we think Jakie not only feels glows and pangs, but has the cogni-
tive powers to perceive the present causes of their future occurrence?
Skeptics about altruism are often moved by the thought that an egoistic
story is less extravagant than a hypothesis introducing some ability to
identify with others. Here, however, skeptical hypotheses about glows
and pangs seem the truly extravagant options.
So we can broaden the domain of psychological altruism in the non-
human world, at least a little. This is important for understanding the
ethical project, because it allows us to attribute altruistic desires to ani-
mals before ethical considerations are on the scene. A central theme of my
approach to altruism is that there are preethical forms of altruism and
that these are realized in animals who have not yet acquired ethical prac-
tice. Yet caution is necessary. Besides the striking—and clear—cases,
there are many instances of primate behavior suggestive of altruism, in
which skeptical challenges are far harder to rebut. Observations of chim-
panzees and bonobos frequently inspire the interpretation that par-
ticu lar pairs form genuine friendships, that the mutual adjustment of
behavior signals an underlying modification of preferences and inten-
tions, prompted by recognition of the other’s wants. When the appar-
ently stable alliance breaks down, when a “friend” deserts a seemingly
close ally, there are two possible reactions: one can see this as revealing
that the parties were calculating all along, using one another to mutual
advantage (or apparent mutual advantage); or one can suppose it exposes
the previously unnoticed limits of altruism along one of the dimensions
(scope) distinguished earlier. Later in this chapter, my preferred expla-
nation of the evolution of psychological altruism will be used to support
the hypothesis that, in some of these cases at least, we find genuine

altruism.
Recent studies of human behavior often suggest that altruism is far
more prevalent in our own species than in our closest evolutionary rela-
tives, attributing the difference to the power of human cultural evolu-
tion. Although this conclusion may be correct, if psychological altruism
is understood as in §§3–5, it cannot be established as easily as experi-
menters often believe. Indeed, as I shall suggest later, experimental re-
sults taken to support the “pervasive character of human altruism” are
not concerned with psychological altruism at all, but with behavioral al-
truism; as we shall see (§11), some of the types of behavioral altruism in-
volved are interesting in their own right.
Participants in interactions where there are possibilities for sharing
are willing to divide a pool of money with fellows, even though they have
the chance to take everything for themselves, and this finding persists
across cultures.26 The behavior counts as psychological altruism only if
these subjects are responding to the wants of their perceived beneficia-
ries and the response is also not the result of an attempt to satisfy solitary
preferences. One might worry about both conditions. First, these par-
ticipants have little knowledge of the wants of their beneficiaries. It is
thus hard to view their response as a modification of preferences through
perception of another’s wants or needs. Second, the skeptical hypothesis
that apparently altruistic behavior is driven by desires to achieve glows
or avoid pangs has considerable plausibility in these conditions. It is hard
to rule out the suggestion that these people share as they do because they
want to accord with (or do not want to violate) canons of approved social
behavior. They are behavioral altruists whose motivations are not read-
ily characterized as either altruistic or egoistic.
Reflection on the experiments raises the disturbing thought that there
is important kinship between the performances of these behavioral al-
truists and those of their counterparts in earlier studies of willingness to
inflict pain and punishment—to administer electric shocks to people
26. I ignore here the variety of ways in which opportunities for sharing arise, and, in
particu lar, the important point that subjects will sometimes give some of their assigned
money to “punish” participants who fail to share. For a more extensive discussion, see §11.
who are allegedly being “trained” or to function as an effective “prison

guard.”27 In both types of psychological experiments, the behavior elic-
ited, whether apparently callous (even “monstrous”) or apparently altru-
istic, may largely express a desire to conform to social expectations.
Perhaps the precise and imaginative experiments on sharing behavior
are not really concerned with psychological altruism. Demonstrating the
conditions for psychological altruism is demanding. One should con-
ceive altruism as covering both the nonhuman examples discussed ear-
lier and the behavior of the experimental subjects, without raising awk-
ward issues about motivations. For some purposes it is surely more
appropriate to concentrate on behavioral altruism—if, for example, one
wants to scrutinize the hypothesis that economic agents always behave
as rational self-interested agents, exploring the possibilities of behavioral
altruism is exactly what is needed.
For our purposes, however, there are two reasons to focus on the more
demanding notion of psychological altruism. Those who recognize and
respond to the wishes of others are different in important ways from
people who are moved to help solely by their desire to be well regarded
or to have the narcissistic comforts of self-congratulation. The conjec-
ture that similar motivations pervade the studies of sharing behavior and
of willingness to torture brings home the point in a dramatic way—even
though we might not want to lump the sharers with those who adminis-
tered “shocks” in the “very dangerous” range, the recognition of an under-
lying propensity to conform in both situations reminds us that aiming
at conformity can blind one to the wants of others with damaging
consequences.
More important, if one hopes to understand how ethical practices grew
out of human capacities for psychological altruism, the conception of psy-
chological altruism will have to be prior to that of behavior done in accor-
dance with, or out of regard for, social norms or ethical maxims. If, as
seems likely, the actions of many of the experimental subjects express their
wish to exemplify norms of sharing, then their “altruism,” if we call it that,
will be a product of their immersion in the ethical practice of their
27. For a concise and informative survey of these experiments, see John Sabini and
Maury Silver, Moralities of Everyday Life (Oxford, UK: Oxford University Press, 1982),
chap. 4.
community. Behavioral altruism of this sort cannot be found in the societ-

ies in which the ethical project began. We cannot trace the project to prior
dispositions to altruism, if we suppose that the prior dispositions are forms
of behavioral altruism grounded in acceptance of ethical maxims.
My persistence in advocating a demanding conception of psychologi-
cal altruism allows for interesting and valuable forms of human action
besides the psychologically altruistic ones. Altruism, to repeat, is a com-
plex notion. As we shall discover later, the taxonomy of human action
has further complications—it would be wrong to suppose that every-
thing else, besides psychological altruism, is undifferentiatedly and bru-
tishly selfish. In understanding the evolution of human ethical practice,
further distinctions and conceptions will be needed (see §11); at that stage
it will be possible to provide a more adequate view of the experimental
research alluded to here.
For the time being, it suffices to acknowledge some examples of psy-
chological altruism, manifested in other primate species and in our own,
besides the fundamental instances of maternal concern. The next step
will be to understand how altruistic dispositions might have originated
and been maintained under natural selection. We turn to the second
part of the task assigned at the beginning of this chapter: to show that
dispositions to psychological altruism were necessary for the type of
society shared by our hominid ancestors, chimpanzees, and bonobos.
§8. Possibilities of Evolutionary Explanation

The most fundamental forms of psychological altruism, concern for off-
spring and, more broadly, altruistic tendencies toward close relatives,
can readily be understood in terms of kin selection (as already indicated
in §6). If an organism tends to adjust its preferences in response to the
perceived wants of others (in accordance with the conditions of §3), if
there is an allele (or alleles) that underlies that tendency, if the others
who benefit from the tendency are relatives, and if the extent of the ben-
efit is sufficiently large with respect to the personal sacrifices (gauged in
terms of reproductive success) made by the altruistic animal, the allele(s)
and the tendency will spread under natural selection.28 Kin selection
28. For details, see Kitcher, “Evolution of Human Altruism.”

allows for psychological altruism as one mechanism for helping behavior

toward relatives, but it will equally favor any mechanism achieving the
same effects. The fact that psychological altruism issuing in aid toward
relatives would have been favored by kin selection does not entail that it
must therefore exist. In §6, psychological altruism was defended as the
best explanation for some types of sharing and helping behavior toward
young. Given the altruistic tendency, kin selection is the most likely ex-
planation of its presence. (Of course, it would count against the original
attribution of psychological altruism to primate mothers if there were
no plausible evolutionary explanation.)
Section 7 began with the poignant example of Little Bee and her
mother, and here too there is a ready explanation in terms of kin selec-
tion. Imagine an original state in which the only form of psychological
altruism is directed toward offspring. Suppose a new variant arises, a ge-
netic change causes (in the pertinent environment) a tendency to broaden
the range of altruism, allowing for possibilities that other animals, be-
sides the young, will provoke that modification of preferences constitu-
tive of psychological altruism. Animals with the variant are less fussy
about those they want to help, but their altruistic responses are always
toward close relatives. For concreteness, assume that an animal with the
variant has the original tendency to respond, when a parent, to the per-
ceived needs of the young, as well as other tendencies to respond to the
perceived needs of parents and siblings. Helping siblings and parents
(although not to the same intensity with which aid is channeled toward
one’s own young) contributes to the spread of the variant allele: for sib-
lings have chances to produce off spring with that allele, and parents
likewise have opportunities for generating further young of the new
type. Hence the broadening of the psychological altruism originally
focused in maternal concern can be favored by kin selection.
The evolutionary scenario just outlined will account for behavior like
Little Bee’s. A tendency to respond to the perceived wants and needs of
one’s mother would be favored by kin selection, for, frequently, the help-
ing behavior produced by the altruistic tendency would increase the
mother’s expected reproductive success and the frequency of the allele(s)
underlying the broadening of psychological altruism. Sometimes, how-
ever, animals with the tendency may make sacrifices that far outweigh
any expected returns—as exemplified by Little Bee’s devotion to Madam
Bee. If their helping behavior were based on calculation, it would be

grotesquely misguided, belying the animals’ manifest intelligence. It is
better viewed as a noncalculational, emotional response, of a type that
normally increases inclusive fitness, but that, in the case at hand, has
negative effects on the spread of the underlying alleles. (Madam Bee’s
predicament arouses altruistic emotions in Little Bee—and the disposi-
tion to be aroused in this way is generally adaptive; the altruistic emo-
tions give rise to particular altruistic desires; on this occasion, acting on
those altruistic desires detracts from reproductive success.)
Will the envisaged evolutionary account extend to the example of
Jakie and Krom? Perhaps. Here the relationship is far more distant, but
the sacrifice made by Jakie is also quite trivial in comparison with Little
Bee’s months-long dedication. A tendency to (mild) psychological altru-
ism toward any member of the ambient social group might be favored by
kin selection, for there is always a (significant?) chance it will direct aid
toward relatives, and thus favor the spread of the relevant allele(s). Any
hypothesis along these lines would have to be carefully elaborated—for
the reproductive costs and benefits are by no means as easy to assess as
in the simpler examples involving close relatives—and it also presup-
poses that evolution of the traits underlying primate social life can be
understood prior to accounting for the spread of psychologically altruis-
tic tendencies to group members. Animals without the broader tenden-
cies would have to be able to evolve capacities for group life, so that, with
the group in place, the stage would be set for kin selection to favor the
expansion of psychological altruism across a broader range. In §9 I shall
directly question this presupposition and argue that psychological altru-
ism is fundamental to primate social life.
Kin selection is only one of the two mechanisms whose recognition
resolved the long-standing puzzle of the evolution of biological altruism.
The other is the disposition to reciprocate. Tendencies to engage in a pat-
tern of interaction with other organisms, in which each participant gives
up something on one occasion and reaps a greater gain in some subse-
quent encounter, can evolve, thus accounting for cooperation among
nonrelatives.29 The initial thought is simple and elegant. If two animals
29. This approach stems from the important work of Robert Trivers, William Hamil-
ton, and Robert Axelrod. The Trivers-Hamilton-Axelrod approach has given rise to an
share a propensity for making small sacrifices (measured in terms of re-

productive success) to promote greater (reproductive) benefits for the
other, if they interact repeatedly, and if the propensity has a genetic ba-
sis, then each may reap (reproductive) advantages from the sequence of
interactions. Suppose you and I are the animals in question. Today I
help you to some significant biological benefit, at much smaller repro-
ductive cost to myself. Tomorrow, you return the favor. Each of us has
made a net gain (measured in terms of reproductive prospects). The lon-
ger we continue, the larger the benefits we garner. The apparently pedan-
tic introduction of qualifying terms—“biological,” “reproductive”—is
important because a mode of evolutionary explanation for biological al-
truism does not automatically provide a convincing account of the evolu-
tion of psychological altruism. With respect to kin selection, the situation
is different, for kin selection is neutral in regard to whether psychologi-
cal altruism underlies the pertinent forms of helping behavior: where
one can argue that psychological altruism is the best explanation of that
behavior (as with the case of maternal concern; see §6), viewing the ten-
dency as the product of kin selection does nothing to undermine the ar-
gument or its conclusion. Reciprocal altruism, by contrast, precisely
because of the simplicity of the idea, invites the skeptical complaint that
calculational mechanisms are at work, and that the anti-Machiavelli con-
dition is violated. To put the point bluntly, whenever a tendency to a
form of behavior can evolve through reciprocal altruism, it looks as
though animals with the cognitive sophistication required for psycho-
logical altruism would also have the abilities to make a calculation re-
vealing how the behavioral propensity would satisfy their own solitary
preferences; hence there would be grounds for skepticism about any al-
leged psychological altruism. At the very least, when tendencies to be-
havior are explained by supposing they evolved through reciprocal al-
truism, skeptics seem to have a forceful objection to the attribution of
extensive series of further investigations. See, for example, Alexander Harcourt and Frans
de Waal, Coalitions and Alliances in Humans and Other Animals (Oxford, UK: Oxford
University Press, 1992); Karl Sigmund, Games of Life (New York: Oxford University Press,
1993); and Ronald Noë, Jan van Hoff, and Peter Hammerstein, eds., Economics in Nature
(Cambridge, UK: Cambridge University Press, 2001).
psychological altruism: the animals can identify the long-term advan-

tages of trading favors.
If reciprocal altruism were the fundamental mechanism through which
cooperative behavior between unrelated animals evolved, we should
have to meet this concern directly, showing that genuine psychological
altruism could emerge and be maintained because of the (reproductive)
advantages of reciprocation. I shall proceed differently. Patterns of re-
ciprocation have to rest on something more basic, tacitly assumed by
accounts of reciprocal altruism. This more basic evolutionary mecha-
nism favors the emergence of tendencies to psychological altruism. Let
us start by reviewing how cooperation among unrelated animals is typi-
cally explored.
Interactions among animals can be seen as games, in which the play-
ers pursue “strategies” (of which they may or may not be conscious).
The outcomes of each combination of strategies are represented by the
“payoffs” to the players, assignments of numbers representing the val-
ues for them of what occurs (for evolutionary studies, these values are
the effects on their reproductive success). Evolutionary game theory
approaches reciprocation among nonrelatives by considering games in-
volving possibilities of cooperation and also of competition. One par-
ticu lar game has received great attention, the famous prisoner’s di-
lemma (PD).
In PD, each player has two options: to cooperate or to defect. If one
cooperates and the other defects, the former obtains the sucker’s payoff,
while the latter enjoys the traitor’s payoff. If both cooperate, they reap
the reward for mutual cooperation. If both defect, they both receive the
punishment for mutual defection. A table shows the outcomes for both
players (with returns to the “Row Player” listed first, and returns to the
“Column Player” given second).
C(ooperate) D(efect)
C <R, R> <S, T>

D <T, S> <P, P>
(Here T is the traitor’s payoff, R the reward for mutual cooperation, P

the punishment for mutual defection, and S the sucker’s payoff.) It is
supposed that T > R > P > S, and that T + S < 2R.30 If the game is played
just once, defection (D) is a dominant strategy for both players, since
T > R and P > S. Rational actors in a socioeconomic interaction of this
form are expected to wind up with the noncooperative outcome of mutual
punishment, rather than achieving the reward for mutual cooperation—
which, if they could be assured of it, they would prefer (since R > P). By
the same token, if animals sometimes engage in interactions with non-
relatives, where the payoffs in units of reproductive success meet the
conditions of PD, natural selection would apparently favor strategies of
defection.
Not, however, if the interactions are repeated. In an iterated prisoner’s
dilemma (IPD), players can adjust their strategies to the previous perfor-
mance of those with whom they interact. A strategy for IPD consists in a
choice of how to play on the first round, together with a set of preferred
responses to the various potential sequences of choices by one’s partner/
opponent. Suppose you know the interaction will be repeated but do not
know exactly how many times it will occur.31 Your strategy is specified
by saying how you will begin, and how you will act given any potential
history of choices by your partner.
Robert Axelrod investigated the success of various strategies em-
pirically, by inviting scholars to submit their preferred proposals for
playing IPD, and staging a computer tournament. In each round of
the tournament, different strategies were paired (as in a round-robin),
and then played a par tic u lar version of PD against each other for a
large number of iterations. 32 The winner was one of the simplest strat-
egies submitted, tit for tat (TFT), which begins by cooperating, an-
30. The second condition implies that, if the game is repeated, it is cooperatively better
for the players both to play C than to adopt a pattern of alternating C and D (so that, on each
occasion, one plays sucker and the other plays traitor, with alternation of roles).
31. This last stipulation is added to address the concern that it will always be prefer-
able to defect on the last round, that once that is a matter of common knowledge it will
be rational to defect on the penultimate round, and so on. There are complications here
that I shall not explore. For present purposes, it is enough to follow the standard
treatment.
32. For details, see Axelrod, Evolution of Cooperation. Note that the number of itera-
tions is close to two hundred, and that the payoffs in the game—the values of T, R, P, and
S—are the same in each iteration and in each round.
swers defection with defection, and responds to cooperation with co-

operation. In the common parlance, TFT is “nice, provokable, and
forgiving.”
Mathematical analyses of populations consisting of variant strategies
for playing IPD suggested that TFT is evolutionarily stable; that is, in a
population in which it is prevalent, it resists invasion by alternatives aris-
ing at low frequencies. 33 The analyses accounted for the maintenance
of cooperative behavior under natural selection, once it has become com-
mon, but did not explain how such behavior might originate, evolving
from an initial state in which it was rare. Unless they are strongly dis-
posed to interact with one another rather than with the rest of the popu-
lation, TFT variants, arising at low frequencies in groups full of non-
cooperators, are driven out by natural selection.
Two problems have now emerged with the hypothesis that psycho-
logical altruism toward nonrelatives (or psychological altruism more in-
tense than would have been favored by kin selection acting alone) might
have evolved through reciprocal altruism. First, while reciprocal altruism
may help us understand cooperation, its amenability to predictive calcu-
lation raises skeptical doubts about psychological altruism as a mecha-
nism for the cooperation. Second, it is hard to understand how disposi-
tions to cooperate (Machiavellian or altruistic) could have obtained a first
firm foothold. There is a third difficulty, too. The IPD scenario imposes
very particular conditions: two animals are designated as partners for a
long sequence of PDs with exactly the same structure; at the end of this,
they are released and assigned to different partners for a repetition of
that sequence of interactions. The idea that anything like this happened
33. For the important notion of evolutionary stability (of an evolutionarily stable strat-
egy), see John Maynard Smith, Evolution and the Theory of Games (Cambridge, UK: Cam-
bridge University Press, 1982). From the beginning it was apparent that there were indirect
ways in which populations of TFTs could be invaded. In such populations, variants that
invariably cooperate would be indistinguishable from the TFTs and could thus enter.
Once there were sufficiently many of them, the stage would be set for noncooperative strat-
egies to invade through exploiting the undifferentiating cooperators. (See my discussion in
Vaulting Ambition, [Cambridge, MA: The MIT Press, 1985], 100–101.) Further research
revealed that combinations of noncooperative strategies can also invade (Robert Boyd and
J. P. Lorberbaum, “No Strategy Is Stable in the Repeated Prisoner’s Dilemma,” Nature 327
[1987]: 58–59).
in our primate past is immensely implausible. Surely no giant hand

swooped down on the savannah, locking animals into compelled inter-
actions that recapitulated the same form.
To address the difficulties with reciprocal altruism, start with the last.
Far more realistic is a different scenario. Suppose our primate ancestors
had recurrent opportunities for interacting with a conspecific, and, on
these occasions, they could either engage in that interaction or act by
themselves. Assume, too, they could sometimes choose partners for in-
teraction, signaling their willingness (or reluctance) to engage in joint
activity. This would replace the standard structure of the IPD, the re-
peated compulsory games, with something different—repeated opportu-
nities for optional games (as I shall call them). The framework of optional
games is both more realistic and resolves some difficulties besetting the
orthodox understanding of reciprocal altruism.
An example helps to fi x ideas. Our primate ancestors had to remove
parasites from their fur. The task was undertaken repeatedly and could
be done in either of two ways. One possibility is self-cleaning—although
that poses problems because it is hard to reach some parts of the body.
Another is to team up with a partner—but that risks exploitation; after
the first animal has provided a thorough cleaning, the second may pro-
vide something superficial and then go off to more interesting activities.
Primates could have signaled to one another their willingness to engage,
issuing, accepting, and turning down invitations, so that partners for
interaction could be chosen.
With some plausible assumptions about the benefits of hygiene and
the costs of spending time, it can be shown that the scenario envisaged has
the structure of an optional PD. If two animals interact with each other,
the cooperative strategy is for each to provide a thorough grooming for
the other; defecting consists in being quick and sloppy. The best of all
outcomes is to receive the thorough attention of one’s partner and to pro-
vide little in return; slightly less good is to obtain a serious cleaning and
to return the favor; significantly less good is to receive a superficial groom-
ing and to give back the same; even worse (although not much worse) is
to clean one’s partner conscientiously but receive a superficial groom-
ing. Not interacting, “opting out,” and cleaning oneself, is intermediate
between mutual cooperation and mutual defection. Hence, with a some-
what arbitrary assignment of numbers, the structure of the interaction is

as follows:
C D
C <9,9> <0,10>
>D <10,0> <1,1>
Interact
Opt out >5
Mathematical analyses reveal that high levels of cooperation are likely

to develop, and to be sustained, in populations whose members have a
sufficiently large number of opportunities for playing optional PD with
one another. More exactly, a strategy of discriminating cooperation (DC)
can originate and be maintained under natural selection.34 Discriminat-
ing cooperators are prepared to interact with any animal that has not
previously defected on them; if their only opportunities for interaction
involve partners who have previously defected on them, they opt out;
whenever they interact, they cooperate. Suppose we begin with a popu-
lation of antisocial animals, beings who interact and defect with one an-
other. In this state asociality will be favored: the solo strategy (always opt
out) does better. In an asocial population (full of solos), however, a lone
DC does equally well; there are no opportunities for interacting, and
DCs are left partnerless to behave like solos. Once a second DC is pres-
ent, however, the two of them team up for a happy life of cooperative in-
teractions that bring large advantages over their asocial fellows. So, from
antisociality, the population proceeds via asociality to a state of high
levels of cooperation. Those high levels will be maintained until the
frequency of nondiscriminating cooperators becomes sufficiently high
(among DCs, nondiscriminating cooperators are invisible—they are never
exploited) to allow antisocial types to enter and take advantage of them.
When that happens, the population can relapse to an antisocial state.
34. The results summarized here were originally presented in Kitcher, “Evolution of
Human Altruism.” I should note that the strategy DC described here is characterized as
DA in the earlier paper (“discriminating cooperator” is a more accurate label than “dis-
criminating altruist”).
Computer simulations reveal that the history of high levels of coopera-

tion is quite long. 35
There are further encouraging results about the mechanisms of coop-
eration. Suppose we abstract from some of the conditions I placed on
psychological altruism in §3, and, in particular, from the Machiavellian
concerns about calculation. Let quasi altruists be individuals who meet
conditions 1–3 but not necessarily condition 4: they adjust their prefer-
ences to align them more closely with what they take to be the wishes of
others, but they may do so on the basis of considerations of their own
expected narrow benefit. As in the discussions of §§3 and 5, it is possible
to gauge the intensity of the quasi altruist’s response, in terms of the
weight assigned to the perceived wishes of the other. Under a regime of
repeated opportunities for playing optional games of various types, se-
lection will favor quasi altruism of a more intense kind, up to golden-rule
quasi altruism because quasi altruists with more intense responses will
participate in a broader class of profitable interactions with others.36
Replacing the scenario of compulsory IPD with the framework of op-
tional games helps. Not only does it offer a more realistic scenario for the
evolution of cooperation, but it overcomes the problem of understanding
how cooperation got going. It even points toward some conclusions about
the mechanisms underlying cooperation: selection will favor tendencies
to respond to the wants of others that give the others’ preferences as much
weight as one’s own. Plainly, however, the shift does not address the most
fundamental difficulty in using reciprocal altruism to explain the evolu-
tion of psychological altruism—for it preserves the simplicity that invites
skepticism. Animals with the cognitive resources to count as psychologi-
cal altruists would be able to see the advantages of discriminating co-
operation and of being prepared to cooperate across a wide range of types
of interaction. The scenario thus shows how Machiavellian calculators
might have evolved to behave like golden-rule altruists.
To address this problem, to show how full-fledged psychological al-
truism of kinds going beyond those favored by kin selection might have
35. See John Batali and Philip Kitcher, “Evolution of Altruism in Optional and Com-
pulsory Games,” Journal of Theoretical Biology 175 (1995): 161–71.
36. This result is derived in Kitcher, “Evolution of Human Altruism.” Note that quasi
altruists resemble behavioral altruists, although some behavioral altruists may not meet
condition 3 of §3.
evolved, requires a more decisive break with the mechanism of recipro-

cal altruism. Analyses in terms of both compulsory and optional games
can play a role in understanding human social practice. The evolution of
primate sociality, however, is based on a different scenario, one favoring
the emergence of psychological altruism.
For optional games presuppose certain forms of cooperative abilities
that have not yet been explained.
§9. The Coalition Game

Worries about the realism of the scenarios so far envisaged should remain.
The primatological work of the past decades queries some assumptions
hidden behind the mathematical analyses. Assuming our evolutionary
cousins serve as good models of our primate pasts, can we really sup-
pose our ancestors behaved like discriminating cooperators? On the one
hand, chimpanzees and bonobos seem not to cooperate anywhere near
as much as the conception of them as discriminating cooperators sug-
gests. Moreover, they often fail to cooperate with the “right” partners—in
joint hunting, for example, those who help bring down the prey are not
always rewarded, while those who have not taken part end up with pieces
of the spoils, and yet the dispossessed appear willing to return the next
day for a similar expedition. 37 More generally, chimpanzee and bonobo
societies are pervaded by asymmetries the account fails to recognize.
Grooming partnerships embody some of these asymmetries, and a more
focused look at grooming shows it to be a far more complicated phenom-
enon than the analysis outlined in §8 pretended. If considerations of
hygiene alone were pertinent, it would be impossible to understand the
enormous amounts of time chimpanzees and bonobos devote to groom-
ing one another. During some periods in the recorded histories of
primate troops, particularly when social tensions are running high, the
animals devote three to six hours per day to plucking and smoothing one
another’s fur. 38
37. See, for example, Goodall, Chimpanzees of Gombe, 288–89.

38. See Frans de Waal, Chimpanzee Politics (Baltimore: Johns Hopkins University
Press, 1984), and Peacemaking Among Primates (Cambridge, MA: Harvard University
Press, 1989).
These features of primate societies point to the more fundamental

presupposition of the explanations in terms of reciprocal altruism: these
are animals who can endure one another’s presence, who can occupy the
same region together at the same time. In the original IPD scenario, that
is simply achieved by force majeure; the organisms are locked together in
their long sequence of interactions. Although the shift to optional games
increased the realism, it took for granted the existence of a pool of po-
tential partners. Animals were supposed to encounter others quite fre-
quently and to be able to signal their willingness to interact. For that,
a minimal form of sociality must already be in place—the animals must
be sufficiently tolerant of one another’s presence to form the pool. Recip-
rocal altruism presupposes an ability to treat others as potential partners
and not as dangerous rivals.
That ability should be the first and fundamental target of evolution-
ary explanation. The processes that gave rise to it generated a capacity
for psychological altruism of a more extensive type than those under-
stood in §8 in terms of kin selection.
Begin with some well-established conclusions about social life among
the apes. Within this relatively small group, the extent to which social
relations, tolerance, and cooperation extend beyond the family varies
greatly. Gibbons divide into small family groups (mother, father, and
young) that are typically hostile to outsiders. Male orangutans are mostly
solitary, ready to defend their territories against incursions from other
males; they interact only perfunctorily with the females whose home
ranges lie within those territories; the extent of female-female associa-
tion is a matter of controversy (with older orthodoxy supposing that fe-
males travel with one or two offspring, and newer observations pointing
to intermittent joining of pairs of females). Groups of gorillas typically
contain several adult females but have only one adult male; to a first ap-
proximation, gorilla social life involves some cooperation among unre-
lated females and only aggressive interactions among adult males.39 For
larger social units, with cooperation among unrelated adults of both
39. A valuable source for discussions of social life among the apes is Barbara Smuts;
Dorothy Cheney, Robert Seyfarth, Richard Wrangham, and Thomas Struhsaker eds., Pri-
mate Societies (Chicago: University of Chicago Press, 1987).
sexes, we must turn to our evolutionary cousins, chimpanzees and

bonobos.
Chimpanzees live in bisexual groups (varying in size from about 20 to
approximately 100), within which there are shifting patterns of alliances
and dominance relations. Among bonobos, the groups are somewhat
larger (roughly 50 to 150), with the same sorts of changing internal struc-
tures.40 A principal difference between the two groups is that the major
associations in the wild seem to be among chimpanzee males and among
bonobo females, although in both species, there are important social in-
teractions among members of the other sex (and between members of
opposite sexes). Study of hominid remains suggests that our ancestors
lived in mixed groups and that their size was of the same order as that
found in living chimpanzees and bonobos. How did the chimp-bonobo-
hominid pattern of sociality evolve?
Any answer to the question must identify the features that distinguish
chimps and bonobos from the other great apes. I shall develop an ap-
proach originally outlined by Richard Wrangham, who proposed that
female behavior is shaped directly by ecological factors, particularly the
distribution of the foods consumed by the species; males have to adapt to
this distribution, adjusting their behavior to increase the chances of cop-
ulating with estrous females.41 Crucial for our purposes is the conjecture
40. I shall tend to take chimpanzees, rather than bonobos, as the model for our hominid
past. This decision rests partly on a sense that many small human societies that live in en-
vironmental conditions closer to those of our ancestors appear to share the relative intoler-
ance for neighbors that is so marked in chimpanzee social life, and, more important, on the
hypothesis that psychologically altruistic tendencies are more prominent and pervasive in
bonobos than in the (common) chimpanzee. Hence I assume that if a compelling story
about the evolution of sociality and its roots in psychological altruism can be given for
chimpanzees, it would be easier to defend a similar account for bonobos. (Here I am in-
debted to a valuable conversation with Frans de Waal.)
41. See Richard Wrangham, “On the Evolution of Ape Social Systems,” Social Science
Information 18 (1979): 334– 68; “An Ecological Model of Female-Bonded Primate Groups,”
Behaviour 75 (1980): 262–300); “Social Relationships in Comparative Perspective,” in
Primate Social Relationships: An Integrated Approach, ed. Robert Hinde (Oxford: Black-
well, 1983); and “Evolution of Social Structure,” in Smuts et al., Primate Societies, 282– 96.
Wrangham bases his analysis on the hypothesis that the principal determinant of female
reproductive success will be her access to food and that the principal determinant of
male reproductive success will be the ability to copulate as frequently as possible with
that mutually hostile communities of chimpanzees have “evolved from a

hypothetical solitary-male system because males could afford to travel in
small parties, even though the optimal foraging strategy was to travel
alone; they were forced to do so because lone males therefore became
vulnerable to attacks by pairs.”42 Abstracting from the emphasis on forag-
ing, one may recognize that, in a world with scarce resources—of what-
ever kind—competition among vulnerable animals may require their par-
ticipation in coalitions and alliances. Addressing that problem is prior to
realizing possibilities for cooperation: for understanding cooperative in-
teractions among unrelated animals, PD (whether optional or compul-
sory) is not fundamental; the framework for the games animals play is set
by the problem of forming coalitions and alliances.43
Imagine a population of solitary organisms (the largest units being
mothers with dependent young) in an environment in which each must
obtain a certain number of resources in order to survive and reproduce.
Suppose the resources are scarce, the animals fight over these resources,
and the stronger typically win. A five-stage process could have led from
the initial situation—no cooperation except for maternal care in early
life—to the kind of social structure found in chimpanzees, bonobos, and
estrous females. So, for example, on his account, orangutans pursue their relatively soli-
tary lives because females can most efficiently forage for fruit by working alone, and males
have physical abilities to defend a territory including the smaller home ranges of several
females. I shall make no such specific assumptions. Instead, I abstract from the particulari-
ties of Wrangham’s discussion, offering a more general model of which his approach would
be a special case.
42. Wrangham, “Evolution of Social Structure,” 290. Compare Hobbes: “. . . the weak-
est has strength enough to kill the strongest, either by secret machination, or by confedera-
tion with others that are in the same danger with himself” (Leviathan, 82). Hobbes, how-
ever, would not have thought that this could apply to the brutes, because, without speech
“. . . there had been amongst men neither Common-wealth, nor Society, nor Contract, nor
Peace, no more than amongst Lyons, Bears, and Wolves” (Leviathan, 20). Hobbes under-
rated the lions and the wolves and knew nothing of the chimpanzees and the bonobos [New
York: Oxford University Press (World’s Classics) 2008].
43. Some primatologists have recognized the point in the context of their studies of par-
ticu lar societies. See, for example, R. Noë, “Alliance Formation Among Male Baboons:
Shopping for Profitable Partners,” in Coalitions and Alliances in Humans and Other Ani-
mals, ed. A. Harcourt and F. de Waal (Oxford, UK: Oxford University Press, 1992),
285–321.
hominids. (Note that what is required here is an account of how a form of

social structure we independently know to exist could have emerged and
remained stable under natural selection: a “how possibly” explanation.)
1. Asociality—animals range alone (at most accompanied by depen-

dent young), finding some resources without contest (“scramble”
competition) and competing directly for others (“contest”
competition).
2. First Coalitions— some animals arise that are disposed to act
together in contest and to share the resources obtained (not
necessarily equally).
3. Escalation—because of the success of the early coalitions, larger
coalitions form, sharing the benefits they earn in contests (not
necessarily equally, and possibly involving interactions among
subcoalitions).
4. Community Stabilization—coalition size is ultimately limited by
the difficulty of defending all the resources in a range, and the
habitat becomes partitioned into ranges defended by stable
communities, within which the resources are divided by the
formation of subcoalitions.
5. Cooperation—by engaging in optional games (some of which
may be optional PD) and behaving cooperatively, members of the
stable communities increase their fitness.44
Without pursuing the technical details, I shall try to show how this pro-
cess might unfold.
Begin with a more benign version of the initial state, a Rousseauian
world that contains more than enough for everyone. As the population
expands, competition enters. Eventually, so long as the competition goes
on in the assumed way, some animals will not find the resources they
need to survive.
44. Note that the fitness values that occur in the payoff matrices for the games played by
community members, whether optional or compulsory, must reflect the consequences of
actions for the underlying alliances to which the animals belong. This recapitulates the
point made earlier that the structure of animal interaction cannot be understood in isola-
tion from the demands of the most fundamental game, here seen as the coalition game.
If the animals pursue solitary strategies for gaining resources, as en-

visaged in stage 1 of the process outlined above, there will be contests for
some resources. Assume, for simplicity, that the contests are resolved
without actual fighting: the animals simply assess one another’s strength,
and the weaker one retires (in cases of equal strength, divisible resources
are shared equally; indivisible resources are assigned to each animal
with probability 1/2). Throughout the course of their lives, the strength of
the animals changes, according to an obvious schedule. Initially, while
an animal is under the protection of its mother, it effectively has what-
ever strength its mother has. Once released from its mother’s care, it is
at its weakest. Thereafter, strength grows as the animal matures, provided
that sufficient resources are obtained; eventually, perhaps, animals that
live long enough undergo a slight decline in strength.
Populations faced with these conditions are vulnerable to extinc-
tion. For a new generation to arise, the young must survive the critical
period after their release from maternal care. During this period, they
are the weakest members of the population, and whatever they achieve
must be gained by fi nding resources currently uncontested by others
and consuming those resources before a stronger individual arrives to
dispossess them. If the competition is sufficiently severe, all resources
will be contested, and, after a brief period of maternal care, all the young
members of the population die. In a very hostile world, populations
stuck at stage 1 are likely to be short-lived. More exactly, the pressure
of mortality will cull the population so it is effectively returned to a more
benign—Rousseauian—environment.
Suppose, however, variants arise that are disposed to team up with oth-
ers. Specifically, imagine a variant that is prepared, when weak, to travel
around with another animal of similar weakness, to collect resources to-
gether and to divide them. (There may be variation in the propensities to
tolerate different schedules of division.) If two such variants encounter
each other, they form a coalition. Because the members of the coalition have
to travel together, the coalition can visit only as many resources as a single
individual can. Assume that strength is additive; that is, the strength of a
coalition is the sum of the strengths of its members. Each of the variants in
a coalitional pair can now increase its access to resources, for the doubled
strength will surely provide victory in contests with other weak young ani-
mals and may be sufficient to win some encounters with older members of
the population. Selection thus favors variants of this type, even if the divi-
sions of the resources acquired are not even.
Plainly, several parameters must be set in developing versions of the
scenario I am envisaging, but it is possible to show that, given almost all
ways of choosing values for non-Rousseauian worlds, any population at
stage 1 will contain at least one pair of organisms who can increase their
fitness through coalition formation. That does not mean, of course, that
the disposition to team up must evolve: there might be no way to gener-
ate any such propensity. I shall suggest shortly that more basic capacities
for psychological altruism provide a way in which the successful vari-
ants might emerge.
Just as stage 1 would favor the emergence of pairwise coalitions, so too
the emergence of pairs puts pressure on animals who are working alone.
The gains of the animals who team up are obtained by dispossessing
those who would otherwise have done better. Any variation that equips
them with a disposition to pair with another animal will be favored. As
the population becomes full of coalitional pairs competing with one an-
other, the weakest pairs will do better if they are prepared to add single
members or merge with other pairs. Selection favors the variants who
unite with others at the size required by the actual escalation of coalition
formation.
Although the origination and escalation of coalition formation is easy
to understand, the termination of the process appears more mysterious.
The rationale, however, is a direct consequence of the fact that coalitions
have to travel together if they are to exert their joint power. No coalition
can visit more resources than a single individual. When the environment
is filled with large coalitions, coalition members who receive the small-
est shares may have no better option than to resume scrambling for re-
sources the large coalitions are not able to visit. The dynamics of the
process leads to a situation in which the habitat is partitioned into terri-
tories controlled by sizable coalitions, occasionally with a floating popu-
lation of individuals who live on the fringes.45
45. The announced results are not hard to derive analytically. They coincide with the
findings of some ingenious computer simulations designed by Dr. Herbert Roseman. See
his unpublished Ph.D. dissertation “Altruism, Evolution, and Optional Games,”2008 Co-
lumbia University.
This is an evolutionary scenario for the emergence of the social struc-

ture found in chimpanzees, bonobos, and hominids, one that will lead to
groups mixed by age and sex, each of which controls a relatively stable
territory that it defends against neighboring rivals. Within these groups,
there will be patterns of alliances bearing on the division of the resources
the group commands. That structure will determine the potential benefits
of various possibilities for cooperation, for there will be gains from
strengthening existing alliances and costs from disrupting them. The ho-
mogeneous pool of partners for optional games, envisaged in the analysis
of §8, is structured by the shapes of previous encounters. Reciprocal al-
truism and interaction in optional games can be understood only against
the background of the coalitional structure of the group.
So far the conclusions address only animal behavior, with no direct
implications about psychological altruism. To go further, it is necessary
to ask how the variants envisaged, with their disposition to team up with
others, might have been psychologically realized. Answer: this ability
to form coalitions, and ultimately to constitute a stable social group, ex-
presses a further expansion of those fundamental psychologically altru-
istic tendencies attributed in the case of maternal care.
Mothers have a propensity to modify their wants and preferences from
what they would otherwise have been, to accommodate the perceived
wants of the young. Primates have evolved to broaden this response to
others, so that preferences reflect the perceived wishes of close relatives, a
broadening supported by kin selection and manifest in the behavior of
Little Bee. I propose a further extension: the disposition to adjust wants
and preferences to the perceived preferences of an age-mate, initially trig-
gered in contexts where both animals are weak and vulnerable. This is a
species of psychological altruism, the capacity for early friendship. Pairs
of animals with this broadened altruistic disposition reap the advantages
just outlined. Young animals, no longer under parental protection, need
allies if they are to gain anything in a competitive world. Psychological
altruism of this special type is one way for them to find friends.
Skeptics will suppose there are self-interested routes to the same end.
What would they be? The coalition game is by no means a simple op-
portunity for reciprocal altruism. It does not present the players with a
compelled or optional iterated prisoner’s dilemma, inviting them to cal-
culate a strategy for success. The coalition game is many-personal—and,

for the players, the number of participants will typically be unknown. It
is not even evident what would count as a “best strategy” for playing it.
Whether someone counts as a good ally or not depends on all sorts of
delicate facts animals have no way of recognizing. Moreover, working
out a good procedure for playing the game challenges the intellectual
powers of mathematicians, economists, and philosophers. The best one
can do is pick a partner, team up—and hope.
That appears to be just what chimpanzees and bonobos do. Their al-
liances do not seem to depend on any tallying of costs and benefits.
Instead, these animals are prepared to support members of their groups
with whom they have a history of interactions, often dating back to peri-
ods early in their lives—the strongest alliances descend from that period
of juvenile vulnerability.46 What sorts of calculations might underlie
their behavior?
It is natural to believe that the clever head can always substitute
for the kindly heart, but that need not always be so. When the prob-
lems posed for reasoned selection of the best strategy are sufficiently
intractable—as they are in the case of the coalition game—it may not just
be that an emotional response to another animal, the transfer of altruis-
tic dispositions to identify with others to a novel sphere, the domain of
“early friendship,” does no worse than the cunning of the Machiavellian
calculator, but that it works better. Animals with a disposition to try to
work out the costs and benefits suffer from too little information to make
good decisions on this basis, and their efforts can easily lead them to
abandon an alliance when there are no serious prospects for doing any
better. Furthermore, they may hesitate more than their blindly sympa-
thetic counterparts, and indeed be recognizable by others as less reliable
and less stable coalition partners.
When weak animals are forced to compete for resources they need,
their inability to win contests by themselves confers a selective advantage
on a disposition to identify with the interests of conspecifics, particularly
with those who are in a similar predicament. That advantage fostered the
spread of propensities to psychological altruism antecedently limited,
46. See Goodall, Chimpanzees of Gombe, 379–85, 418–24.

first toward young and then toward close relatives. The broadened pro-
pensities allowed for the formation of those loose coalitions found in our
evolutionary cousins. Far from being anthropomorphic, sentimental, or
self-deceiving, the hypothesis advanced here looks like the best explana-
tion of the form of sociality of our hominid past. It also explains why the
friendships of youth are so deep and enduring, both in human beings and
in other primates, and why newcomers are sometimes accepted into pri-
mate social groups when a resident animal has formed social bonds with
them in a shared past as juveniles together.47
Psychological altruism is the kernel from which ethical practice grows—
because it lies at the heart of the type of sociality our hominid ancestors
experienced. As we shall discover, however, the plant is far more elaborate
than the seed.
47. De Waal relates a striking instance, in which a relatively unprepossessing male (Ji-
moh) was accepted into a chimpanzee troop because of his prior association with two older
females in the group. See de Waal, Good Natured, 131–32.
chapter 2
Normative Guidance
§10. The Limits of Altruism

Imagine a population of organisms with altruistic dispositions. For each
of these organisms, there is a variety of contexts and a range of other
members of the population such that the psychological states of the focal
organism—specifically the desires and the emotions—will adjust to re-
flect that organism’s perceptions of the wants, needs, and feelings of the
others. These dispositions enable the organisms to function as a popula-
tion, to live in the same place at the same time and to encounter one an-
other daily without too high an incidence of social friction and violence.
But the dispositions are limited: cooperators are sometimes exploited,
returns are uneven, and, when there is an opportunity for large selfish
benefits, even long-standing allies are sometimes left in the lurch. Defec-
tions threaten to tear the social fabric, and, in their wake, much signaling
is required; our organisms engage in prolonged bouts of mutual groom-
ing and other forms of physical reassurance.
I shall call these organisms “hominids,” although it would be equally
apt to dub them “chimpanzees.” The limitations of their psychological
altruism cause the tensions of their social lives and prevent them from
gathering in much larger groups and participating in more complex
67
cooperative projects. A look at their evolved descendants some quarter

of a million generations later discloses that the limits have been tran-
scended. Ten thousand years before the present, those descendants have
formed settlements that sometimes contain a far larger population; they
have learned to interact peacefully with many conspecifics whom they
do not encounter on a daily basis; and they have constructed complex
systems of cooperation that involve marked differentiation of roles. How
has all this been achieved?
One possibility is that they have acquired some new and stronger
mechanism for psychological altruism. Conceiving hominid societies as
exactly like those of contemporary chimpanzees (or bonobos) is plainly
implausible, for the members of the later hominid societies had diverged
from their evolutionary cousins five million (or more) years ago. Perhaps
as hominid brain size increased, it was necessary for babies to be born at
developmentally earlier stages (so their heads would still pass through
the birth canal), with the consequence that they were more dependent
for a longer period of time. The resulting selection pressure may have
favored enhanced altruistic tendencies in the specific context of provid-
ing care for helpless young.1 Without underestimating the importance of
steps like these, it is evident that neither hominids nor contemporary
human beings have escaped entirely from the difficulties and tensions,
the rivalries and conflicts, of chimpanzee social life. If human societies
are less vulnerable to breakdown than those of our primate relatives, it is
because other modifications have taken place.
What modifications? The hypothesis to be explored is that other
changes that have occurred, including in particular the acquisition of
language, have made it possible for human beings to reinforce the origi-
nal limited altruistic tendencies, so members of human societies no lon-
ger falter quite as frequently in their cooperation. Because defection is
1. Arguments along these lines have been developed by Kristen Hawkes and her col-
leagues (see, for example, Hawkes, James O’Connell, Nicholas Blurton Jones, Helen Alva-
rez, and Eric Charnov, “The Grandmother Hypothesis and Human Evolution,” in Evolu-
tionary Anthropology and Human Social Behavior: Twenty Years Later, ed. L. Cronk, N.
Chagnon, and W. Irons [New York: De Gruyter, 1999]) and by Sarah Hrdy, Mother Nature
(New York: Pantheon, 1999), and Mothers and Others (Cambridge, MA: Harvard Univer-
sity Press, 2009).
Normative Guidance 69
more often prevented, less time has to be spent in reknitting the social
fabric. The cumbersome peacemaking of our original hominids is re-
placed by a new device, one preempting rupture rather than reacting to
it, and in principle capable of operating in a wide variety of contexts.2
That device is necessary for what we think of as ethical practice. I shall
call it a “capacity for normative guidance.”
The previous chapter was at pains to defend attributions of psycho-
logical altruism and to rebut the skeptical insistence that sees Machiavel-
lian intelligence behind apparently helpful or kindly actions. Its account,
however, was entirely consistent with the thesis that the psychological
altruism of our hominid ancestors was limited. Recall two of the dimen-
sions of altruism: range and scope. An animal may be disposed to re-
spond altruistically to particular other members of its social group (“close
friends”) across a relatively broad set of contexts, and to respond to all
members of its social group in some contexts (banding together against
outsiders, for example), although there are occasions on which it would
act selfishly even toward its closest friends and staunchest allies. The
limited quality of chimpanzee-hominid altruism, in both range and scope,
set the stage for the emergence of normative guidance.
The limits of altruism are most starkly and spectacularly visible when
the selfish rewards for deserting erstwhile allies are extremely high—as
when a male has the opportunity to achieve dominance in the social group.
A study of “chimpanzee politics” in the colony at Arnhem (an environ-
ment allowing the animals to retain important features of their life in the
wild, but, at the same time, providing opportunities for systematic ob-
servation of them) revealed the ways in which three high-status males—
Yeroen, Luit, and Nikkie—related to one another and to the high-status
females, during times of transitions in power.3 Each male exhibited social
behavior readily interpretable as aimed at retaining dominance, achiev-
ing dominance, or, at worst, serving as the principal lieutenant of the
dominant male. In the early phases of the struggle, Luit aided the newly
2. It is thus more than a special-purpose mechanism, like the hypothetical emotional

disposition that underlies alloparenting.
3. Frans de Waal, Chimpanzee Politics (Baltimore: Johns Hopkins University Press,
1984).
adult Nikkie in achieving dominance over the females, while Nikkie’s

diversionary tactics enabled Luit to dethrone the previously dominant
Yeroen. Once he had attained alpha rank, Luit’s policy changed. He
consolidated his position by siding with the females and with Yeroen
against Nikkie. The abruptness and decisiveness of the switch can easily
inspire the conclusion that chimpanzee politics is thoroughly Machia-
vellian. Apparently, friendships among chimpanzees are situation linked.4
The subsequent twists and turns of the story seem to underscore that
judgment. Yeroen deserted Luit to form a coalition with Nikkie, so that
Nikkie eventually became dominant with Yeroen as his lieutenant. After
a subsequent period of tension between the two allies, Luit reemerged at
the top of the hierarchy, apparently in a weak coalition with Nikkie. The
uneasy situation was ended by a night fight, in which Luit was fatally
injured by the other two.5
To say there are no stable friendships within chimpanzee communi-
ties is too strong, for some alliances endure for years, even for virtually
the entire lifetimes of the animals—as §9 insisted, the friends of one’s
vulnerable youth are often one’s lifelong companions.6 Moreover, before
the political instabilities in the Arnhem colony, Yeroen and Luit had
been longtime allies. The fascinating (but sad) story of the months of
conflict reveals—as do similar examples, less fully documented, among
wild chimpanzees—how the presence of a clear opportunity for self-
advancement can expose the limits of altruistic dispositions. Observers
4. De Waal offered a sober judgment of the relationships he observed: “Coalitions based

on personal affi nities should be relatively stable; mutual trust and sympathy do not appear
or disappear overnight. . . . If friendship is so flexible that it can be adapted to a situation at
will, a better name for it would be opportunism” (Chimpanzee Politics, 128). Readers of de
Waal’s subsequent books (Peacemaking Among Primates [Cambridge, MA: Harvard Uni-
versity Press, 1989]; Good Natured [Cambridge, MA: Harvard University Press, 1995];
Primates and Philosophers [Princeton, NJ: Princeton University Press, 2006]) may be sur-
prised by his early emphasis on hard calculation—for the later work is softer in tone and
more inclined to highlight the “good-natured” aspects of primate behavior. The account I
offer in the text supplies a perspective from which all of his evaluations can be endorsed.
5. De Waal, Peacemaking Among Primates, chap. 2. De Waal makes the important point
that Luit’s desire to remain with his social group was so strong that it was difficult to re-
move him, even after he had been severely wounded.
6. See Jane Goodall, The Chimpanzees of Gombe (Cambridge, MA: Harvard University
Press, 1986), chap. 8.
have seen enough varied contexts in which two animals respond to each
other to assign them to each other’s range of altruism—until the animals
encounter a new type of context, in which an altruistic response would
require the forgoing of huge potential gains. The selfish action in that
context is a sign not that everything in the past has been opportunism,
but just that the altruistic disposition is incompletely pervasive. Even for
animals who are central to the range of the altruist’s altruism, there are
circumstances outside the scope of that altruism.
The conception of psychological altruism offered in §§3–5 reveals
what is occurring. Chimpanzees (and our hominid ancestors) have regu-
lar psychological propensities for making an altruistic response to an-
other member of their group, with the intensity dependent on salient
features of the circumstances. Even though an animal frequently displays
a tendency to accommodate the wishes and needs of a particular band
member—a “friend”—there are environments in which the intensity of
the altruistic response drops to zero. In those environments, altruism
suddenly vanishes. Friendship is “situation linked” because there is no
fi xed value of the intensity of the altruistic response depending solely on
the strength of the relationship.7 Even in the most committed mutually
altruistic relationships, circumstances offering one party the chance of a
huge advantage diminish the intensity of the response. When the stakes
are high enough, it disappears entirely.
The struggle for dominance presents in high relief contours visible in
more mundane settings. Every day in chimpanzee troops, members who
are not one another’s principal allies act in blithe indifference to their
fellows’ obvious plans. Attempts to obtain a valued object are blocked
or thwarted, requests to share food are turned down, appeals for aid in
conflict are ignored. The animals involved are not entirely indifferent to
one another, for they would band and bond together in the face of an
externally presented threat. Rather, the scope of their mutual altruism is
7. The approach adopted here has some kinship with Walter Mischel’s emphasis on the
failure of cross-situational consistency in people who have stable personality profi les. See
W. Mischel and Y. Shoda, “A Cognitive-Affective System Theory of Personality: Recon-
ceptualizing Situations, Dispositions, Dynamics, and Invariance in Personality Struc-
ture,” Psychological Review 102 (1995): 246– 68. I am grateful to George Mandler for the
suggestion that I explore Mischel’s work.
very limited; only under the most threatening situations is it exercised.

For the rest, they operate on the basis of tolerance of one another’s pres-
ence, though, when one’s indifferent course collides too strongly with
the plans of the other, conflict may erupt.8
The limitations of psychological altruism thus show up both in the
breakdown of close ties under special conditions (the Yeroen-Luit-
Nikkie saga) and in the everyday frictions of animals whose altruism to-
ward one another is limited in scope. The bounds of altruism are revealed
in a third way. Even when an altruistic response is made, and when it
directs a helpful action toward another animal, there are sometimes
signs of psychological division. Conflict within is occasionally visible.
Chimpanzees are openly torn between selfish and altruistic courses of
action, making it apt to attribute to them two desires, both expressed in
facets of their behavior. An animal hesitates. Holding a branch rich in
leaves, he is poised to strip them off and eat, and, simultaneously, the set
of the body acknowledges the presence of an ally; eventually, the arm is
extended, thrusting a small bunch of leaves toward the friend, while the
rigidity of the gesture and the averted face show the presence of a con-
trary desire. The configuration of limbs and muscles is genuinely a mix-
ture. The tension of the moment is apparent.9
Animals can have stable dispositions to respond with quite different
intensities of altruism to environmental cues that can simultaneously be
present. Some features of recurring situations trigger an altruistic response
at a particular intensity; in response to different features the animal is
disposed to react with a different intensity of altruism, or perhaps to re-
act with zero intensity. No conflict appears until the animal encounters
circumstances with both sets of features: the begging gesture elicits the
disposition to share; the lushness of the leaves excites the tendency to
consume. The conflict may be resolved through an action expressing
only one of the desires, or there can be a compromise, a minimal sharing,
or the muscular tension expressing a psychological struggle.
8. In terms of the discussion of §9, the extremely limited altruism profi les displayed in
such cases express membership in different—and often competing—subcoalitions.
9. Because it is so banal, this phenomenon is rarely described in studies of chimpan-
zees. Even a few hours of observation will provide instances.
Human behavior reveals similar phenomena. People trying to lose

weight are tempted by the aromas from the kitchen. They describe
themselves both as wanting and as not wanting the food, and the incom-
patible wishes are expressed in the active salivary glands and the hasty
retreat. Although there is a philosophical temptation to tidy up such
cases, to discover a single preference capturing what the person “really”
wants, there are, as with the chimpanzees, examples challenging the idea
of a single consistent disposition. People who struggle to master a new
language or to set themselves a regular regime of exercise can, with equal
justice, sometimes be seen as either weak in resolve or healthily unwilling
to drive themselves. To find a “real self” free from conflict, we should
have to decide which of the candidates is Jekyll and which is Hyde.10
If the altruistic dispositions of chimpanzees (and hominids) were lim-
ited in the three ways I have described (through breakdown of the most
intense responses in extraordinary situations, through the everyday fric-
tions of more casual friends, and through internal conflicts), their social
lives would be very difficult. They are (and were). Peace and mutual tol-
erance are typically hard-won. Precisely because of this, observations
of chimpanzee societies disclose periods of intense social interaction,
lengthy bouts of grooming undertaken to reassure friends who have
been disappointed by recent behavior. At times of great tension within a
group, chimpanzees can spend up to six hours a day huddled together,
vastly longer than any hygienic purpose demands. Even when daily life
is relatively smooth, the minor difficulties and irritations stemming from
the incompleteness of altruism, specifically the indifference to one another
of animals who belong to different subcoalitions, require an expendi-
ture of time and effort in mutual reassurance. Psychologically altruistic
dispositions make it possible for these animals to live together, but the
limitations of those dispositions subject their social lives to strain. Day
after day, the social fabric is torn and has to be mended by hours of
peacemaking.
10. I draw the examples considered here, as well as the helpful Jekyll-Hyde metaphor,
from Thomas Schelling’s valuable discussion in Choice and Consequence (Cambridge, MA:
Harvard University Press, 1984), particularly chap. 3, “The Intimate Contest for
Self-Command.”
Once, that was the predicament of our ancestors, too.11 They over-
came it through acquiring a mechanism for the reinforcement and re-
shaping of altruistic dispositions, and for the resolution of confl ict.
The evolution of that mechanism, the capacity for normative guid-
ance, was an important step in the transition from hominids to human
beings.
§11. Following Orders

An ability to apprehend and obey commands changed the preferences
and intentions of some ancestral hominids, leading them to act in greater
harmony with their fellows and thus creating a more smoothly coopera-
tive society.12 A capacity for following orders can be expressed in all
sorts of actions, many of which have nothing directly to do with making
up for the limitations of altruism. Self-command, a familiar human ca-
pacity, can address the kinds of problems just discussed.
Those problems, altruism failures, are constituted by occasions on
which an animal A, belonging to the same social group as an animal B
toward whom A is in other contexts inclined to make an altruistic re-
sponse, fails to respond altruistically to B, either forming no altruistic
preference at all or acting on the basis of a selfish desire that overrides
whatever altruistic wishes are present. The simplest—and original—
form of normative guidance consists in an ability to transform a situation
that would otherwise have been an altruism failure, by means of a com-
mitment to following a rule: you obey the command to give weight to the
wishes of the other. A and B belong to the same social group, and, for a
range of contexts R, A forms preferences meeting the conditions on psy-
chological altruism (the conditions of §3). Under circumstances C, how-
ever, A does not respond altruistically to B but retains the desire present
in C*, the solitary counterpart of C (or, for examples of internal conflict,
it is this selfish desire that leads A to action). Under normative guidance,
11. If our hominid ancestors lived in societies more akin to those of contemporary bono-
bos, then their situation would have been less tense than under a chimpanzee form of soci-
ality. The differences, however, are matters of degree, not of kind.
12. Eventually it also modified our ancestors’ emotional lives.
A obeys a command that enjoins behavioral altruism: A is to act in

the way a psychological altruist would; that is, the desire expressed in
the action is more closely aligned with B’s wishes than the selfish desire
would have been.
Just as the discussion of psychological altruism began from a special
example (the sharing of food), so here too a particular case is helpful;
complications come later. Imagine two members of the same social
group, A and B. They share with each other across a wide variety of cir-
cumstances. Faced with an extremely rich and attractive food item, how-
ever, A is not disposed to form the altruistic preference generated in
other sharing situations; the intensity of A’s altruistic response vanishes
entirely. (In terms of the averaging model, although A sometimes sets the
value of wAlt at a value greater than 0, under this particular circumstance,
C, the value of wAlt is 0.) If A is now capable of normative guidance, and
if the normative guidance takes the very special form of A’s commitment
to a command that orders food sharing in C (perhaps it is the command:
“Always share equally with B!”), then the preference A forms in C will
take B’s wishes into account, by setting wAlt > 0 (if the command enjoins
equal sharing, wAlt = 1/2).13 If the preference formed leads to action, A no
longer commits an altruism failure but is behaviorally altruistic. The
newly formed desire satisfies conditions 1 and 2 of the account of psycho-
logical altruism (§3), but not necessarily conditions 3 and 4. A, following
orders, need not be responding to any perception of B’s wants, nor need
A be free of the taint of Machiavellianism. Normative guidance trans-
forms the animal’s psychological life so that something that looks, from
the outside, like an altruistic preference is formed (or is operative) across
a broader range of contexts.
Psychological altruism was characterized in terms of the difference
made to one’s own wishes by the perceived presence (and needs) of
others; now normative guidance is conceived in terms of the differ-
ence made to one’s action-guiding preferences by the recognition of
13. In the discussion of psychological altruism, where A’s own perspective is crucial to
the formation of the altruistic preference, I saw that preference as incorporating A’s percep-
tion of B’s wants. Here I imagine the command as requiring alignment with B’s actual
wants. There will be no discrepancy, when A has an accurate perception, and, for the time
being, I shall assume that mistakes are not made.
commands.14 The modified preferences, however, need not be fully psy-

chologically altruistic—they just are different from the blatantly selfish
wishes that would have prevailed in their absence. The critical idea is
the replacement of a desire that fails to incorporate the perceived wants of
another individual with an action-guiding desire that gives the other’s
preferences some weight. That can be achieved even though the desire is
not generated by the perception of the wishes of another, and even though
it violates the anti-Machiavelli condition. Behavioral altruism (directed
by preferences modified so they are closer to the wants of the beneficiary)
will sometimes do.
Normative guidance produces surrogates for psychological altruism
in animals who can follow orders. The products of normative guidance
(in its simplest and original form) are desires that issue in behavioral al-
truism. To understand the process of normative guidance, the following
of orders that replaces altruism failure by behavioral altruism, it is nec-
essary to probe psychological causes more thoroughly than has yet
been done, both with respect to the lives of normatively guided indi-
viduals and with respect to psychological altruists. For it is tempting to
adopt an oversimplified (and overly neat) picture of the distinction be-
tween normative guidance and the mechanisms behind full psychological
altruism.
On this oversimple view, psychological altruism is generated by an
emotional response to the beneficiary, whereas normative guidance in-
volves the operation of a cognitive faculty (“reason,” perhaps). Psycho-
logical altruism is “hot,” normative guidance “cold.” Both subtheses
should be rejected. Start with the varieties of psychological altruism.
Different kinds of psychologically altruistic individuals are possible.
Imagine an altruist who reacts in context C by modifying his or her wishes
from those occurring in the solitary counterpart C*, because of his or
her perception of the wishes of B; the new desire may be accompanied
by the presence of an emotion toward B, and, if present, the emotion may
or may not cause the new preference. Even if we use a crude and unana-
14. Plainly, one can recognize commands and act in response to them in ways that have
nothing to do with psychological altruism. That will concern us later. For the time being,
normative guidance is tied directly to the reshaping of altruism.
lyzed concept of emotion to consider the situation, we can distinguish

four cases:
a. A’s new desires are caused by an emotional response to B.

b. A’s new desires are not caused by, or accompanied by, any emo-
tional response to B.
c. A’s new desires are not caused by any emotional response to B,
but the factors that generate the new desires also produce in A an
emotional response toward B.
d. A’s new desires are not caused by any emotional response to B;
an emotional response to B accompanies those desires, but it is
independent of the causal process that generates the new
desires.
The oversimple view supposes that cases of type a represent the most
fundamental (primitive) form of psychological altruism; cases b–d dis-
play responses that could emerge only from normative guidance.
Why should one think this? Underlying the view is an apparently
plausible line of argument: the adjustment of desire could result only from
the operation of an emotion or the outcome of a process of reasoning; prior
to the articulation of ethical practice, the only forms of reasoning available
to an agent (human or nonhuman) would have to be calculations of selfish
advantage; hence, preethical adjustments of desire based on reasoning
would fail the anti-Machiavelli condition; by the same token, the only
ways in which obeying commands could produce altruism involve the
recognition of reasons for modifying desire.
On the account of §3, all four types count as instances of psychologi-
cal altruism. The argument just outlined denies that the modification of
desire constitutive of psychological altruism could occur in cases b–d. To
assess it, consider the examples that occupied us in the last chapter.
Some of them fit easily into the simple view. Prominent instances of psy-
chological altruism among primates express an emotional reaction to the
plight of another animal: mothers’ immediate responses to the discom-
fort of the young, or Little Bee’s patience with her mother. It is far from
evident, however, that the example of Jakie and Krom can be so easily
assimilated. Further, as §6 argued, maternal concern is not always a
matter of being prompted by emotion. The primate mother who stum-

bles across a carcass and views it as an occasion for seeking out her
young appears to be undergoing more complicated psychological pro-
cesses, which are not easily captured in a venerable—but crude—
philosophical practice of opposing reason to the passions.
On the ecumenical view adopted in §4, emotions are complex pro-
cesses typically involving both cognitive and affective states. The
causal relations among these states can be quite various, and there is
no reason to suppose that the cognition cannot be primary. Perhaps a
cognition—recognizing that Krom wants the tire and that she has failed
to remove it, seeing that this carcass is food for the young—induces a
new affective state. Or perhaps that cognition leaves the prior affective
condition of the perceiver unaltered—there is no upsurge of emotion
at all, but simply the formation of a new desire on the basis of affective
dispositions already present. Animals can have propensities for form-
ing new desires that do not depend on their entering into a new affec-
tive state. Consequently, versions of b– d can count as psychological
altruism.
Not only can cognition cause affective states, or produce new desires
without modifying the affective background, but there can also be intri-
cate chains of causation in which perceptions give rise to new beliefs, the
new beliefs generate affective states, these affective states, in turn, lead to
altered beliefs, the altered beliefs to novel affective states, all this entan-
gled with the formation of desires: indeed, this may be the stuff of much
of our more complex emotional life. The simple vocabulary employed in
the examples a–d is inadequate to present clearly all the ways in which
psychological altruism can arise (even though we do not yet know just
what form a fully satisfactory conceptualization of the emotions would
take). Moreover, there is no basis for denying at least some of the com-
plex possibilities to nonhuman animals.
This brief for taking the complexities of our emotional life seriously
subverts one-half of the simple view. Troubles also beset the other half,
the proposal that normative guidance must be a matter of reasoning.
Recent work in neuropsychology suggests that the opposition of “cold”
reason to ardent passion is highly problematic and that there is evi-
dence for the role of emotion in what have often been viewed as cool
deliberations.15 Beyond this general point, there are grounds for attrib-
uting a major directive role to emotions in some instances of normative
guidance.
Consider, first, the way in which the psychology of a normatively
guided individual can develop. Initially, a human being, a member of
one of those small bands in which our ancestors lived, is disinclined to
respond to the predicament of one of his fellows. Capable of normative
guidance, he obeys a command to make a behaviorally altruistic response,
and his reacting in this way generates in him an emotional response to
the beneficiary, a primitive feeling of sympathy (as in case c previously).
That feeling is reinforced by the beneficiary’s reaction to his behavior,
and, perhaps after a few further interactions, this person is able to en-
gage in the behaviorally altruistic conduct either on the basis of the origi-
nal process or through a full—psychologically altruistic—identification
with the other. An emotional change may thus be the direct product of
the commitment to following an imperative: as you come to endorse
the command to treat your brother in a particu lar way, your emotions
toward the brother are modified, and the new fraternal feeling gives
rise to the desire to treat him in ways you would previously have
avoided (or resisted). Initially, normative guidance operates to pro-
duce behavioral altruism, but it eventually issues in full psychological
altruism.
How is that first step taken? Must it be on the basis of reasoning—
perhaps through a Machiavellian recognition of the benefits of comply-
ing? Not necessarily. Endorsing the command can embody emotions,
sometimes emotions directed toward the commander: you may accept
it because you are afraid.
The point may provoke an obvious reaction. If the notion of norma-
tive guidance is liberal enough to allow for conformity grounded in fear,
acquisition of the capacity for normative guidance cannot be the decisive
transition to ethical practice. A dilemma seems to loom. If the ability
to follow commands, to obey rules and precepts, is the decisive step in
acquiring a genuinely ethical practice, then this special sort of ability
15. See Antonio Damasio, Descartes’ Error (New York: Putnam, 1994), and Marc
Hauser, Moral Minds (New York: Ecco, 2006).
requires an explanation—for it cannot be rooted in emotions of fear or

prudential calculation. On the other hand, if processes in which people
comply because they fear the consequences of disobedience were avail-
able to our human ancestors and initiated the practice of normative
guidance, then only a simulacrum of ethical practice has been con-
nected with the prior preethical state; the people in question have not yet
made the transition to the real thing. These individuals, allegedly “sub-
ject to normative guidance,” have not yet achieved the distinctively “eth-
ical point of view.” The broad conception of normative guidance allows
for an evolutionary transition from hominids lacking the capacity to hu-
mans who enjoy it, but this continuity is purchased at the cost of losing
contact with the proposed goal, to wit, the emergence of ethics. To make
normative guidance relevant to ethics, one needs a propensity to act in
accordance with commands grounded in a different (and purer) form of
psychological causation.
There is no such purer form to be had. At least since the eighteenth
century, philosophers who have disputed the character of ethical agents
have envisaged an “ethical point of view” in which people give them-
selves commands—commands that are not external but somehow their
own, the “moral law within”—and have regarded this point of view as
requiring the subordination, if not the elimination, of emotion.16 Others
have regarded the operation of emotion as central to ethical agency. It is
often assumed that the major challenge for a naturalistic approach to
ethics consists in showing how the achievement of the “ethical point
of view” might have evolved from more primitive capacities; inspired
by this thought, naturalistically inclined thinkers frequently address the
challenge by attempting a reduction of that “point of view” to the feeling
of special types of emotions. Their disputes with their opponents rest on
a shared mistake.
The acquisition of a capacity for normative guidance—understood, as
above, as an ability to follow orders that issues in surrogates for altruism—
16. A prime source of this view is, of course, Kant, and the most sophisticated elabora-
tions of it are offered in the Kantian tradition of ethical theory. Yet Kant’s opponents, who
often protest the denigration of the emotions, share the emphasis on a distinctively ethical
point of view. I am proposing that we reject a precondition of their debates.
does not mark the transition to the “ethical point of view.” That is not
because there is some further move that does the trick awaited by the
critics, one that shows how a very special kind of normative guidance (a
special way of internalizing the orders, say) constitutes the “ethical point
of view,” but because the entire conception of the “ethical point of view” is
a psychological myth devised by philosophers. There are plenty of ways in
which human beings can be led to recognize and to conform to com-
mands. While it is undeniable that some kinds of causal processes make
ethical progress over others (in ways Chapter 6 explores), we should not
infer a binary distinction between those processes that constitute genu-
inely ethical motivations and those that do not.
Most of the people who have ever lived have embedded their ethical
practices in a body of religious doctrine, viewing the precepts to be fol-
lowed as expressions of the will of gods, spirits, or ancestors (or occa-
sionally as capturing the tendencies of impersonal forces). Fear, awe, and
reverence have been parts of the emotional backdrop to most of the
important decisions and deliberations these people have made, and
virtually all those decisions have been subject to felt concerns about
the attitudes of transcendent beings. The fact that these people have
presupposed massively false beliefs about the universe does not under-
mine their status as ethical agents. Neither should the fact that what they
want, intend, and do are partially caused by emotions of fear and awe.
To insist on an “ethical point of view” liberated from such emotions is
to reserve that point of view for a very small number of cool secularists.
Moreover, it is reasonable to worry that the alleged ethical point of view
is itself only available because of the perspectives previously adopted by
those no longer counted as full ethical agents. The ability to “revere the
moral law” probably depends, in the evolution of culture and in the de-
velopment of individuals, on prior emotions, simpler feelings of rever-
ence now written off as ethically primitive.
There are many different ways in which people can be led to behav-
ioral altruism through their commitment to obeying a command. They
may explicitly represent to themselves the consequences of disobeying,
and find those consequences unpleasant or frightening because of future
interference with their bodies, behavior, or projects. They may make no
such explicit representation, but be moved by fear, or respect for the
commander. They may regard the source of the command as a being

greater than themselves, one whom it is important to obey. They may
actively want to be in harmony with the wishes of some such being.
They may regard the source of the command as part of themselves, and
fear the psychic disharmony caused by disobeying it. They may have a
general idea of the worth of the situations brought about by commands
of a general type to which this particular imperative belongs. They may
want to be the sort of person who lives in accordance with a general class
of commands they have previously endorsed. They may want to live in
harmony with others who expect that commands of this sort will be
obeyed. They may have a general ideal for themselves that involves obey-
ing commands current in their social group. Or they may conceive of
themselves as members of a joint project, in which commands are issued
and obeyed. These surely do not exhaust the possibilities, and some of
the considerations can be present together, with different degrees of
force.
The merits of a liberal articulation of the concept of normative guid-
ance should now be apparent. Our decisions involve a hodgepodge of
considerations and feelings, and it is foolish and unnecessary to limit the
full range of psychological possibilities, taking some to be importantly
free of emotion and others not, some to be constitutive of “the ethical
point of view” and others not, some to accord with the anti-Machiavelli
condition and others not. Emotions are complex processes typically
involving both cognitive and affective states (§4), causation can run from
affect to cognition or in the opposite direction, and our actions some-
times result from intricate cycles involving different types of states. The
simple view, against which I have been campaigning, formulates the pos-
sibilities using language we know to be inadequate (even though we surely
still lack a clear and precise vocabulary for categorizing the relevant states
and processes).
Psychological altruism occurs when perception of the wishes of
another modifies desires to align them more closely with the perceived
wishes. Normative guidance comes about when the recognition of a
command leads someone to act in accordance with it and (in the condi-
tions studied so far, the context of the beginnings of the ethical project)
to replace altruism failure with behavioral altruism. Emotions, desires,
and cognitive states can be entangled in both cases. The causes of psy-
chological altruism and of normative guidance are probably highly het-
erogeneous. There are many ways to be a psychological altruist and,
equally, many ways to undergo normative guidance. None of these latter
modes is especially privileged as definitive of an “ethical point of view.”
No doubt there are extreme cases. Someone who forms the wish to
help another, simply because he is commanded to do so and because he
recognizes that disobedience will bring painful punishment on himself,
is no psychological altruist and (at best) at a rudimentary stage of ethical
practice. At the other extreme, a person who has a general conception
of the wishes of others, who follows a rule because it is taken to promote
the desires of someone else, may be viewed as at least an approximation
to psychological altruism and as participating in a more advanced form
of ethical practice, despite the fact that the wishes, and even the situa-
tion, of some of those she aids are unknown to her, and even though she
has a standing desire to be the sort of person who contributes to the
satisfaction of others’ desires. Normative guidance, as explicated here,
applies to individuals of both types, generating behavioral altruism in
the one instance and something akin to full psychological altruism in the
other.
Given the diversity of causal possibilities, why would one want to take
a stand on which of them has to be realized in a genuinely ethical agent?
The “ethical point of view” emerges as a challenge for naturalism because
it opposes the idea of ethical agents as those sympathetic individuals
who respond to the needs of others. While superficially attractive, these
people suffer a defect that makes them less than fully worthy.17 Their
kindly emotions are unreliable: it is reasonable to fear that the mind of
“the lover of humanity” will sometimes be “clouded,” and that, under
17. The classic source for the reaction is Kant, Groundwork of the Metaphysics of Morals,
Mary Gregor, trans., (Cambridge, UK: Cambridge University Press, 1998, Akademie
pagination 398). This passage is often viewed as expressing an opposition to Hume, but I
suspect that Kant actually had Adam Smith in mind. Not only does Smith develop the no-
tion of sympathy much further than Hume did, but his Theory of Moral Sentiments (Knud
Haakonssen, ed. Cambridge, UK: Cambridge University Press, 2002) [unlike Hume’s
Treatise (Oxford, UK: Oxford university Press, 1978)] is a work Kant is known to have
read.
such conditions, fellow feeling will no longer operate and the person will
act selfishly. Yet if we take the concern about reliability seriously,
“proper” motivation appears impossible. What basis is there for suppos-
ing that carefully restraining the passions and engaging in abstract moral
reasoning (of any of the sorts philosophers have commended) will prove
reliable? Can’t our faculties of reasoning sometimes be “clouded,” too?
Abstract reflection and reasoning are hardly more reliable than the
emotional responses dismissed as capricious. Many of the most horrific
deeds of the twentieth century were carried out in the name of abstract
principles.
As we shall appreciate later, reliability is the issue (§21)—the worry
about the “clouding of the emotions” expressed an important point. Yet
the search for a single type of psychological causation, invariably reliable
or at least always more reliable than its rivals, is foolishly utopian. Dif-
ferent ways of inducing people to modify their preferences and actions
through obeying orders have different merits and deficiencies. Normative
guidance would work better by taking advantage of the ways in which
different psychological processes are suited to different situations. Per-
haps normative guidance evolved in parallel fashion to familiar types of
organic change, where initially crude systems for producing some im-
portant outcome are supplemented with further devices: the organism
has a variety of ways of generating what is required and is thus buffered
against catastrophe.
Normative guidance almost certainly began with crude external or-
ders, followed out of fear; much normative guidance may have been
mediated by respect for the supposed commands of transcendent be-
ings, respect tinged with hopes and fears (§17). Out of those hopes and
fears have come quite other emotional resources for motivating obedience,
feelings of awe and respect, of social solidarity and of contentment in
acting jointly with others, of pride in one’s conduct and of responsibility
to one’s fellows. The history of modes of normative guidance embodies
certain kinds of progress, and attempts to act through following dictates
the agent sets for himself, considers, and endorses have often been
progressive with respect to earlier and cruder forms of psychological
causation. These differences, however, are matters of kind rather than of
degree. Some processes (perhaps processes involving an especially pure
form of emotion, perhaps processes that rein in emotion entirely) are

valuable additions to our repertoire, but they have no special standing
setting them apart from the modes of normative guidance preceding them.
Their merits can be recognized without supposing them to constitute an
“ethical point of view,” which counts as the last word.18
The approach defended here allows a more systematic treatment of
the behavior of subjects in economic experiments (§7). These people are
recruited by researchers, know little, or even nothing, of one another’s
wants or needs, and are placed in situations in which they can decide
what fraction of a monetary reward to share with fellow participants or
how much they will give to punish those who do not act cooperatively.
One thing is clear. The participants’ preferences cannot be adequately
represented by supposing them to be concerned with money and nothing
but money: they do not belong to the fictitious species Homo economic-
us.19 So why do they share, or give money to punish? Not because they
are moved by the plight of people who would otherwise leave empty-
handed, for they lack any basis to make judgments about the impact on
these strangers. One explanation, consistent with the evidence, is that
some form of normative guidance is playing a role. The participants do
what they do, sharing with others, because they follow an order, one
they have accepted and endorsed or one they view as current in their
society.
If they were genuinely moved by a dedication to fairness, a clear-eyed
vision of the value of equality in dividing goods, if this and this alone
moved them to want to share (or to punish noncooperators), we might
18. See Thomas Nagel, The Last Word (New York: Oxford University Press, 1997),
chap. 6; also The View from Nowhere (New York: Oxford University Press, 1986), chap. 9.
19. This is already to demonstrate something that is very important for economic re-
search, for it entails that models imputing utility functions that are increasing functions of
amounts of money, and of this alone, are unlikely to accord with the behavior of actual
agents (for whom other things are important). Indeed, for the project of advancing econom-
ics, any concerns about the ways in which the subjects come to the wants they express in
their actions are entirely irrelevant. What is far less clear is how these ingenious experi-
ments bear on philosophical concerns about altruism and its role in ethical practice. For an
illuminating presentation of the experimental work, see Ernst Fehr and Urs Fischbacher,
“Human Altruism—Proximate Patterns and Evolutionary Origins,” Analyse & Kritik 27
(2005): 6–47.
count their preferences as altruistic. Although they know nothing of the

needs of those they reward, they have a general view that outcomes in
which those people received nothing (or even received less than half)
would be, from the perspective of the beneficiaries, unsatisfactory; the
sense of fairness endorses the complaint, and so, without any selfish
background motive, they want an outcome of equal division. This con-
jecture might tell the whole truth about some of the experimental sub-
jects, but we are by no means forced to accept it. For the available evi-
dence leaves open alternative modes of normative guidance: perhaps the
participants want the “glow” (or to avoid the “pang”); perhaps they want
to be the sorts of people who accord with prevalent social norms of shar-
ing; they know their parents, spouses, friends, or children would dis-
approve of their greedily making off with everything they can; they may
want the approval of the experimenter and not want to go down in his or
her records (even if only mentally kept) as “one of those stingy people”;
without any clear sense of the virtues of equity, they know this is the sort
of thing of which people approve, and the sacrifice does not seem too
large (they are going to leave the lab with something in their pockets).
Elaborated versions of these psychological scenarios raise serious doubts
about whether the anti-Machiavelli condition is satisfied. Even more
obviously, the modified wants are not responses to another person; in-
deed, in some experiments, the actual beneficiary is invisible; the dia-
logue is between the agent and the ambient society (perhaps embodied
in the experimenter).20
Normative guidance can generate full psychological altruism in situa-
tions that would otherwise be altruistic failures. Initially, it almost always
generates behavioral altruism. Human motivation is sufficiently complex
that, in many circumstances including those of the economic experi-
ments, we just cannot tell (at least not without a lot of work—and maybe
some luck besides) how exactly to classify people who act to benefit
others.21
20. Subjects whose primary motivation is to impress (or to avoid disappointing) the
experimenter are easily linked to the experimental subjects who were prepared to infl ict
pain on others.
21. This conclusion motivates the attitudes of the researchers who carry out these ex-
periments, who suppose the important concept is that of behavioral altruism. My account
§12. Punishment
To treat normative guidance in this way has an obvious presupposition.
Behind the disposition to follow orders, whether delivered externally or
from internalized commands, must stand practices of punishment. Un-
less there were sanctions for disobedience, fear could hardly be central
to the initial capacity for normative guidance. Conversely, when punish-
ment is present in a group, it can make possible the evolution of elabo-
rate forms of cooperative behavior (and much else besides).22
Can this presupposition be defended? The actual beginnings of the
ethical project have been seen as a transition from a state of limited psy-
chological altruism to one in which commands are followed out of fear.
The plausibility of that view would be undermined unless there were an
explanation of the possibility of punishment.23
Begin with chimpanzee societies in which a crude precursor of pun-
ishment is already present. Conflicts within these groups are often set-
tled through the interventions of a dominant animal.24 Here rank or
physical strength (or both as concomitants of each other) prevail, and a
dispute is settled—not always, of course, through the infliction of pain
or discomfort on the animal whose initial defection gave rise to the con-
flict. Allies who might have intervened to protect some of those who re-
ceive the rough discipline of the dominant animal anticipate the costs to
themselves and hold back.
of the ethical project also recognizes the important role of dispositions to psychological
altruism. Different concepts are needed in different forms of inquiry and there need be no
quarrel about which notion of altruism is the “right one.”
22. Here I rely on a brilliant essay by Robert Boyd and Peter Richerson, “Punishment
Allows the Evolution of Cooperation (or Anything Else) in Sizable Groups” (originally
published in Ethology and Sociobiology 13 [1992]: 171– 95; reprinted as Chapter 9 of Boyd
and Richerson, The Origin and Evolution of Cultures [New York: Oxford University Press,
2005]).
23. Here, it is important to recall the methodological points of §2. A hypothesis about
the actual origins of the ethical project is supported by evidence about the prior hominid
state, and recognition of familiar human capacities to address its social difficulties. That
hypothesis must be defended by showing that its presuppositions are compatible with the
constraints acknowledged by pragmatic naturalism.
24. Goodall, Chimpanzees of Gombe, 321– 22; and de Waal, Peacemaking Among
Primates.
Punishment need not always take so dramatic a form and can be pres-
ent simply when animals recognize opportunities for cooperation with
one another. Once the basic dispositions to altruism toward nonrelatives
that underlie chimpanzee-hominid society are present, optional games
(§8) are available. There is a pool of potential partners who can be re-
cruited for joint ventures. Because of tendencies to bond with close
friends and allies, some kinds of defections in the ventures will be
tolerated—animals will not behave with the rigor of discriminating co-
operators, refusing invitations to joint activity, when the potential part-
ners are targets of psychological altruism and longtime allies. Neverthe-
less, as the ties are weaker and the history of interaction more limited, it
is to be expected that a strategy like discriminating cooperation will be
favored. The altruistic dispositions emerging from the coalition game
incline animals to give weight to benefits received by their allies, and
thus to increase the value attributed to outcomes in which the ally gains
and the focal individual loses; consequently, animals will be less rigor-
ous in dismissing their close friends as potential partners for interaction;
as the relationship becomes more distant, however, the deviation from
the basic structure of the optional game (for example, optional PD) is
much smaller, and the strategy favored will more closely approximate
discriminating cooperation, refusing further interaction on the basis of a
single defection.
That itself is a form of punishment. To deprive an animal of opportu-
nities for cooperative interaction is to force it sometimes to pursue sub-
optimal ways of meeting its needs. So long as there are occasions for
joint activity with others, allies who remain willing to enter partnerships
with the animal in question, the impact need not be severe. If the allies
are often unavailable, however, or if the refusal to interact spreads more
broadly, life may become quite difficult. Ostracism can be a serious
punishment.25
The practices just mentioned turn on the responses of individuals to-
ward actions by others, actions they do not like. Those individuals can
25. Social confi nement and exclusion are used as forms of punishment in small human
societies. For a vivid depiction of the effects, see Jean Briggs, Never in Anger (Cambridge,
MA: Harvard University Press, 1970).
effectively cause pain for the perpetrators, either through their strength
(or through force that is unchallenged because of considerations of rank)
or through refusal to interact (a response even the weak can usually man-
age). Social participation in these events is minimal: in the one instance,
bystanders behave as mere spectators because of the physical power (or
the rank) of the punisher; in the other, their attitudes or actions cannot
completely undermine the punisher’s success—they may continue to co-
operate with the animal whom the punisher has blackballed, but they
typically cannot compel the punisher to do so.26 More sophisticated sys-
tems of punishment emerge, as animals form social expectations about
the circumstances of punishment.
For an action to be a kind, even a crude kind, of punishment, rather
than simply another contribution to the melee, it is important that
bystanders not be drawn in. Thus, a first step in the direction of punish-
ment requires that other members of the group, even allies of the threat-
ened animal, should not intervene. There is a regularity—friends of the
animal(s) targeted in punishment let it proceed. The next stage couples
the mere regularity with an expectation, shared across the population,
that others will not interfere in such contexts. The expectation suppresses
resistance on the part of the target; the animal picked out expects others
not to intervene and merely suffers what happens. A further refinement
would be the existence of a regularity concerning the animals who carry
out the aggression: perhaps they are animals who bear a particular relation
to the context; perhaps they play a particular social role. Finally, there
arises an expectation about the identities of the animals who initiate ag-
gression. At this last stage, we have reached the systems of punishment
found in contemporary human societies (and in societies for which we
have historical records).
The actual evolution of punishment may have diverged from the
sequence of steps just envisaged; nor is it necessary to specify a point
26. In principle, just as there could be escalation of violence when some animals physi-
cally punish others, so too there could be escalation of noncooperation when a discrimi-
nating cooperator crosses another individual off the list of potential partners. In the former
case, obvious strength or recognition of rank stops the arms race; in the latter, the refusal of
A to play optional games with B is, I suspect, often not recognized and, when it is, does not
inspire B’s allies to forgo potentially valuable opportunities for cooperation with A.
at which “real” punishment is present; nor has it been explained why

any hominid lineage went through these stages. Firm views on the last
issue ought to be grounded in precise models of the advantages of
moving from one stage to the next, and constructing such models would
require far more information than we can probably hope to acquire
about the causes of reproductive success in the ancestral environment(s).27
The challenge is not to understand the actual evolution of punishment,
but to respond to concern that no such evolution is possible. Decompos-
ing punishment into conditions that can be sequentially achieved suf-
fices to demonstrate the possibility of gradual evolution. Crucially, to
buttress the account of normative guidance, the emergence of punish-
ment does not require the prior achievement of ethical practice.
The early stages of the envisaged sequence could have originated
without language: as noted, chimpanzees sometimes resolve conflict by a
crude form of punishment, and the possibility of optional games gives
rise to another. By contrast, the later steps would be facilitated by prior
acquisition of linguistic skills. The emergence of more sophisticated
forms of punishment is probably intertwined with the evolution of
language—and both are probably entangled with the acquisition of
normative guidance.
Suppose a type of altruism failure, keeping food items for oneself, say,
regularly elicits aggressive retaliation from others. Chimpanzees and
hominids could recognize the regularity, thus allowing for variants who
recognize the potential threats to them if they fail to share, and whose
fear generates compliance. With the advent of language, descendants of
these variants can formulate the command for themselves and for others.
Mothers train their young by commanding them to share, and, because
of the command, the young stay out of trouble and avoid risks of injury.
The repeated commands leave an echo on later occasions, and the
27. It is not hard to construct models allowing for the possibility of adaptive advantages
in initiating and refi ning systems of punishment. Those models serve the function of pro-
tecting the hypothesis of a gradual evolution of schemes of punishment against the charge
that they are idle fantasies, incompatible with Darwinian evolutionary theory. Yet, without
far greater knowledge of the ancestral environments, and hence of the values of pertinent
parameters, it would be unjustified to propose that any model of this sort picks out the ac-
tual course of the evolution of punishment. Modesty is appropriate here.
original disposition to share is reinforced by the memory of maternal

instruction.
Through explicit command and fear of punishment, even the primi-
tive punishment of the earliest stages, normative guidance can obtain a
purchase. Animals with a capacity for recognizing and following orders
have advantages over their fellows who lack that ability.28 Once the ca-
pacity is present, it can operate to yield the socially coordinated behav-
ior required by the more advanced forms of punishment. Animals—
now surely human beings—can formulate descriptions of regularities
about the consequences of alternative forms of behavior on the part of
bystanders. Bystanders who intervene are seen to encounter the same
sorts of trouble as the first-order offenders who perpetrate the failures
of altruism that invite punishment. Group members formulate, for
themselves, their kin, and their friends, orders to stand back and let the
discipline proceed. When these rules become prevalent, each can rec-
ognize others as complying, yielding a social expectation that bystand-
ers will do no more than watch. Perpetrators, aware of the expectation,
see the futility of resistance, commanding for themselves a strategy of
docile submission less dangerous than trying to fight back. So norma-
tive guidance, once present, can figure in transitions to more refi ned
forms of punishment. As punishment is refi ned, further regularities
become salient, providing scope for additional occasions of normative
guidance.
Recognizing the painful consequences of particular—and tempting—
courses of action, our ancestors, prompted by fear of the outcomes,
ordered themselves (and their offspring) to hold back. The next step will
be to consider how the grip of this capacity for self-command and self-
control might be intensified.
28. Once again, whether the capacity will be advantageous turns on the details of
the situation. If punishment carries even a small probability of serious damage, and if the
order-following variant is just slightly more likely to avoid the altruism failure, then the
expected gains in terms of staying intact and healthy can outweigh the loss of food that re-
sults from sharing. Once again, we cannot know whether this scenario is plausible; this is a
“how possibly” explanation.
§13. Conscience
Two prominent Shakespearean figures present a view of conscience.
Richard III offers a conjecture about the origins of internal checks on
our conduct:
Conscience is but a word that cowards use,

Devised at first to keep the strong in awe.
Hamlet, while using similar words, worries about the effects of con-
science on behavior, once the tendency for self-regulation is already
present:
Thus conscience doth make cowards of us all,

And thus the native hue of resolution
Is sicklied o’er with the pale cast of thought.
Together, the passages suggest an obvious picture: strong people with

self-interested intentions are held in check by an internalized mode of
normative guidance that substitutes fear for their “native resolution.”
That picture has sometimes moved thinkers to lament the crippling
effects of internalization.29 Whether or not they are right, pragmatic
naturalism needs an explanation of how internalized commands became
possible.30
The first forms of normative guidance, considered in §11, focused on
the capacity to follow explicit orders. Human beings (rather than homi-
nids, since they have acquired language) learned the local rules in child-
hood and later remembered the commands passed on to them. As they
grew in strength, however, the memory of older commands might prove
29. Nietzsche’s complaint is most evident in the fi rst two essays of On the Genealogy of
Morality (Cambridge, UK: Cambridge University Press, 1994); similar themes are sounded
by Freud, in many later works, but especially in Civilization and Its Discontents (New
York: Norton, 1989), as well as by William James in his writings on the “strenuousness” of
the moral life (James “The Moral Phi losopher and the Moral Life” in William James Writ-
ings 1878–1899 (New York: Library of America, 595– 617).
30. Once again, the methodological points of §2 are relevant here.
too weak to overlay the “native hue of resolution.” They might lapse
into the altruism failures from which normative guidance promised
liberation.
As more sophisticated systems of punishment are elaborated, how-
ever, the ineffectiveness of remembered commands becomes costly both
for those who fail to be normatively guided and for other members of
their societies. Variant individuals, with a tendency to respond to modes
of socialization that reinforced the disposition to self-discipline, would
cooperate more thoroughly and encounter less trouble. This extension
of normative guidance involves both social innovations and psychologi-
cal changes in the individuals. On the social side, it requires practices of
training the young members of the group so that the prospects of flout-
ing a command become associated with emotions they find unpleasant.
On the individual psychological front, it consists in refinements of the
emotional lives of these individuals.
The Shakespearean suggestion that fear lies at the root of this process
of internalization need not be exclusive: other emotions might be avail-
able for recruitment to the cause of normative guidance. Imagine a social
group of early humans, able to issue and remember commands, but vul-
nerable to the flouting of those commands by individuals who think of
themselves as strong. An innovation in the training regimes customary
among this group, the practice of issuing orders to the young, promotes
an enduring fear: perhaps they are lured into violating one of the precepts
and then subjected to some extraordinarily harsh and memorable pun-
ishment; perhaps this occurs at an especially impressionable age. There-
after, even as they grow, those trained in this way remain haunted by a
sense of dread as they contemplate disobeying certain commands. Con-
science does make cowards of them. Yet, similar effects can be achieved
in different ways. If the young are induced to identify with some of the
orders current in their group, if they see obeying those orders as partly
constitutive of belonging to this distinctive social unit, they may feel more
complex reactive emotions—pride, perhaps, when they continue to carry
out the commands, shame or guilt when they do not. As these reactive
feelings attach to outcomes considered in prospect, they may substi-
tute for the raw fear of punishment, promoting the same types of coop-
erative behavior on a different basis.
We know too little about the intricacies of human emotions to elabo-

rate this scenario in any great detail, but the outline is clear. The simplest
modes of internalization trade on the ability of programs of socialization
to exploit human fears. More sophisticated methods of training people
can foster other emotions, perhaps emotions unavailable in different de-
velopmental environments, whose association with potential courses of
action reinforces tendencies to behavioral altruism. The result is a soci-
ety in which cooperation is more broadly achieved and in which costly
episodes of punishment are less frequently needed. Further, even at early
stages of the ethical project, different groups may have cultivated differ-
ent emotions, founding their ethical practices in distinctive ways. There
may be several ways to build a conscience.
However it is formed, conscience is the internalization of the capacity
for following orders. The ably socialized individual does not simply hear
the voice of an external commander, or remember the injunctions ad-
ministered in childhood. The commanding voice seems to come from
within, initially and crudely as the expression of fears, later perhaps as
the representation of membership in a particular social group. In either
mode, it provides a more effective anticipation of the costs of deviating
from the approved regularities in conduct than the original tendency
to follow and remember external orders. The conscience-ridden human
being fits more easily into the social niches, provides less provocation to
punishment, and encounters much less trouble.
If, to borrow another phrase from Hamlet, society plays upon the indi-
vidual as on a pipe, it need not always be the same tune. Successful social
inculcation of normative guidance may work through quite different emo-
tional complexes, even though variant group techniques succeed equally
in securing cooperative behavior. Although conscience begins in fear, it
may later be dominated by shame or guilt, pride or hope, emotions avail-
able only in social environments where normative guidance, in some
cruder form, has already taken hold.31
31. In accordance with my strategy of outlining a scenario, I offer no detailed claims

about how any of these emotions is to be understood, or whether, as some anthropologists
and phi losophers influenced by them have suggested, there are cultures in which the emo-
tion of shame is central and others in which the emotion of guilt is central. As noted in the
Nothing follows about the evaluation of internalized normative guid-

ance. Modes of conscience fueled by fear (or other negative emotions)
can surely distort and cripple human psychological lives, 32 but whether
self-regulation from internalized fears of authorities must always be so
baneful in its effects is by no means clear. The consequences from har-
monious interactions with others can outweigh sacrifices in expressing
selfish desires—indeed, the social involvement may be viewed as a deeper
and more significant articulation of what is properly one’s own set of
wants and aspirations. Much depends, plainly, on the particular orders
that the human with a conscience feels compelled to obey, and whether
they interfere with yearnings central to a person’s life. There are two di-
mensions to the internalized forms of normative guidance, one charac-
terized by the emotional basis through which compliance is obtained and
one depending on the content of the commands. Repressive forms of
conscience can be generated along either dimension, if conscience devel-
ops in unhealthy ways. Social inculcation that couples all deliberation
to fear, shame, and guilt can warp the socialized individuals; equally, mas-
sive prohibitions, however backed by emotional responses, can confine
someone completely. 33 On the other hand, a person whose conscience
expresses itself in a variety of ways, including sometimes through fear,
guilt, and shame, can achieve, and recognize herself as achieving, a richer
emotional life through the social exchanges conscientious cooperation
promotes.34
text, I do not think these exhaust all the possibilities; nor do I think they exclude one an-
other in the ways often suggested.
32. The point is eloquently expressed by Nietzsche in his critique of the “herd moral-
ity” based on ressentiment. How to foster forms of conscience that yield the important
benefits of internalization without deforming individuals is, of course, a question the ethi-
cal project continually has to decide.
33. One way of reading Freud’s Civilization and Its Discontents is to view him as claim-
ing that any way of achieving the measure of social cooperation required for civilization
will have to involve both prohibitions on a massive scale and pervasive negative emotions.
His claims rest on very particu lar ideas about our fundamental desires and drives.
34. This is obviously akin to the Hobbesian perspective on the constructive role of fear
that permeates Leviathan.
§14. Social Embedding

Members of the human groups envisaged (small societies, akin to the
hominid bands preceding them) are socially embedded in two impor-
tant ways. First, as just supposed, the particular way in which normative
guidance is internalized depends upon the training regimes present
within the group. Second, the content of the orders given depends on
discussions among members of the group. The character of the discus-
sions has varied considerably from group to group, time period to time
period, with different degrees of involvement according to age, rank,
and sex. Originally, however, an agreed-on code, articulated and en-
dorsed after discussions around the campfire, 35 was transmitted to the
young through training regimes that had also been socially elaborated
and accepted.
Equality, even a commitment to egalitarianism, was important in the
earliest phases of the ethical project. In formulating the code, the voices
of all adult members of the band needed to be heard: they participated
on equal terms. Moreover, no proposal for regulating conduct could be
accepted unless all those in the group were satisfied with it.
Although these theses may appear implausibly strong, they rest on
three sources of evidence. Anthropological studies of societies whose
ways of life are closest to those of our early human ancestors show the
types of equality ascribed.36 Further, if normative guidance is to resolve
the social tensions, discussions must end old conflicts, not generate new
ones. Lastly, for a small band, one that must work together and unite
against external threats, no adult member is dispensable. These groups
are products of the coalition game, and the dynamics of that game create
egalitarian pressures.
Equality survives in those contemporary groups whose societies are
small and whose relations with neighboring bands are often tense. Our
35. Here my views are close to those of Allan Gibbard, Wise Choices, Apt Feelings (Cam-
bridge, MA: Harvard University Press, 1991).
36. See Christoph Boehm, Hierarchy in the Forest (Cambridge, MA: Harvard Univer-
sity Press, 1999); Richard Lee, The !Kung San (Cambridge, UK: Cambridge University
Press, 1979); Raymond Firth We, The Tikopia (Boston: Beacon, 1961), Marjorie Shostak,
Nisa (Cambridge, MA: Harvard University Press, 1981).
ancestors lived like that until roughly ten to fifteen thousand years ago.
Consequently, more than three-quarters of the period through which
the ethical project has evolved was spent in social circumstances now
quite rare. Small societies reasonably fear the interference and preda-
tions of neighbors. Social cohesion is vital, and no adult can be margin-
alized in normative discussion. As the coalition game (§9) already re-
vealed, the hominid bands out of which early human societies grew
resulted from the partitioning of the physical environment through coali-
tion building. The stability of the partition depends on the approximate
balance among neighboring groups, and, where the groups are small, the
contribution of every member is necessary. Discussions that involve all
adults, that aim to answer to the needs of all adults, and that blur distinc-
tions of rank and ability were crucial to roughly the first forty thousand
years of the ethical project.37
Those discussions would have issued in agreed-upon rules for life
together—but not merely on that. Ethical codes are multidimensional:
besides explicit rules, they involve categories for classifying conduct,
stories that describe exemplary actions (both commended and frowned
upon), patterns of socialization, and habitual forms of behavior. At the
earliest stages, we should think of all these elements as accepted by all
members of the group. Around the campfires, they reached agreement
on precepts, on stories of model behavior, on ways of training the young,
on practices of punishment, on sanctioned habits, perhaps occasionally
on changes in the concepts hitherto employed. This form of socially
embedded normative guidance set the stage for the evolution of the ethi-
cal project.
Ethical codes can pronounce on their own amendment, fi rmly dis-
allowing any possibilities of change or welcoming revisionary discussion.
Perhaps at early stages, there was a common insistence on clear rules,
to be followed obediently and never to be modified. The difficulties of
37. My estimates here are speculative. I suppose that the ethical project began with the
acquisition of full language, at the latest fi fty thousand years ago, and that human societies
were small until, at the earliest, fi fteen thousand years ago. I conclude that the social egali-
tarianism observed in contemporary hunter-gatherers, and the kinds of social discussions
in which they engage, was central to the ethical project for at least the first thirty-five thou-
sand years.
earlier hominid/human social life were surely sufficiently extensive that

initial proposals were incompletely successful, and the social groups
that went furthest in resolving their altruism failures almost certainly
did so by permitting attempts to adjust what had already been achieved.
The codes thus devised and amended are social products: they represent
a joint reaction to the altruism failures previously afflicting the group
and they aim to diminish the frequency of similar failures in the future.
They presuppose the individual capacity for normative guidance, but
how the members are to be guided is a matter for all to decide. The ini-
tial function is to reduce the incidence of altruism failures, and codes are
fashioned by social apprehension of the ways in which cooperation has
broken down.
Does this overemphasize the social character of the ethical project?
According to an alternative—“biological”—hypothesis about the origins
of ethics, not only did our early human ancestors acquire a disposition to
respond to orders—eventually a disposition to command from within—
but also the content of the commands given, rather than being fi xed
through social discussion, embodied shared biases toward particular
kinds of rules. Instead of a capacity for normative guidance to be steered
in various directions, depending on the ways in which altruism failures
are seen as arising (and probably reflecting the actual history of failures
of a particular group), the rival conjecture views individuals as evolu-
tionarily biased toward specific modes of self-command. 38
The biological hypothesis envisions psychological changes. People
acquire dispositions to behave in different ways (perhaps sharing more
frequently than hitherto), and concomitant capacities to feel particular
emotions or to render particular kinds of judgments (negatively directed
toward those who do not share). They are furnished with a moral sense
that redirects some of their own conduct and is expressed in reactions to
the actions of others (and sometimes to their own prior behavior). But
the acquisition of this sense would not yet give rise to the ethical project.
Armed with it, members of the group act more frequently in accordance
with standards we—we who are participants in the ethical project—
38. The type of view considered here is most clearly expressed by Marc Hauser. See his
Moral Minds (New York: Harper Collins, 2006).
approve, but they, the original agents, do not yet have these standards or
yet see a distinction between the behavior they used to exhibit and that
which they now perform. From our perspective they may be more just
than their predecessors, or kindlier perhaps, but this is not an assess-
ment they can make.
For them to initiate the ethical project they must come to see certain
types of behavior as exemplary or particular rules as commanding their
obedience. Could they derive any such recognitional ability from their
own dispositions and capacities, or from reflection on what they are
moved to do? How would they come to see one desire or action-prompting
emotion as different in status from others? They feel many kinds of sen-
timents (although the emotions available to them depend on the social
environments in which they live), but how do they ascertain which ones
belong to the “party of humanity”? 39 To identify something as a genuine
command, they need to distinguish commands from other pressures, and
the most evident possibility is to identify a source—a commander. Given
their environment, the only available source consists of their fellow group
members. If there were an explicit practice of discussing and formulat-
ing rules for the group, they would be able to draw the critical distinc-
tions. Nothing else in their psychology or in the ambient environment
can confer that ability on them. The ethical project can only begin, then,
when normative guidance is socially embedded.40
Even if there are dispositions to behave in ways we think of as ethi-
cally progressive—to refrain from violence, to share more, to comfort
the suffering, or whatever—these are merely “nice tendencies,” ways of
conforming to regularities (regularities the ethical project, once it gets
going, will approve), but they are not abilities to obey rules or precepts.
To be the beginnings of the ethical project they must be coupled to a
capacity to discern and be governed by rules and commands that re-
ceive some sort of authority. The ethical project requires normative
39. I borrow the phrase from Hume, Enquiry Concerning the Principles of Morals (In-
dianapolis, IN: Hackett, 1986), 77. It serves as a useful reminder of the fact that those who
believe in the existence of particu lar moral sentiments—or moral judgments—need to ex-
plain how agents are able to identify which ones these are.
40. There are affi nities between the line of argument in this paragraph and Wittgen-
stein’s famous private-language argument (Philosophical Investigations §§243 ff.).
guidance, and because there are no rival sources of authority to the

group (or some subset of it), it demands that normative guidance be
socially embedded.
The biological hypothesis needs further refinement if it is to illumi-
nate any aspect of the ethical project. For the novel capacities it posits
depend on the social environment.
Consider various forms of the hypothesis. The very strongest would
suppose that human beings acquired a tendency to obey particu lar
kinds of rules— or, more properly, to conform to particu lar kinds of
regularities—quite independently of any social backing for those rules.
So, for example, with respect to sharing behavior, it might declare that,
beyond the limited primate tendencies for sharing, humans acquired a
broader disposition compensating for certain kinds of altruism failures.
As noted, in this story, normative guidance is not playing any important
role; rather, the more extensive human capacities for sharing result from
an extra mechanism for psychological altruism. Possibly, our ancestors
acquired some such additional mechanism, but no such mechanism
could rival the social inculcation of norms in the complex work of enlarg-
ing human cooperative tendencies. That is made plain by the prominent
part ethical reminders, whether self-given or public exhortations, play
in promoting human cooperation—as well as by the controlled experi-
ments on sharing. Effectively, the strong hypothesis must maintain that
the large differences between human and nonhuman forms of psycho-
logically altruistic or behaviorally altruistic behavior come about in two
distinct ways, some from a strengthened version of the tendencies to al-
truism already present in other primates and some from human capaci-
ties for self-command.
Weaker versions of the hypothesis suppose that evolution under natu-
ral selection has equipped people with biases that operate through the
capacity for normative guidance. Perhaps human beings, placed in any
social environment, will develop to feel specific emotions in response
to particular types of behavior—positive emotions to sharing (one’s own
sharing or the sharing actions of others), negative emotions toward fail-
ures to share, for example. Social injunctions that direct sharing will
thus be more likely to “take” than putative rules prescribing more selfish
courses of conduct. At the extreme, it may be supposed that some sets of
commands would be impossible for us to follow; they would be analo-

gous to languages we cannot learn.41
Experiments in sharing reveal that, in the actual environments in
which people grow up, where they acquire from their societies norms
prescribing certain types of sharing, laboratory subjects will share with
others and will punish those who do not share.42 Cross-cultural confir-
mation of the results takes us a little way across the space of potential
environments, but it cannot rule out the possibility that common fea-
tures of contemporary socialization are playing an important causal
role. To demonstrate that contrary behavior is impossible for human
beings would require showing that no environment allows human de-
velopment to follow a different path. Conclusions of that form are no-
toriously hard to defend rigorously, because of our massive ignorance
of the potential environments.43 Additionally, we know already that in
some environments—unhealthy ones, to be sure—the norms we are sup-
posedly predisposed to follow are violated by human behavior. The
ruthlessly self-directed actions of the Ik, the struggles in concentration
camps, and the willingness of subjects in psychological experiments to
infl ict pain on others remind us that, under the right (or, more prop-
erly, the wrong) conditions, the supposedly universal effects will not
be forthcoming.44
41. Hauser (Moral Minds) uses the analogy, and supposes that there is an ethical coun-
terpart to “universal grammar.” For reasons given in the text, I am dubious.
42. The most systematic body of results comes from the work of Fehr and his associates;
see the reference in note 19. Hauser lucidly summarizes this.
43. The problem is exactly analogous to one that bedev ils many sociobiological and ge-
netic determinist claims—the difficulty of extrapolating a norm of reaction from a small
sample of cases. For diagnosis, see Kitcher, Vaulting Ambition: Sociobiology and the Quest
for Human Nature (Cambridge, MA: The MIT Press, 1985) and “Battling the Undead” in
Rama Singh, Costas Krimbas, Diane Paul, and John Beatty (eds) Thinking About Evolu-
tion: Historical, Philosophical and Political Perspectives (Cambridge, UK: Cambridge Uni-
versity Press, 2001, 396–414).
44. See Colin Turnbull, The Mountain People (New York: Simon and Schuster, 1972);
Primo Levi, Survival in Auschwitz (New York: Touchstone Books, 1996); and John Sabini
and Maury Silver, The Moralities of Everyday Life (Oxford, UK: Oxford University Press,
1982). Turnbull’s ethnography is controversial, but unless all his observations are thor-
oughly false, there would still be grounds for wondering about the hypothesis that our
predispositions make contrary norms impossible for us.
Our tendencies to behavior are most likely quite plastic. Given the
hypothetical genomic change that underlies the supposedly broadened
altruistic tendencies, there would probably be a range of dispositions
to action across the (largely uncharted) space of social environments in
which people can live. If the conclusions drawn earlier (§11) about the
explanation of the behavior of subjects in experiments on sharing are
correct, propensities for conduct are likely to depend on the presence of
socially embedded normative guidance and the forms that guidance
takes. The weaker version of the biological hypothesis is implausible so
long as it insists on a specific type of emotional reaction available across
all environments and very particular ways in which that emotional reac-
tion is directed independently of the social milieu.
Far more plausible is the idea that, because of our evolved psychology,
not all attempts to inculcate norms will do equally well. Perhaps we do
have tendencies for emotional responses to types of actions, so that, in
the environments that prevail, following one norm might be uncomfort-
able for us (in the way experimental subjects feel discomfort as they are
following the experimenter’s order to inflict “pain”), while following an-
other might be accompanied by feelings of ease. To modify the linguistic
analogy, given those social environments so far created, some languages
might be more difficult to learn—and some sets of commands similarly
hard to follow. Human evolutionary history may have bequeathed to us
forms of blindness that make reliable compliance with some prescriptions
difficult. Without a proof of impossibility, pragmatism counsels societ-
ies to work hard at training their members to follow the precepts they
deem most important.
Our early human ancestors, equipped with a capacity for normative
guidance, were able to explore various possibilities for social exercise
of that capacity. Those explorations proceed along two dimensions, one
concerned with the ways in which the young are trained in the ethical
code, the other focused on the content of the code. Because we know, as
yet, so little about any biases with which our evolutionary past might have
equipped us, my account will attend to the more visible, social, features
of ethical exploration. To proceed in this way is not to conceive of human
beings as infinitely plastic, or (to switch images) as blank slates on which
societies can write what they please. The history of the ethical project,
from the acquisition of normative guidance to the present, is a history of

experiments, carried out by social groups who sometimes may have faced
difficulties precisely because they rubbed against the grain of human
nature in ways of which neither they nor we are aware.
To recapitulate: hominid societies were confronted with recurrent
altruism failures, a predicament limiting their size and level of coopera-
tion. Through the acquisition of normative guidance and its social
embedding, these failures could be addressed by elaborating ethical
codes. The subsequent ethical project is a sequence of ventures in devel-
oping such codes, in which—as the next chapter will explain—the domi-
nant mechanism is a cultural analogue of natural selection. It is possible
that a small portion of the original altruism failures were corrected by an
alternative mechanism, some strengthening of the altruistic tendencies
already present among primates (although where we have evidence for
any such mechanism, the effects are specific to a range of contexts).45 It is
also possible that human psychological evolution equipped human be-
ings with biases (as yet uncharted) that interfered with or reinforced spe-
cific types of ethical codes. Neither possibility undermines the enterprise
of trying to understand the main features of the cultural evolutionary
process that the acquisition of normative guidance made possible for us.
45. A prime example is the case of cooperation in child care. See the references in note 1.
chapter 3
Experiments of Living
§15. From There to Here

At the dawn of the ethical project, our ancestors lived in bands small
enough so that all adult members could participate in discussions in
which each could speak and all could be heard. Around the campfire, in
the “cool hour,” they sought ways of remedying the altruism failures
from which their social lives had suffered. What kinds of problems did
they discuss?
Scarcity of resources is a likely candidate. Perhaps times have been
hard, and they have often wrangled about the few food items garnered.
Suppose today has been a good day; for once each member of the band
has had plenty to eat. As they gather together and reflect on their recent
squabbles, all of them are able to detach themselves, at least temporarily,
from the difficult circumstances, and think in general about possible
outcomes when the amount available is too small to give everyone what
he or she would like. They imagine possible distributions of that in-
adequate amount, each considering not only his or her own share but
also those of the others, and attempting to recognize the felt consequences
for the others. From their reflections and exchanges comes an agreed-on
104
Experiments of Living 105
vision of which distributions are preferred and a rule enjoining the di-
viding of the spoils.1
Whether or not they would go as smoothly as just supposed, conver-
sations about sharing are readily imaginable. Equally, the discussants
might agree to aim at increasing the food supply, viewing each band
member’s wish to assuage hunger as something to be supported, or they
might all concur in repudiating acts that initiate violence. Socially em-
bedded normative guidance can begin the ethical project, but the pre-
cepts it is likely to generate appear simple and crude. How could the
project of these pioneers blossom into the ethical richness of contempo-
rary life? How did we get from there to here?
There is no serious chance of answering the second question, of de-
fending some narrative as providing the actual evolution of the ethical
project. The clues are too scanty. For the fifty thousand (or more) years
of the ethical project, we have written records only for the last five thou-
sand. Already, at the dawn of writing, elaborate systems of rules are in
place. Evidently, much happened in the Paleolithic and the early Neo-
lithic, leaving only indirect indicators of social change. Knowing the
starting point (the small bands of discussants) and the late phases (ethi-
cal life today and the historical records of the past few millennia), one
can identify what changes occurred, even if it would be folly to pretend
to know how they came about.
Here are some obvious modifications. By five thousand years ago, hu-
man beings had assembled in societies vastly larger than the groups in
1. I do not suggest that the rule agreed on need be the first choice of every band member,
nor that it take any specific form: perhaps it demands equal division, or equal division
among those who have gathered the resources, or division by subunits with special regard
to the needs of younger members. The point is both that there is pressure to agree on some-
thing, and that each of the discussants attempts to accommodate the views of others. This
last point is one way in which the approach I favor diverges from that taken by John Rawls
(A Theory of Justice [Cambridge, MA: Harvard University Press, 1971]). Instead of suppos-
ing that the discussants are rational egoists who consider the consequences for themselves
under conditions of (partial) ignorance, I take them to be psychological altruists, able to
refi ne that psychological altruism in contemplating a general problem that they face, who
deliberate using their knowledge of one another. Further differences lie in the facts that this
is no hypothetical contract, and that it is not directed at any “basic structure of society.”
which the ethical project began. In those large settlements, the egalitari-
anism of the early phases had given way to complex hierarchies. Ethical
life had become entangled with religion. It had also come to address is-
sues beyond the conceptual horizons of the pioneers: citizens of the polis
who inquire into the good life inhabit a different world from those mak-
ing decisions about how to share scarce resources. New roles and in-
stitutions had emerged, generating precepts about property and about
marriage. In more subtle ways, an expansion of the notion of altruism,
beyond the concepts so far considered, permitted new ethical ideas
about human relationships.
There is no doubt that these changes occurred. Acknowledging the
difficulty of explaining how they actually happened leaves pragmatic
naturalism with a problem. Skeptics charge that the account of the ori-
gins of ethical practice works only by changing the subject—something
is shown to emerge, but it is not really ethics. Versions of the accusation
surfaced in the previous chapter: How did we acquire the “ethical point
of view”? How was the commanding voice internalized? How did a sys-
tem of punishment evolve? The questions gain force by sowing doubt
about any possibility of explanation in the terms pragmatic naturalism
permits. No available route leads from there to here.
Doubt is settled by telling a story meeting all the constraints. The
skeptic denies that something is possible, and an adequate response is
to provide a “how possibly” explanation (§2); claiming that this is how
things actually happened is not required. In the previous chapter, the
challenges were turned back by denying the need for any “ethical point
of view” (§11), by offering a scenario for the emergence of systems of
punishment, crude and more refi ned (§12), and by suggesting several
possible ways to build a conscience (§13). The goal of the present chap-
ter is to offer something similar for the rich features of the ethical life
apparently so far removed from the small groups of the pioneers. Not,
then, “how we actually got from there to here” but “how we might have
done so.”
Fundamental to the “how possibly” explanations to be developed
is the increased power of cultural transmission in a species that has
acquired language. The fi rst task is to consider this mechanism for
change.
§16. Cultural Competition

During the first forty thousand years of the ethical project, our species
consisted of a population of small bands, each elaborating a socially em-
bedded mode of normative guidance. Those who framed the first clus-
ters of rules responded to the altruism failures most salient for them, and
perhaps there were already intergroup differences here. Or, even if the
bands shared a common set of altruism failures—troubles in apportion-
ing scarce resources or controlling violence, for example—the com-
mands accepted in social deliberation varied from group to group. Vari-
ation set the stage for a new process: cultural competition.
Differences surely arose both in the content of the rules and in the
systems of socialization and enforcement. For simplicity, consider only
variation in the rules adopted. Assume the population contains bands
with equivalently effective systems of socialization and punishment. One
group declares: food acquired is to be equally shared among all; another:
food is to be shared only among the participants in foraging efforts; yet
another: food is to be divided in accordance with consensus judgments
about effort. Each group has the same expected compliance with whatever
rules it adopts. These bands engage in “experiments of living.”2 Cultural
competition results from the fact that some experiments work more suc-
cessfully than others.
What does lesser or greater success mean here? One measure reca-
pitulates the fundamental currency of Darwinian evolution, the repro-
ductive success of the members of different bands. So the success of an
ethical code is gauged by the extent to which people living in groups
adopting that code leave descendants in subsequent generations.3 That
does not imply that there will be an increasing number of subscribers to
the code in subsequent generations, for greater success in leaving de-
scendants might be offset by a propensity to desert the code. Imagine
two codes, E and F. People in societies adopting E leave, on average,
2. The phrase is John Stuart Mill’s: see On Liberty [Oxford: Oxford University Press
(World’s Classics)], 1998 chap. 3.
3. I shall ignore issues about how far one must look into the future to achieve a reliable
measure for assessing success. Oversimplifying again, we can suppose that the relative
proportions in the first generation of descendants are preserved in subsequent iterations.
three descendants for every two descendants of people living in societies

adopting F; if both societies invariably transmit their code to biological
descendants, and if there is biological competition in which all individu-
als prove equally adept, the E societies will grow at the expense of the F
societies. But if the F societies invariably transmit their code to biological
descendants, while one-sixth of the descendants of people in E societies
migrate to an F society, the proportions will remain constant. Thus, codes
may have one sort of success (in biological reproduction) without having
another (success in commanding adherents).
Cultural competition concerns the latter type of success and is prop-
erly measured by the number and size of the groups in which a code is
adopted.4 Separating cultural success, expressed in terms of adherence,
whether by individuals or by societies, from reproductive success (the
currency of Darwinian evolution) may seem artificial. For, you may sup-
pose, occasions on which the spread of a form of culture is not correlated
with any ability to foster reproductive success are likely to be rare and
transitory. The fantasy of a striking effect in terms of leaving descendants
coupled to, and offset by, a tendency to desert the code is just that—a
fantasy. Tendencies of this sort would be opposed by natural selection:
variants with a disposition not to switch, but to remain with the biologi-
cally more successful culture, would leave more descendants. Hence we
should expect a loose correlation between cultures securing many ad-
herents and cultural practices advancing biological reproduction. In a
famous slogan, “genes hold culture on a leash.”5
4. Here there is further room for decision, for, if a code commands the obedience of a
smaller number of large groups, should it be counted as more or less successful than one
that is followed in a larger number of societies whose combined population is less? It is
important to appreciate the distinct possibilities, but, for the purposes of the current dis-
cussion, no decision about which is the real measure of cultural success is required.
5. The slogan was advanced in classic discussions of human sociobiology—see, for ex-
ample, Charles Lumsden and E. O. Wilson, Genes, Minds, and Culture (Cambridge, MA:
Harvard University Press, 1982). The details of the argument for thinking the slogan is
correct turn out to be much more complex than the simple presentation in the text sug-
gests, and controversial assumptions are required for its derivation [cf. the review by John
Maynard Smith and N. Warren, and chap. 10 of my Vaulting Ambition (Evolution, 6, 1982,
620–27)]. The most important rebuttal of the idea that cultural success and biological suc-
cess are likely to be coupled came from articulated accounts of the coevolution of genes and
Detailed attention to mechanisms of cultural transmission and their

interaction with Darwinian evolution reveals how the advantages of
learning from others allow for processes of imitation, stable under natu-
ral selection and sometimes giving rise to biologically maladaptive ten-
dencies.6 To recognize historical possibilities in the elaboration of ethi-
cal practice, no general account is needed—we can manage without any
sophisticated theory of gene-culture coevolution. It is, however, impor-
tant to appreciate the lack of any tie between biological and cultural suc-
cess. Codes commanding obedience need not be those that further re-
productive success. That important point notwithstanding, on occasion
some Darwinian consequence of a particular ethical code, for example,
the fact that the children of those who subscribe to it tend to survive
and flourish, plays a role in the acceptance of that code by other groups.
Cultural competition does not entail that successful codes march en
bloc from group to group, in the fashion of colonial conquerors. A col-
lection of rules can spread piecemeal, some of its constituent items being
accepted, others rejected. The rule espoused by one band can influence
the code accepted by another, even though the latter group does not take
over the rule intact: we who apportion the spoils of the hunt according to
the perceived contributions of the hunters discover that our neighbors
reward all members of the tribe equally, and are inspired to amend our
practice in a way that combines aspects of both the extant codes (“Di-
vide the gains equally among those who take part in joint projects!”).
New immigrants bring ideas about normative guidance to be aired in
discussions, sometimes modifying extant prescriptions, even when they
are not taken over wholesale.7
culture. A seminal analysis was provided by Robert Boyd and Peter Richerson in Culture
and the Evolutionary Process (Chicago: University of Chicago Press, 1985), and developed
further in several essays in their Origin and Evolution of Cultures (New York: Oxford Uni-
versity Press, 2005).
6. For a succinct explanation of this important thesis, see Boyd and Richerson, Origin
and Evolution of Cultures, 8–11; more technical amplifications are provided in chaps. 1 and
2 of the same book.
7. Although the oversimplification I have adopted invites the thought that individual
rules are the atomic units transmitted in the history of ethical practice, the possibilities of
more subtle influences and modifications reveal that that is inadequate: a search for cultural
“atoms” must slice codes more finely. Moreover, there are good reasons for not thinking of
For the first forty thousand years of the ethical project, small bands of
human beings regulated their lives by socially embedded ethical codes.
Faced with perceived difficulties the extant versions of their codes failed
to address, they tried new ideas. Sometimes, they interacted with other
bands, in whose practices they saw something to inspire revision of their
own rules. Eventually, some groups merged, and aspects of one or both
of the antecedent codes endured in the practice of the subsequent soci-
ety. Some bands simply died out, or dispersed, and their ethical prac-
tices withered with them, even though survivors may have brought fac-
ets of the previous code into the groups they joined. Sometimes new
arrivals, accepted perhaps as mates, brought novel ideas to the camp-
fire discussions, producing a synthesis previously envisaged by neither
of the (“parent”) groups. Processes of these general types (and probably
many more) combined to cause some kinds of rules to be prevalent, others
rare.
The most widely shared features of contemporary ethical codes prob-
ably emerged in many different ways. If human beings have evolved psy-
chological tendencies to acquire certain kinds of norms, a common rule
might reflect these propensities (subject to the qualifications of §14). A
rule might be the simplest response to a difficulty faced by all social
groups. A rule (or a preliminary version) might originate in a single group
and spread to others because it promises to satisfy widely shared de-
sires. Alternatively, groups failing to acquire the rule might suffer some
severe disadvantage, so that they had a tendency to die out or to be
taken over by outsiders. The features of the ethical codes transmitted to
us emerge from these sorts of episodes—and no doubt many more
besides.
cultural practices generally and ethical codes in particu lar as collections of discrete atoms
that can be shuffled and rearranged in novel combinations. Interactions among such al-
leged units may be crucial to the nature of the cultural practice, so that there is no stable,
practice-independent contribution a cultural “atom” generates. Perhaps only in the context
of an entire ethical code does a rule have specific meaning. See, for example, the writings of
Dan Sperber, in particu lar, Explaining Culture (Oxford: Blackwell, 1996). I have tried to
show the complexities of attempts to build serious theories of culture that mimic biological
evolution—theories of the transmission of “memes”— even at the level of kinematics; see my
“Infectious Ideas,” Chapter 10 of Philip Kitcher In Mendel’s Mirror (New York: Oxford
§17. The Unseen Enforcer

To scotch suspicions about pragmatic naturalism’s ability to allow for
the emergence of ethics (“real ethics”), it is necessary (and sufficient!) to
show how processes of cultural competition could have led from simple
early ventures in socially embedded normative guidance to the complex
ethical practices of recent millennia. Begin with the entanglement of
ethics and religion.
Ideally, discussions in the “cool hour” liberate and expand prior ten-
dencies to psychological altruism. Realistically, however, full engage-
ment with others may only rarely (if ever) be achieved. The deliberations
may be conducted by people weary of constant squabbles and yearn-
ing for a consensus that will bring peace. They seek shared rules as a
matter of convenience, hoping to discipline their fellows who lapse from
cooperation—but they are quite ready to break the rules when they think
they can get away with it. The discussants engage in a bargain, giving up
some limitations on the actions they would like to perform (genuinely like
to perform) for the sake of the benefits accruing from similar restraints
imposed upon others.
It may be a good bargain, in that, with a practice of punishment in
place, a significant class of potential altruism failures may be avoided,
simply because onlookers can see what is occurring and enforce the
agreed-upon rules. Yet when other members of the group are in no posi-
tion to check whether you are conforming to the rules, you prefer to dis-
obey. If there is an accusing voice from within, it does not sound with
any great insistence or volume.
The early history of normative guidance was almost certainly one in
which a population of human bands reaped the advantages of public
rules, publicly applied in public situations, but in which many—maybe
all—individual members were willing not to conform to the rules when
they took themselves to be unobserved. Given thoughtful choices by
those who introduced and revised the rules, obedience would typically
contribute to the average reproductive success of the members of the
band—consider, for example, sharing rules that generally ensure food for
everyone. Groups would gain in the Darwinian struggle for existence and
in cultural competition, through socially embedded normative guidance,
even though conformity to the rules was confined to instances in which

actors could expect others to monitor their conduct. Advantages in cul-
tural competition might come about in either, or both, of two ways:
through others’ perception that members of this band could satisfy widely
shared wishes, and through the assessment of them as healthier, better
fed, or whatever proximate cause contributes to the extra Darwinian fit-
ness. An ability to achieve conformity across a broader range of contexts
would yield an extra edge in cultural competition, while typically also
adding to the expected reproductive success of individuals.8 Techniques
for enhancing compliance promote cultural (and probably biological)
success.
What techniques? As they reflect together on their ethical practices,
the deliberators will recognize noncompliance as often caused by the
belief that one is unobserved. They remember rule violators whose con-
fidence that they were not seen—and thus could avoid retribution—turned
out to be false. Within some groups, adult members refine programs for
socializing the young. Perhaps they inculcate enduring fears of the ef-
fects of rule violations, instilling some crude form of conscience (§13) to
keep people on track even when there is no obvious observer around. To
the extent they are effective in doing this, later generations of the band
will tend to comply at higher frequency, with positive Darwinian and
cultural effects. But how exactly is this fear to be triggered? Prevalent in
human cultures—in the successful surviving experiments—is an appeal
to unobservable entities that respond to breaches of ethical codes. West-
ern monotheisms use the device: there is an omniscient deity who ob-
serves all, who judges, and who punishes lapses from commandments.
Variations on the theme occur in most other religious traditions. The
ancestors continue to observe the actions of the descendants and to
8. The plausible assumption that rules directed at remedying altruism failures are likely
to promote the average Darwinian fitness of individuals as well as to enhance the cultural
competitiveness of the ethical practices realized in social groups enables me to sidestep
worries that the forces of cultural and Darwinian competition might tug in different direc-
tions. A more general account would explore the ways in which rules antithetical to average
Darwinian fitness might be supported because of their efficacy in cultural competition, but
my purposes here can be satisfied with far less. Boyd and Richerson, Origin and Evolution
of Cultures, provide the elements out of which the more general account can be fashioned.
retract their favors if the commands are broken. Spirits are associated
with particular places or particular animals and will wreak vengeance
on the group if rules are violated. There are hidden forces in nature,
with which people must align themselves to be successful, and to deviate
from the prescriptions is to endanger or destroy this alignment. Ethnog-
raphies testify to the popularity of the idea of unobserved enforcement
(typically, but not always, personified): as when informants tell of an “all-
father” who “from his residence in the sky watches the actions of men
[and] ‘is very angry when they do things they ought not to do, as when
they eat forbidden food.’ ”9
Once the idea of an unseen enforcer is in place, fear of punishment
can be embedded in a complex constellation of emotional responses.
Commands promulgated by elders can be identified with the wishes of
the gods or spirits (or with the tendencies of the impersonal forces affect-
ing human success). If the gods are local, they may be seen as prescrib-
ing particular rules for the group, rules that both express the favor of
the deities and constitute the identity of the band. Later phases of a
group’s ethical practice look back on an episode in which the ancestors
obtained the favor of a particular divinity and were also given the divine
command(ment)s.10 Crude fear of punishment is transmuted into more
positive emotions—awe, reverence—and the commands are welcomed
as a mark of the favor of an extraordinary being. Group members see the
rules as constitutive of who they are.
Religious beliefs, beliefs in some kind of “transcendent” reality, tied
to the origin and reinforcement of ethical prescriptions, are almost uni-
versal across known human societies—at least until recently. Why is this?
As noted (§16), there are many possible ways to succeed in cultural com-
petition, and the would-be explainer of the prevalence of any aspect of
9. Edward Westermarck, The Origin and Development of the Moral Ideas, 2nd ed.
(London: Macmillan, 1926), Vol. 2, 671; the entire chapter (chap. 50) is full of fascinating
examples of “Gods as Guardians of Morality.”
10. This conception is plainly present in the earliest versions of the Judaic tradition;
moreover, as the preambles to the legal codes of the ancient Near East make very clear, very
similar ideas appear in societies throughout Mesopotamia and Egypt. I conjecture that
these are simply written out versions of oral traditions that thrived and developed over tens
of thousands of years.
human culture does well to tread cautiously. In this instance, however,

there are grounds for tentatively embracing a historical conjecture: reli-
gion permeates human history because groups that did not invent some
form of the unseen enforcer were less able to reap the benefits (Darwin-
ian and cultural) of socially embedded normative guidance; with lower
levels of cooperation and social harmony, they were losers in cultural
competition.
A rival possibility is that Darwinian selection has generated a propen-
sity for conceiving and adopting ideas about transcendent entities who
are both the sources of prescriptions and the supervisors of conduct. Yet
any thought of a genetic variation inclining individuals to so specific a
form of religious belief is utterly implausible: variations, whether point
mutations or shuffl ings of the genome, produce, as their proximate ef-
fects, differences in the structures or relative proportions of proteins
present in cells, and this kind of change, inserted into some early human
environment, could not yield so particular an effect. Nor is cultural dif-
fusion of the idea of the unseen enforcer from some ur-society in which it
was first articulated at all likely. Groups lacking this idea, learning of the
stories told by outsiders about how beings who especially favored them
commanded them to behave according to their particular rules, would
hardly be inspired to think that the structure of the account, though
none of the details, was applicable to their own case.
The proposed conjecture is far simpler. In a world of apparently
unpredictable phenomena and seemingly inexplicable changes, our an-
cestors responded by invoking unseen entities with extensive powers.11
Some groups took a further step, attributing to these beings a connec-
tion with the social order: impersonal forces would react against those
who broke the rules, ancestors or spirits would wreak vengeance on those
who failed to conform to the code, deities expressed their wishes in
the commands recognized by the band and were able to inspect behav-
ior, even when agents conceived of themselves as “alone.” Groups who
11. Dewey favors an anthropological account along these lines, viewing it as the ultimate
source not only of religion, but also of philosophy and of science. See the opening chapters
of The Quest for Certainty vol. 4 of John Dewey: The Later Works (Carbondale IL: Univer-
sity of Southern Illinois Press). See also Pascal Boyer, Religion Explained (New York: Ba-
sic Books, 2001), and Daniel Dennett, Breaking the Spell (New York: Penguin, 2006).
took this step gained a powerful mechanism for securing compliance

and did better than rivals whose invocation of unseen powers was not
connected to the ethical sphere. Religiously entangled ethics is preva-
lent because the very specific link between unseen powers and ethical
conduct bestows significant advantages in cultural competition.
Philosophers have often been unsympathetic to the almost universal
historical embedding of ethics within religion. Their arguments, articu-
lated from Plato on, demolish the thesis that religion can provide a par-
ticular type of foundation for ethics.12 They do not, however, touch the
thought that religion may be valuable to, even essential for, ethical prac-
tice, in virtue of its power to increase compliance. Far from being an ir-
rational idiosyncrasy, divine-command approaches to ethics may reflect
a deep fact about cultural competition. Yet, for all the short-term advan-
tages it brings, invoking an unseen enforcer amends the ethical project
in potentially dangerous ways. For it threatens the equality that origi-
nally reigned in normative deliberations. Those who can convincingly
claim to have special access to the will of the transcendent policeman—
shamans, priests, and saints—come to have an ethical authority others
lack.
Our next task is to examine the breakdown of initial equality more
generally, considering divisions by status and role, and the origin of in-
stitutions that expand those divisions. How might the ethical project
have introduced, tolerated, even favored these differences?
§18. Some Dots to Be Connected

For tens of thousands of years, egalitarian distribution of basic resources
was crucial to the ethical project. Vulnerable small groups required the
participation of all adults. They surely deployed precursors of the many
clever strategies contemporary hunter-gatherers use to promote equality
among their members. The !Kung, for example, take steps to ensure that
differences in hunting ability are not manifest. They impose serious
sanctions for boasting about a kill, cultivate a practice of joking designed
to check feelings of pride and arrogance, and have a custom of crediting
12. The exact character of the arguments will occupy us in §27.

the kill to the owner of the arrow, which, when combined with a wide-
spread practice of arrow sharing, effectively reduces differences in hunt-
ing yield.13 Violation of these conventions is regarded as a way of court-
ing bad luck. Under the circumstances of early human life, groups failing
to develop similar strategies would forfeit the advantages normative
guidance had brought.14 The societies visible in the first written records,
however, contain fine differentiation of rank and status. What might
have produced them?
Archeological evidence of early cities (Jericho, Çatal Hüyük) makes it
apparent that, by eight thousand years ago, human beings were able to
live in groups far larger than those present at the early stages of the ethi-
cal project.15 When a thousand or more people live within the walls of
the same city, strategies of peacemaking through face-to-face reassur-
ance are no longer applicable. There must be a system of agreed-upon
rules for forestalling potential conflicts and for dealing with people who
are relative strangers. Some extension of the prevailing injunctions to
cover transactions with individuals outside the small group of regular
associates must have been achieved substantially earlier. By fifteen thou-
sand years ago, at the very latest, bands of human beings were periodi-
cally uniting temporarily, for the deposits at some sites testify to a larger
association.16 Moreover, there is indirect evidence for peaceful intergroup
associations at earlier stages—and possibly even for the existence of
trade between different bands.
13. See Richard Lee, The !Kung San (Cambridge, UK: Cambridge University Press,
1979).
14. For general discussion of the importance of egalitarianism, see Cristoph Boehm,
Hierarchy in the Forest (Cambridge, MA: Harvard University Press, 1999), and for telling
examples see Lee,!Kung San. The case for a period of egalitarianism in human prehistory,
between the hierarchies of apelike hominids and those of the societies for which we have
historical records, is succinctly made by B. M. Knauft, “Violence and Sociality in Human
Evolution,” Current Anthropology 32 (1991): 391–428; see, in particular, the famous U-shaped
curve.
15. See James Mellaart, Çatal Hüyük (New York: McGraw-Hill, 1967).
16. See Clive Gamble, The Palaeolithic Societies of Europe (Cambridge, UK: Cam-
bridge University Press, 1999), chap. 8; Paul Mellars, “The Upper Palaeolithic Revolu-
tion,” in The Oxford Illustrated Prehistory of Europe, ed. Barry Cunliffe (Oxford, UK:
Oxford University Press, 1994), 42–78.
The earliest hominid technologies were disposable. Two hundred

thousand years ago our ancestors made tools as they needed them and
left them behind when they moved on. For them, tools posed no impor-
tant constraint on mobility (people did not need carrying gear), nor did
tools figure as a type of property (if someone takes an ax, the maker can
easily replace it). But as hominids dispersed, they frequently left the
sources of their tools behind them, and, by twenty thousand years ago,
bands were foraging in regions a significant distance (a hundred kilo-
meters or more) from the nearest places in which raw materials for their
tools were found (the case of tools made from obsidian is particularly
striking). Those bands would have needed carrying devices (for under-
standable reasons, not preserved in the record), and they would also have
needed to coordinate their behavior with one another and with other
bands so as to make possible either a long-distance trade network or a se-
ries of journeys to gather the materials they required.17 Either instance
threatens obvious possibilities of exploitation and aggressive interven-
tion, and the codes of the groups involved would have had to be modi-
fied to cope with these dangers. Even if they were not yet practicing trade
with one another, their ethical codes would have had to contain rules
that forbade harming outsiders, at least under some circumstances. Rules
of this sort anticipate the possibilities that flower in the later cities, in
Jericho and Çatal Hüyük, Ur, Uruk, and Babylon.
Long before people came together to build pyramids or ziggurats, our
ancestors were crafting tools that depended on distant materials, bring-
ing special substances deep into caves to paint animals, and burying their
dead with special artifacts. By fifteen thousand years ago, human groups
were fashioning statues and leaving them at grave sites, a practice hard to
explain without supposing conceptions of transcendent beings whose
welfare is a matter of practical concern. Thousands of years earlier,
people took time to isolate the pigments needed for decorating the walls
17. The hypothesis of Paleolithic trading networks was originally advanced by Colin
Renfrew and his colleagues, based on the discovery of obsidian tools at considerable dis-
tances from the nearest source. See C. Renfrew and S. Shennan, eds., Ranking, Resource,
and Exchange (Cambridge, UK: Cambridge University Press, 1982). The trade-network
hypothesis seems superior to the rival idea of long journeys undertaken by members of a
band, given the obvious problem of explaining how such journeys might be navigated.
of caves, developing techniques of painting, and producing the extraor-

dinary art of the French and Spanish sites.18 These activities are unlikely
in any society still struggling to satisfy basic requirements of food and
shelter, improbable also if there is not some incipient division of labor.
By thirty thousand years before the present, the enterprise of framing
rules for life together, the ethical project, must have been quite well
developed.
The early law codes provide the clearest indications of the evolution
of ethical codes that occurred late in prehistory. Ancient Near Eastern
texts include stories embodying ideals of behavior, myths about the
afterlife, and partial codes of laws. The Gilgamesh epic, for example,
provides a picture of what is expected of high-ranking people in the
pyramidal societies of Sumer and Babylon; similarly, the protestation
of innocence in the Egyptian Book of the Dead19 shows us what kinds of
actions were counted as ethical transgressions and thus illuminates the
structure of the ethical code; most obviously, the lists of rules found in
the Mesopotamian codes, from the Lipit-Ishtar code of the early second
millennium, through the code of Hammurabi (a century later) and be-
yond provide us with a sense of the conduct requiring explicit prohibi-
tion and of the relative importance of various social breaches.
The preambles to the law codes constantly emphasize that the law-
giver brings peace and resolution of conflicts; the law is seen as a method
of transcending a social life in which brute force prevails and the strong
oppress the weak. The surviving tablets and stelae do not offer any com-
plete account of the laws in force. They amend a body of existing law,
offering revisions and extensions that address problems arising in the
creation of social order. These “codes” represent a multistage process
of development of social rules extending back to the dawn of writing and
beyond. Their fragmentary character is immediately obvious. Provisions
are made for very specific types of occurrence—whether a “senior” strikes
the daughter of another “senior” and causes a miscarriage, whether an ox
18. See Jean-Marie Chauvet, Eliette Brunel Deschamps, and Christian Hillaire, Dawn
of Art (New York: Harry N. Abrams, 1996).
19. The Egyptian Book of the Dead, trans. E. A. Wallis Budge (New York: Dover, 1967),
194– 98; also in James B. Pritchard, ed., Ancient Near Eastern Texts (Princeton, NJ: Prince-
ton University Press, 1969), 34–36.
gores a passerby, whether a woman crushes the testicle(s) of a man who

is fighting her husband. The particularity points to new troubles of an
increasingly complex society.
The Neolithic pastoralists and farmers of Mesopotamia had already
worked out rules for restraining violence, protecting the fruits of their
labors, and organizing sexual relations. As they were integrated into
larger units in a world dependent on social coordination to supply ade-
quate irrigation, new issues arose—how are measures to be standard-
ized, how does one ensure that land is properly used, how are the public
canals and dikes to be maintained? The surviving codes lavish great
detail on these questions, as well as addressing the various kinds of
violence and sexual relations that emerged from the social friction of
large numbers of people occupying a relatively small space. They occur
against the background of a general understanding of the ways in which
violence is to be contained, sexual relations regulated, and property
protected.
Later diff usion of rules from the Babylonian codes reveals the cultural
transmission prevalent throughout prehistory (although diff usion would
have gone swiftly only once the ethical project had evolved to allow peace-
ful interactions among bands). The Hebrew Bible takes over parts of the
law we find in Sumer and Babylon: Exodus 21:28–29—concerned with
control of oxen—recapitulates articles 250–51 of the code of Hammu-
rabi, and Deuteronomy 25:11–12 reaffirms the Mesopotamian prohibi-
tion against wifely testicle crushing. Mesopotamian theocracies plainly
had complex rules for religious ritual and ser vice to the gods (or their
surrogates, the ruler-priests). The code of Lipit-Ishtar already links the
law to divine command, and the Egyptian Book of the Dead sees the pros-
pects of the afterlife as dependent on present conduct. The idea of the
unseen enforcer permeates all these texts.20
20. There is even a Babylonian wisdom literature, dating to before 700 b.c.e. (possibly
to significantly before this period), in which the attitudes we associate with Christian ethi-
cal conceptions are articulated. It reads: “Unto your opponent do no evil / Your evildoer
recompense with good; / Unto your enemy let justice [be done].” Another text, of uncertain
date but possibly very early, offers the same theme: “Do not return evil to your adversary; /
Requite with kindness the one who does evil to you.” There is no reason to think that the
authors of these texts invented the idea. They, too, like the writers who borrowed the
This fragmentary record provides a small number of dots that must

be connected by any adequate account of how the early phases of the
ethical project could have given rise to the ethical practices of the pres-
ent and the historical past. The pieces of evidence constrain a “how pos-
sibly” account but are insufficient to yield any confidence that only one
narrative will accommodate them (§2). The following sections construct
a potential explanation, answerable to the demands of pragmatic natu-
ralism, tracing the emergence of social divisions, trade, the institution
of private property, and ultimately of societies in which the most privi-
leged can speculate about the good life.
One important point needs advance consideration. Previous para-
graphs have focused on the first written legal codes, as if these offered
insight into ethical practice. Yet, as every beginning philosophy student
learns, legal and ethical prescriptions are quite different: there are laws
for which compliance is not an ethical matter, as well as ethical maxims not
translated into law. To trace the possible evolution of the ethical project,
is it legitimate to begin with social discussions of regulations for conduct
and end with legal codes?21
For our purposes, boundaries should be blurred. Almost all societies,
at almost all times, have socialized new members by inculcating more
than ethical resources—at least as contemporary philosophy under-
stands the ethical. The young are informed about what is a matter of
religious duty, what is a matter of law, what is a matter of politeness and
social custom—that is, what we see as falling under these categories.
The specific conception of the ethical figuring in philosophical discus-
sions grows out of a historical process. Later judgments distinguishing
theme of forgiveness from them, probably drew on previous traditions. The ethical codes
of prehistory survive in these early texts, accompanied by regulations that deal with novel
problems.
21. Many prominent thinkers have been willing to advance views about law quite simi-
lar to my proposals about ethics: witness H. J. S. Maine, Ancient Law (Tucson: University
of Arizona Press, 1986); Benjamin Cardozo, The Growth of the Law (New Haven, CT: Yale
University Press, 1924); and H. L. A. Hart, The Concept of Law (Oxford, UK: Oxford Uni-
versity Press, 1961). Like Dewey, I take the kinship very seriously (as did Cardozo). For
encouragement to think about the development of law, I am indebted to Jeremy Waldron,
and I am also grateful to Sam Rothschild for valuable conversations.
some obligations as ethical emerge from earlier practices blind to differ-

ences among categories of norms.
Divisions of rules into types often makes sense, for rules can conflict
and it is sometimes (though not always) handy to supply a general way of
deciding what has priority. People are commanded to engage in a partic-
ular ritual, but, in the middle of the ceremony, participants hear that the
lives of other group members are threatened and immediate attention
is required. Does the rule to protect others in the band take priority over
the prescription to finish the ritual? Different groups may decide that
question differently. Some, perhaps the culturally most successful ones,
declare that the command to aid and protect has higher status than the
rule requiring the ritual to be carried through to its conclusion. Many
societies, contemporary and historical, have divided prescriptions into
three (rough) categories. The most fundamental are the commands as-
sociated with transcendent beings; these can be used to elaborate, and
sometimes override, rules emerging from social discussion, (something
like matters of law); both categories take priority over the least important
directives, those taken to govern manners and customs. Division into these
categories does not settle all issues of priority, for it is possible for two
divine commandments to conflict (the rules for worship clash with pre-
scriptions to save others).
These categories, and the ways of deploying them in subordinating
some rules to others, are products of the cultural elaboration of nor-
mative guidance. There is no inevitability about the outcome. Commit-
ment to a particular hierarchy of types of norms—or, indeed, to any such
hierarchy—is a matter for potential scrutiny. (There might be an invari-
ant relation among types of commands—type 1 always takes precedence
over type 2—or the relative status could be context dependent.) Occa-
sions of conflict among norms provide a spur to the practice of differen-
tiating types of norms, including the very particular practice contempo-
rary philosophy sees as constitutive of ethics.22
22. It is worth noting that, even within traditions differentiating the ethical from matters
of law, religion, and etiquette, some voices speak differently. Pioneering secularists of the
past (Hume, Adam Smith) often seem to blur the distinction between morals and man-
ners: eighteenth-century accounts of moral sentiments surprise readers by grouping wit,
cheerfulness, and elegance with honesty and generosity. Even thinkers who allow for a
§19. Divisions of Labor

The ethical pioneers lived in the fashion of contemporary hunter-
gatherers, in egalitarian societies where almost all adults carry out the
same range of daily tasks to ensure individual and social survival, and
in which the contributions of all are typically necessary (§§14, 18). Per-
haps there was a modest sexual division of labor, centered on special
involvement of women with young children and (perhaps) different
types of foraging. Deliberations about how to share scarce resources
surely acknowledged the basic desires of all members of the band, and
endorsed those desires, in the sense of preferring everyone’s desires to
be satisfied provided there is enough to go round. Attitudes of endorse-
ment create pressure to transform conditions of scarcity into a state
of greater abundance. Moreover, to the extent that more resources are
available to the group, the task of avoiding altruism failure becomes
easier. Societies with codes fostering cooperation are more likely to en-
gage in joint projects that garner valuable resources, and the increase in
resources yields an enlarged class of occasions on which a socially en-
dorsed outcome is readily seen and relatively attractive, thus promoting
more cooperation. As groups develop new strategies for cooperative
projects that increase their joint resources, they enter a feed-forward
cycle.
One obvious strategy for obtaining more of the things everybody
needs is a form of the division of labor. If one of us is better at finding
roots and another makes superior arrows, we are likely to acquire more
food items, or to acquire the same amount in a shorter time, if the first
person concentrates on root finding and the second on arrow making.
This is not yet a matter of decomposing tasks into subroutines and
separation between religious commands and the requirements of ethics do not always as-
sign priority to the ethical. Kierkegaard is famous (notorious?) for maintaining that the
greatness of Abraham, as “the knight of faith,” consists in his “suspension of the ethical.”
To suppose he can be refuted by declaring that it is constitutive of ethical maxims to take
priority over religious injunctions is no more convincing than specifying that your favorite
rule (or strategy) of nondeductive inference must be adopted because it is constitutive of
rationality. Better to try to understand why one, rather than the other, might be incorpo-
rated in our practices.
assigning people to repeat particular actions.23 Rather, the spectrum of

jobs is partitioned to take advantage of the distribution of skills. Imple-
menting the strategy depends on each individual’s doing his or her part,
and that requires normative guidance to constrain potential shirkers.
Under the aegis of normative guidance, however, the strategy may be to
everyone’s benefit: the group finds more roots, and its sharper arrows
bring down more (or larger) prey; through equal division of the spoils,
everyone gains a bigger share.
Sometimes the environment in which the band finds itself is benign,
and refined productive strategies are unnecessary. For many, perhaps
all, of the Paleolithic bands, times were surely sometimes hard and divi-
sion of labor correspondingly important. Recognizing the possibility of
hard times, some groups could have instituted a practice of storing re-
serves for future conditions in which even the most efficient distribution
of tasks would bring in less than they needed. In a fluctuating environ-
ment, division of labor, accompanied by a practice of storing a surplus
when life goes relatively easily, promotes both the Darwinian fitness of
members of the band and also contributes to the satisfaction of basic de-
sires across a broader period of time.
So far, the bands remain egalitarian: the requisites of life are found or
made, and divided among the members according to their needs. The
ethical code expands to regulate collective activity, requiring individu-
als to carry out the tasks assigned to them and to labor to acquire more
than is needed to meet current demands. But groups committed to the
production of surplus resources prepare the way for a second division of
labor. Suppose the demands of a variable environment are sufficiently
rigorous that those who do not save do not survive. The result will be a
population of small bands all practicing surplus production. Relations
among these bands may be wary and suspicious, even aggressively hos-
tile. Nevertheless, if neighboring bands are well matched in size and
strength, they will see that little is to be gained by attempts to encroach
on others’ territory or to take advantage of others’ resources. It may also
become clear that adjacent groups have analogs of the distribution of
23. As envisaged by Adam Smith, Wealth of Nations (New York: Modern Library,
2000), book 1, chap. 1.
skills that made intragroup division of labor profitable. Because of fea-

tures of the local environment, or because of specially developed tech-
niques, one group may have food, or carrying gear, or tools of kinds or
quality the other does not possess. Without endangering their respective
abilities to survive periods of environmental challenge, each can recog-
nize a gain in variety of resources—and in overall value of the stock of
resources—from giving up some part of its surplus to acquire surplus
items from its neighbor. Trade is born.24
Once trade begins, there is an impetus to exploit further the initial di-
vision of labor. If our group is to exchange with the band across the river,
and if we are to retain enough to ensure our survival through possible hard
times ahead, we shall have to have more than more than enough. More-
over, the stable pursuit of trade will require a new form of cooperation,
peaceful interaction among individuals from different bands whose an-
cestors viewed one another with suspicion or even hostility. There are new
demands on the versions of normative guidance practiced in the neigh-
borhood, additions to the ethical code that regulate behavior with respect
to people not previously considered within the framework of commands.
This is an important step. With the addition of norms governing interac-
tions with members of other bands, the set of people covered by normative
guidance is extended—the circle expands.
This scenario leads from the original small societies, with their rules
for remedying the failures of altruism within the group, to later commu-
nities, still small, each of which has a stock of collective resources and
each of which engages in limited interactions with neighbors. These com-
munities have to extend the division of labor in ways that appear small
but are socially and culturally consequential. First, the need to preserve
their stock for hard times and to use part of it in barter with neighbors is
likely to bring new forms of work. More important, the performance of
some of the tasks now carried out requires particular tools or equip-
ment, and if some members of the band spend large periods of time on
24. In telling this story, I diverge from Smith, who appeals to an innate propensity to
“truck and barter.” Unlike Smith, I have also not assumed that the effect of the division of
labor must be greater productivity—some groups may settle for less work and more leisure.
See Smith, Wealth of Nations, book 1, chap. 2, esp. 16.
the pertinent tasks, while others never do, it will be important to ensure
that the users’ access to the needed implements is not impeded by the
activities of nonusers. These developments introduce an embryonic con-
ception of property.
Even though the basic resources of the group (food, materials for shel-
ter, and so on) are divided, the group as a whole owns the surplus. This
is to be used in hard times and to be available for exchange, and it will be
important to defend it against those who would take it. The rules allow-
ing for trade must specify that items brought by one trading partner are
not simply to be seized by others. Similarly, if the equipment required
for performing a particular task is to be available to one group member
and not to others, there must be rules allowing the user to keep it and
forbidding nonusers to interfere. Those rules will not yet permit the user
to transfer the items in question to anyone else he or she chooses—rather,
they insist that tools be passed on to the next performer of the task—but
they will protect a temporary power to employ the equipment. Commu-
nal property and a limited form of individual property have emerged.
§20. Roles, Rules, and Institutions

Divisions of labor assign different tasks to different members of the com-
munity and thereby create roles. The band relies on one individual in
tracking game, on another in negotiating with neighbors, on yet another
in finding or constructing shelter. For these roles to be filled efficiently,
their occupants have to be well selected and to behave in ways promot-
ing the ends toward which the role is directed. Groups will make better
use of divided labor to the extent they are able to identify reliably the
physical and psychological capacities needed for a particu lar role, and
to articulate and enforce rules applying to those who occupy the role.
Even though resources acquired through the division of labor con-
tinue to be split equally among group members, the search for an effi-
cient way to apportion roles already brings the beginning of social ine-
quality. Members of the band must attend to individual differences, to
the “talents” that some have and others lack, that fit some for particular
roles. The simplest ethical prescriptions applying to those roles embody
the band’s acquired knowledge of how the tasks are best performed,
enjoining care, for example, at stages known to be especially crucial. Yet

many kinds of performance improve with practice and training, and
with respect to these the assignment of roles will go better insofar as the
assigners can spot incipient talent and subject its development to rules.
Within the socialization of youth, there can arise an appreciation of
difference that is valuable to the community, provided rules for self-
development are in place.
Becoming good at tracking game or making carrying devices requires
the novice to obey the instructions of those who know how and to be
diligent in carry ing out the exercises prescribed. Some rules govern-
ing training will be specific to the task. Others will apply across a wide
range of roles. The young are to be obedient and not willful, attentive
and not distractible, industrious and not lazy. More general still is a form
of prescription combining the idea of differences in propensities for nec-
essary forms of work with the general characteristics required for profi-
ciency in any role: “Develop your talents!” Coordinated group projects
thus exert pressures on individual performance, prompting the appre-
ciation of important virtues—industry, courage, prudence, temperance.
Although normative guidance began as a remedy for altruism failures,
the ethical codes found in the historical record, from ancient times to the
present, contain directives to act in ways without evident impact on the
lives of others: people are supposed to be prudent and resolute, even
when their imprudence or irresolution would affect only themselves.
Where do self-regarding principles come from? A possible answer: divi-
sions of labor introduce the conception of differences in talent, and such
differences are potentially valuable in promoting group welfare; the ben-
efit requires talents to be properly developed; once that is understood,
ethical codes elaborate to enjoin the development of promise (with de-
rivative rules forbidding laziness). Lurking in the background is still a
connection to the original goals, for neglecting one’s self-development
can be viewed as a kind of altruism failure. But a more personal basis can
ensue.
Attention to differences in the propensities of members of a social
group probably began long before the articulation of normative guid-
ance. Strategies for playing optional games require recognition of the
characteristics of potential partners (§8): at a bare minimum, you have to
discriminate those who have been unreliable in the past, and it may also
pay to choose individuals with particular physical traits.25 With devel-
oped division of labor, finer scrutiny became pertinent. Scrutiny starts
to perturb the egalitarian attitudes of the group—for some of the tasks
assigned can be carried out successfully by any of a relatively large sub-
group (they require no special talent or training), while others may be
difficult to fill well. Roles are more or less demanding, according as they
have more or fewer potential occupants.
Suppose a particular role—tracking, say—is both demanding in this
sense and requires extensive training and effort in performance. Good
potential trackers are rare, their skills need to be honed over a relatively
long time, and their tracking activities require searching attentively
through a broad area. The group inculcates stringent rules for learning
and performing this role. The solitary nature of the task, however, makes
enforcement difficult, and the bare idea of unseen enforcement may not
prove enough. A new idea is added: significant contribution of success-
ful tracking to the tribal stock is particularly favored by the entities who
are the source of the ethical code. Especially pleasing to the gods are
those with rare talents, who develop those talents to the full and use
them energetically in ser vice of the common good. In societies elabo-
rating the division of labor in this fashion, new human desires readily
emerge: people come to want the approbation, even admiration of their
fellows; they wish to enjoy the favor of the gods.
So elaborated, the code begins to advance a new conception of the
good human life. Its earlier forms identified the common good in terms
of basic desires, viewing human lives as going well when those basic de-
sires were met. Early stages of the ethical project introduced rules whose
intended effect was to improve the prospect of satisfying more of the ba-
sic desires and thus living better. Introducing the unseen enforcer con-
nected the rules with the wishes of a great being with special concern for
25. Consider a joint hunting venture. If two equally reliable partners are available, and
one is quicker than the other, that person may be a better bet for bringing home the game.
That can easily be offset by other considerations: if the slower individual is a longtime as-
sociate, and failure to interact on this occasion would prompt the person then to refuse
future invitations, the physical superiority of the rival candidate is better ignored.
the group.26 Now the development and exercise of rare talent—in ser vice
of meeting the basic desires of members of the band—is seen as favored
by this being. Well-socialized group members want this sort of approval.
For the specially gifted, at least, to live well involves gaining the being’s
favor.
Once the broader idea of a good human life has been introduced as
an incentive for the rarely talented, it can be extended to others. Al-
though it is less important to encourage those who fi ll less demanding
roles, general diligence benefits the band. The myth that divine ap-
proval descends on those who fit themselves to their station and dis-
charge its duties with energy is a valuable extension of the idea of un-
seen enforcement. It detaches the rules enjoining development of talents
from their derivative status, as consequences of more general principles
about contributing to cooperative projects, and locates them in the
direct command of the unseen lawgiver. Perfecting one’s talents may
contribute to the success of the group, but it is required of each member
because it is the divine will.27
Cultural competition can favor an evolutionary transition from an
initial stage at which ethical codes are directed only at altruism failures
to more internally complex societies, with divisions of labor, prescrip-
tions for interactions with members outside the group, specific roles, rules
for carrying out those roles, and injunctions for behavior even when the
impact on others is not of central concern. At this stage in the ethical
project, continued discussion of the prevalent ethical code will some-
times need to consider the institutions of the group: prescribed pat-
terns of behavior focused on some domain of the band’s life. My “how
possibly” story concludes with the emergence and evolution of an insti-
tution presupposed by well-known ethical maxims: property.28
26. To keep things simple, I offer a version in which unseen enforcement is personified.
27. Once the injunction to develop talent has been detached from the consequences for
society, it can be maintained as a freestanding, self-regarding maxim, even when the idea of
a divine backing for it is abandoned. The attitude of the citizens of Plato’s Republic is,
presumably, one of seeing lives as good in terms of the perfection of talent—for this is what
the orga nization of the city aims to do— even though they have read the Euthyphro.
28. Both the seventh commandment and the tenth presuppose the institution of private
property.
Trade already introduces a notion of communal property, for the

partners each have to subscribe to rules forbidding violent appropria-
tion of the other’s resources. Division of labor gives rise to a weak con-
ception of private property within the group, in that those assigned to
certain tasks are awarded privileged access to whatever equipment those
tasks require. The privilege depends on proper use by those who occupy
the relevant roles: the maker of carrying equipment or the digger for roots
is not free to abandon, or blunt, or misuse the special hand ax or the
spade—or even to let the tools sit idly by; the items assigned are to be
employed in the pertinent tasks, and, after the labor is over, to be pre-
served in a socially expected modification of their original form.29 Nor
can the user transfer the equipment to anyone he or she chooses; once
the user’s career in this role has ended, the new occupants of the role will
acquire the privileges he or she now enjoys.
How might a stronger conception of private property, one that allows
owners to use and dispose of their belongings as they choose, have
emerged? One obvious thought: productive performers of particular
tasks, particularly tasks playing a large role in the group’s life, might be
rewarded by giving them power to dispose of resources previously owned
by all; their production might be encouraged by such rewards. 30 So long
as one considers contexts of privileged access to tools, any transition of
this sort looks mysterious, for it is hard to conceive how a power to dis-
pose of equipment previously owned by all might be any significant re-
ward or motivation to diligent exertion. If we examine stages close to the
historical present, the mysteries dissolve.
By ten thousand years before the present (five thousand years before
the invention of writing), our ancestors had learned how to domesti-
cate plants and animals. The plants and animals owned by a group
29. Often, tools will be supposed to show a little, but only a little wear; there may be
occasions, however, when extensive per for mance involves considerable damage to a tool.
The point is that the extent to which this is to occur is not for the user to decide.
30. This is a standard conception in the history of political economy (developed by
Smith and by many others), and the thought that some kind of private property is a neces-
sary incentive for hard work occurs even in a thinker as worried about the notion as John
Stuart Mill (see Principles of Political Economy, Works [Toronto: University of Toronto
Press, 1963], 2:207, 2:225–26, 3:742–55).
constitute resources temporarily in the charge of occupants of specific

roles—tenders of gardens, herders. Unlike the equipment assigned in a
privileged fashion to particular workers, these resources are attractive
potential rewards. Imagine a social innovation. Within a pastoral soci-
ety, the deliberators resolve that shepherds may hold back a small frac-
tion of the healthy lambs born, disposing of them as they see fit. Assum-
ing the members of the group would like to have this new power, 31 the
shepherds would be motivated to work in ways conducive to the repro-
ductive success of the flock, expending extra effort on protection against
predators, discovering good pasturage, and nurturing young lambs.32
The innovation could increase the success of the group, as measured by
the growth of the communal and the individual stock; it could generate
ever-more successful trade with other bands—as well as producing pro-
nounced inequalities within the pastoral society. If asymmetries in com-
mand of resources translate into differences in power, greater weight
given to some voices in deliberations about rules, the apportioning of
resources between public and private ownership could be increasingly
tipped in the direction of the latter, until the communal stock was in-
significant in comparison with private ownership. In tandem with this
movement, the idea of private control of resources could easily be ex-
tended into other spheres—most notably when domesticated animals
are exchanged with the products made by members of other groups.
A word of caution is appropriate here. The thought of private owner-
ship as a motivator to group productivity has been so overused in eco-
nomic discussions that it is essential to be specific. The “how possibly”
story does not view private property as an essential outgrowth of human
society, supposing that once we had engaged in normative guidance, di-
vision of labor, trade, and the domestication of animals it was inevitable.
Instead, under particular conditions, against a scheme of socialization
31. This is a nontrivial supposition, for one can imagine groups in which practices of
socialization rendered it quite repugnant.
32. My account echoes a famous Genesis story, the deal Jacob strikes with his father-in-
law, Laban. Similarly, the earlier discussion of saving surplus in good years, in preparation
for hard times, recalls the policies Joseph institutes in Egypt. Might we consider the myths
of ancient religions as embodying records of transitions in human prehistory retrospec-
tively seen as important forms of social and ethical advance?
producing certain kinds of desires—desires for the control of resources

previously communally owned—an institution of private property could
succeed in cultural competition.33
§21. Altruism Expanded

Previous sections have attempted to show how stepwise evolution of the
ethical project might transform the social environment and thereby
make new kinds of desires, aspirations, and emotions possible. The final
task is to show how these transitions can expand the scope and the char-
acter of altruism. Originally, normative guidance was seen as generating
behavioral altruism in response to altruism failures. As the ethical proj-
ect evolves, it can generate psychological altruism, even in more elaborate
forms than hitherto considered.
Altruism failures can be remedied by harnessing a number of emo-
tions: fear, dread of the unseen enforcer, awe and reverence, a positive
desire to be in harmony with the deity’s plans and wishes, even a sense
of identity with the society blessed with divine favor. The same ends can
be achieved by inspiring people not simply to simulate altruism but to
have altruistic propensities across a wider set of contexts. Well-socialized
people then act to help others through a mixture of motives—through
taking others’ wishes seriously, through sympathetic emotions, through
respect for the supposed source of the ethical code, through a sense of
identity with a group, through worries about the results of breaking the
rules. No special sort of psychological process is likely to be better at pro-
ducing appropriate behavior across all circumstances; the mind of “the
friend of humanity” may cloud over, but, equally, his or her reason may
go astray (§11). Reliability is an entirely appropriate measure, for, from
the perspective of achieving cultural success, the goal is to arrive at strat-
egies of socialization for eliciting preferred behavior on as many occa-
sions as possible. Pluralism has evident advantages. The group that
supplies a variety of psychological dispositions for altruistic response
obtains greater relief from altruism failures.
33. Much later (in §62) we shall take up this institution with more critical eyes.
Cultural success exerts pressure to develop schemes of socialization

extending the scope of psychological altruism. That can result from
effective techniques of promoting behavioral altruism. Change behavior
patterns so people engage in a larger number of friction-free interactions
with one another, and extended psychological altruism may follow. Par-
ticipating in cooperative ventures with B inclines A to think of B’s wishes
as good to satisfy and engenders feelings of warmth toward B. Skillful
socialization reinforces the effects. Parents acquire an arsenal of techniques
to induce rivalrous siblings to get along—and their methods are not all
recent inventions.
How far successful projects in socially embedded normative guidance
extend genuine psychological altruism, rather than replacing some pre-
vious altruism failures with behavioral altruism, is unclear—as unclear
as the categorization of contemporary people, in many everyday con-
texts of helping and in the special circumstances of experiments in shar-
ing, into real psychological altruists and others (§11). Under normative
guidance, psychological altruism is also extended in other ways, and the
rest of this section is concerned with modifications of the notion that in-
troduce further complications to the account previously offered (§§3–5).
I begin from an obvious point, one that recognizes familiar types of
altruism not involving any positive response to the actual desires of
the beneficiary (or to the desires that would actually be attributed to
the beneficiary). The altruistic mother does not align her wants with the
wishes of the young child who vigorously resists the medicine. Yet the
mother is surely responding to some sort of attributed wish: it is as
though she envisages the future life of her child, recognizing wishes that
would arise later, given various sorts of response now. In parallel fashion
to the account of §3, we can approach altruism in terms of responses to
the perceived interests of others.
How should the distinction between interests and mere wishes be
drawn? Many thinkers are tempted to identify interests with the wants
those others would have if they were clearly (and coolly) to deliberate
on the basis of all the facts, but this approach threatens to collapse into
triviality. (One of the facts is about what we would want if we knew every-
thing.) Yet there is an insight here: we separate the wish someone ex-
presses from his or her genuine interest by attributing to the person
some type of ignorance. A current wish diverges from a real interest

when the wish would give way to the preference marked by the interest,
were the person relieved of some current misconception or form of igno-
rance, and when the modified preference would be retained in light of
further knowledge.34
In responding to young children, accommodating wishes rather than
interests is often a defective form of altruism—perhaps not even worthy
of the name. Does this require the account of altruism to be rewritten to
focus on attributed interests rather than attributed wishes, so paternal-
ism would be preferred across the board? Reflecting on our ordinary
notion of psychological altruism, framed as it is by the ethical project,
you might say this: to be an altruist is to identify with the other person,
and that is to take the person seriously as an agent (at least once he or she
is mature); hence, even if you think the person’s wishes misguided, as
unlikely to promote what he or she would want were he or she to know
more, those actual wishes are to be respected. Or you might say some-
thing different: to be an altruist is to care about the other’s good, and that
is not what the person actually—and myopically—wants, but rather what
he or she would want were he or she better situated to judge; so one
should align one’s desires with the person’s interests, the person’s
wants as they would be if he or she were aware of crucial facts.
Once the ethical project has introduced the ideas of “identifying with
others” and of “the other’s good,” both thoughts are available: there are
paternalistic and nonpaternalistic forms of altruism. Ethical consider-
ations now figure in decisions about psychological altruism. On some oc-
casions, it would be arrogant to substitute one’s own judgment about what
the intended beneficiary would want, given the benefit of an idealized
34. A fully rigorous account would need further qualifications, since it is possible for
someone to acquire misleading information that subverts the modified wish. It would do
so, of course, by coming with another type of misconception or ignorance, from which yet
further knowledge could relieve the person. Perhaps the best way to approach the notion of
interests is to start with the idea of a remedy for ignorance, conceived as the clearing up of
misconceptions or a new piece of knowledge. An interest is a wish one would have, given a
remedy for current ignorance, and a wish that would survive any further acquisition of
knowledge provided the acquisition was supplemented with an appropriate remedy for ig-
norance. Whether this disposes of all the difficulties is not obvious. In any event, for our
purposes, the simpler approach of the text will do.
perspective. If A has evidence that would support the judgment that B

has thought hard about his or her valuations of outcomes, if A’s own re-
flections on those outcomes are hasty and uncritical, A is quite wrong to
override B’s expressed wants, even though, by chance, A’s particular
judgment on this occasion would be closer to what B, given more infor-
mation and cool reflection, would actually desire. By the same token, if A
has excellent evidence that B is missing a crucial item of information, if
there is no opportunity to present the salient facts to B—and thus induce
a change in B’s desires with which A’s own valuations could then be
aligned—then responding to B’s actual wants would seem to rest either
on indifference to B’s welfare or on disrespect for B’s powers of ratio-
nal revision.
There is a preethical notion of psychological altruism, out of which
the ethical project grows, but also ethically charged notions of psycho-
logical altruism emerging later. The latter can revise previous judgments
about “altruistic” responses. Before the introduction of the first agreed-
on rules, an agent may have been inclined to respond positively to the
desire of a fellow in a particular context: B*, a member of the band who
does not often associate with A and B, has found a coveted resource and
offers to share with B on an equal basis; A perceives B as wanting more
than an equal share and acts to help B grab the whole.35 Once a socially
embedded system of normative guidance has included the command that
those who volunteer to share what they have found should not be inter-
fered with (in the way B intends), the status of A’s intervention is changed.
A no longer endorses B’s desire: A has agreed to a rule that distinguishes
B’s wish from the desires A wishes to be satisfied. The first modification
of the concept of psychological altruism consists in moving from a no-
tion that treats any desire of the beneficiary as an occasion for positive
response, to one that restricts psychological altruism to those desires
that accord with the ethical code. Initially, because the original rules are
35. Note that the imagined scenario does not involve any direct benefit for A; for A there
are only costs (use of time and energy, risk of harm from the resentful B*). In terms of the
approach of §9, we should take A and B to belong to a subcoalition that does not include B*;
B* is a more distant member of the band, in that the first subcoalition that includes A and
B* is bigger and more inclusive than the first subcoalition that includes A and B (and simi-
larly for B and B*).
to remedy altruism failures, desires ruled out by this restriction will be

those embodying a failure of altruistic response: you do not count as an
altruist if you respond positively to the desire of someone whose wish
represents an altruism failure.36
Once ethical practice includes self-regarding commands (§20), some
desires are viewed as defective—and thus not to be promoted by others—
because they prevent a person from developing in a particular way. So
arises the vision of mature, well-socialized people, the people who en-
gage in the deliberations about the commands of the code, whose “real”
desires are to be identified by the things they would express in such de-
liberations. The deviant wishes expressed on different occasions can
even be conceived as altruism failures, where the agent and the potential
beneficiary are one and the same person. To want to behave in this way
now is to fail to respond to the wishes of the person you wanted to become.
Paternalism enters the picture, as altruists respond not to the actual
wishes of those buffeted by fluctuating forces but to the wishes endorsed
by people in “the cool hour,” wishes in accordance with the ethical
code.
Part of altruism consists in advancing the (endorsed) altruistic wishes
of others: A can be an altruist in virtue of wanting to advance B’s altruis-
tic wish to help B*. One of those toward whom B has altruistic desires
can, of course, be A, and this forms the basis for a distinctive sort of al-
truism, higher-order altruism, as I shall call it (although, if the term had
not already been preempted, “reciprocal altruism” might be better).
Sometimes it is altruistic to allow others to express their altruism to-
ward you, even though your own solitary wishes are thereby satisfied.
You have a long-term history of psychologically altruistic interaction
with someone else. Often you wish to do something together, although
each of you has different ideas about what this should be. Were both of
36. Here I am indebted to Jennifer Whiting, whose perceptive comments on an earlier

discussion of psychological altruism brought home to me the importance of this kind of
restriction. As Whiting noted, the “infection” of what superficially look like altruistic re-
sponses can proceed along a chain of indefi nite length (A promotes B1’s wish to advance
B2’s wish to . . . to promote B n’s wish to do something that would be counteraltruistic to-
ward B n+1), making the provision of sharp conditions on psychological altruism extremely
messy.
you to act selfishly, you would perform the individually preferred ac-
tions, but you would forfeit what is primary for each of you—to wit, be-
ing together. Were both of you to act as psychological altruists, as so far
construed, the situation would be even worse: each would do the less-
preferred action and still not have the benefit of acting together. To escape
the bind, one of you has to be a different type of psychological altruist,
an altruist who adjusts wishes to align them with the other’s altruistic
desire.37
Although it is evident that the concept of higher-order altruism can be
abused, providing cover for people to pursue their selfish wants, the
anti-Machiavellian condition (§3) discriminates cases. Egoists simply have
their solitary wants, or see the simulation of an altruistic response to
others as a good strategy for achieving those wants (“Of course, if you
really want to help me by doing that, I don’t want to stand in your way”).
Psychological altruists reflect on their partners’ wishes, factor in their
own desires to promote those wishes, and, if they accept the altruism of
a partner, do so because they view it as based on a wish more central to
the partner’s life than any they would express by promoting the part-
ner’s nonaltruistic desires.38
Often-repeated interactions among people, in which altruistic re-
sponses are expressed by both parties, bring with them the possibility of
an importantly different form of higher-order altruism, one in which the
processes through which outcomes are reached become sources of happi-
ness for the participants. The original solitary value ascribed to an out-
come is sometimes negligible in comparison to the value that whatever
outcome is reached results from a serious process of mutual engagement
with the wishes of another person. Adjusting our actions to one another
can be more important for us than what those actions actually achieve.
37. For a more developed account, see Kitcher, “Varieties of Altruism.”

38. I strongly doubt that the considerations in play are ever very precise— one simply
has a feeling that expressing altruism in a particu lar way really matters to a friend, a
spouse, a parent, or a child. Because those judgments are not exact, considerations of turn
taking often play a role. You allow your friend to be generous because you were able to ex-
press your friendly feelings on the last similar occasion. This is one place in which recipro-
cation does play a role in genuine altruism—simply because what is reciprocated is a genu-
inely altruistic response.
Our experiments of living began when a primitive system of rules was

used to make up for salient altruism failures. The desires targeted by
those first efforts were quite basic. The elaboration of normative guid-
ance generates new desires, eventually desires for interactions revealing
the mutual expression of altruistic responses. Successful extension of al-
truism can produce series of occasions on which people promote one
another’s wishes, taking pleasure both in the pleasure given and re-
ceived and in social approval of that pleasure. Institutions like marriage
(and perhaps other forms of partnership) give rise to such series, and
those who participate come to view efforts at accommodation, even when
not completely successful, as valuable. In turn, the recognition of value in
mutual response can reshape the institution of marriage or the most im-
portant kinds of friendship. From individuals for whom normative guid-
ance is a way of ameliorating social trouble, a sequence of experiments of
living can produce people for whom mutual recognition in an enduring
relationship is central among their desires.
The expansion of human desires was surely coupled to the refinement
of our emotional lives. Through the evolution of the ethical project, even
if our affective responses remained unaltered, new types of cognitions
and desires became attached to the affective states. Positive affective re-
sponses might be triggered, even amplified, by the recognition of the
ways of attuning our desires to those of others and theirs to ours. We can
dimly apprehend the origins of love.
All this changed our ancestors’ conceptions of what it is to live well.
To be secure, to be healthy, to eat, and to copulate is no longer enough.
My “how possibly” story ends with a vastly enriched notion of the good
life. Desires to develop one’s talents become central, the active contribu-
tion to our community is important to us, and particular relationships are
more significant than anything else. By gradual steps, the ethical project
could evolve, from the simple beginnings of socially embedded normative
guidance to the ethical sensibilities we discern in ancient Greece. Plato
is a footnote to the history of ethical practice.
chapter 4
One Thing after Another?
§22. Mere Change?

Pragmatic naturalism aims to understand the character of the ethical
project by exposing major features of its evolution. Probing the deeper
past is difficult, for clues are fragmentary. The invention of writing,
however, enhances the opportunities to investigate the evolution of
ethics: the records of the past five thousand years might reveal how con-
temporary societies have come to their present practices. More specifi-
cally, historical investigation promises to address challenging questions,
issues of immediate concern.
Is the evolution of ethics a matter of mere change? Is it analogous to a
Darwinian picture of the history of life, revealing only local adaptations
without any overall upward trend? Do ethical codes diff use and meta-
morphose through processes having no connection with truth or knowl-
edge or progress? Is it just one damn thing after another?
These worries express in a temporal context concerns much bruited
with respect to cultural variation. As anthropologists documented the
diversity of cultural practices (often framed only with difficulty by using
the concepts of Western ethical thought), and as they argued for under-
standing these practices on their own terms, rather than dismissing
138
One Thing after Another? 139
them as primitive, ethical relativism began to be taken seriously.1 The

core relativist idea denies any standard, or measure, independent of the
ethical practices of different societies, against which the code of one so-
ciety can be judged as superior to the code of another. The idea provokes
an obvious reaction. Consider groups of people you view as having done
horrible things. Familiar examples: the Nazi attempt to purge Europe
(and potentially the world) of “vermin” or the killing fields of the Khmer
Rouge. Many people feel a powerful urge to protest the behavior and
whatever ethical prescriptions are brought forward in its defense, to say
there is something objectively wrong about what was done, to deny that
condemnation only expresses a local perspective, to protest that those
condemned cannot, with equal justice, criticize their critics. There must
be some external standard to which ethics is answerable.
Exploring ethical variation across time avoids some of the tangles fig-
uring in cross-cultural debates about relativism. Historical study prom-
ises examples of societies, not merely distantly related, but actively en-
gaged with rival options for ethical transition. It might show the “mere
change view” to be correct or, by disclosing how people make “objec-
tive” decisions, provide clues about the constraints ethical deliberators
sense. As we shall learn, the task is harder than it might initially seem,
but two useful conclusions emerge. First, there are compelling examples
of transitions that look progressive—the mere-change view is hard to
sustain. Second, to the extent that decisions made by the pioneers
who first took progressive steps can be scrutinized, they are not readily
viewed as responses to external constraints. These points generate the
predicament Part II addresses.
The crucial episodes are those in which a society makes an ethical
innovation that appears not simply to articulate ideas already in place,
and that also seems to represent ethical progress. Initially, I shall look
1. Characterizing what relativism actually claims, and assessing its credentials, turns
out to be a complex matter. For perceptive discussions of the cluster of problems here, see
Gilbert Harman, The Nature of Morality (New York, Oxford University Press, 1977); Mi-
chele Moody-Adams, Fieldwork in Familiar Places (Cambridge, MA: Harvard University
Press, 1997); and Carol Rovane, “Relativism Requires Alternatives, Not Disagreement or
Relative Truth” Blackwell Companion to Relativism, ed. S. Hales (Blackwell, 2011);
Rovane’s views about the claims of relativism are developed further in a forthcoming book.
briefly at three examples challenging the mere-change view, but not pro-
viding insight into the processes underlying the apparently progressive
transitions: the transformation of the lex talionis in the ancient world,
the change from a heroic ethos in ancient Greece to the ideal of the citizen
of the polis, the emphasis on compassion introduced by Christianity.
For clearer ideas about how the participants made their decisions, we
shall need more recent cases.
§23. Three Ancient Examples

In the earliest legal codes, the idea of exact retribution—eye for eye,
tooth for tooth, life for life—is construed in an oddly literal (and, by our
lights, repugnant) fashion. If someone causes the death of “the daughter
of a senior,” “that man’s daughter” is to be put to death.2 Although the
surviving references more often concern daughters, the law does not ap-
pear to rest upon the invisibility of women as independent people—there
are similar formulations about sons.3 Analogous laws sometimes do em-
body conceptions of women as property, whose lives and bodies are
controlled by male relatives: a law on rape declares that the wife of the
rapist is to be raped by whomsoever the father of the victim chooses.4
A few centuries later, this literal construal of exact retribution has
vanished. Now it is the perpetrator of the deed who must pay in the man-
ner, and to the extent, of the damage inflicted: his or her life must be ex-
acted to pay for the life of the victim.5 A transition in ethical practice (not
only in law) has occurred: where it was previously supposed that, when
harm has been done to a member of one family, it is right to inflict the
same injury on the corresponding relative in the family of the perpetra-
tor, it is now the doer of the deed who should suffer. More than two mil-
2. James Pritchard, ed., Ancient Near-Eastern Texts (Princeton, NJ: Princeton Univer-
sity Press, 1950), 170, 175.
3. Ibid., 176.
4. Ibid., 185.
5. Of course, not all relatively ancient societies maintained the lex talionis. Nordic and
Saxon groups developed the notion of “wergeld,” a monetary payment compensating for
lives taken. In the ancient Near East, however, the idea that murderers must forfeit their
lives remained central. Witness the Hebrew Bible.
lennia on, we may demur at the thought that this is the final word on the
matter, but it is hard to resist seeing the change as an improvement. We
envisage cases (the overwhelming majority?) in which the correspond-
ing relative knew nothing of the crime, cases where he or she was a child
or even a friend who mourns the victim. However that may be, if the rela-
tive was not involved in the killing, justice miscarries if the relative loses
his or her own life while the murderous relation goes free. Even if the
perpetrator is punished “through him or her,” that fails to support the
practice, for the relative cannot be treated as part of the machinery of
punishment, as if his or her life were not important to him or her as well
as to the perpetrator. When societies go after the criminal directly,
how can it not be a progressive step?
Great myths and poetry of early civilizations celebrate figures whose
recorded deeds express their devotion to an ideal of honor and greatness
overriding considerations that move later ethicists. Prominent examples
are Homeric heroes.6 We do not need to know if the Iliad has a historical
basis; the crucial question is whether the ethical attitudes expressed are
those prevalent in some Homeric past. One basis for supposing they are
is the improbability of oral presentations of a clearly defined ethical per-
spective, popular across many generations, if the ethical ideas failed to
reflect the actual outlook of the audience (or an audience hearers could
identify as part of their history).
One shift in the period between Homer and Solon replaced the em-
phasis on personal honor as the principal ethical end with the idea of
a contribution to the common good.7 The Homeric hero’s wartime life
was directed toward acquiring personal glory; his prowess might be
embodied in trophies (often given away in acts that simultaneously
marked the hero’s generosity and his previous exploits).8 To appreciate
6. Others include the noble warriors from German, Norse, and Japa nese traditions.
7. This is by no means the end of the idea of honor in Western ethical traditions. The
concept recurs again and again, in chivalric codes in the Middle Ages, in illuminating pas-
sages in Shakespearean plays, in standards for eighteenth-century gentlemen and ladies, in
the military ideals of affluent nations in the nineteenth and early twentieth centuries.
8. For clear presentations of the central features of the heroic code, see Walter Donlan,
The Aristocratic Ideal in Ancient Greece (Lawrence, KS: Coronado Press, 1980), chap. 1;
Moses Finley, World of Odysseus (Harmondsworth, UK: Penguin, 1980); chap. 5; and Joseph
the transition, juxtapose a passage from the Iliad with Thucydides’ later
“account” of Pericles’ funeral oration.9 Hector responds to various pleas
not to engage Achilles in single combat by affirming the demands of
honor. He knows his death would spell disaster for his city (and his fam-
ily), but he cannot accept the dishonor resulting from refusing the chal-
lenge.10 By contrast, when we read Thucydides’ “Pericles,” the common
good comes first. “Pericles” says of the fallen:
Some of them, no doubt, had their faults; but what we ought to

remember is their gallant conduct against the enemy in defence of
their native land. They have blotted out evil with good, and done
more ser vice to the commonwealth than they ever did harm in their
private lives.11
These words are meant to honor a group, not an individual, and they do
so by highlighting individual devotion to the good of the group.
Between the time recorded in the Iliad and the events commemorated
by Thucydides, Greek warfare had changed profoundly. Military ac-
tions were now dominated by the orga nization of armed troops into
the phalanx. (Men bearing heavy armor and a large shield were arrayed
shoulder to shoulder and marched forward together, presenting long
spears.) Success in battle depended no longer on the strength, endurance,
and skill of an outstanding individual—an Achilles, a Hector—but on
disciplined maintenance of one’s place in the line. Conduct routine in
the Iliad—Achilles’ refusal to participate, Hector’s rejection of the
counsels of prudence, Diomedes’ private treaty with his guest-relative
Glaukus—now appears selfish, irresponsible, capricious, and quirky.
The predominance of honor gave way to the virtues of moderation, self-
discipline, and loyalty.
Bryant, Moral Codes and Social Structure in Ancient Greece (Albany: State University of
New York Press, 1996), chap. 2.
9. Thucydides clearly warns that he reconstructs speeches by combining the sense of
what was said with the thrust of what it would have been appropriate to say (Peloponnesian
War [Harmondsworth, UK: Penguin Classics, 1972], 47).
10. See Finley, World of Odysseus, 115–17, for an excellent discussion of this episode.
11. Thucydides, Peloponnesian War, 148.
Pace Nietzsche, the substitution of solidarity in pursuit of a common

project for individual ventures dominated by the thirst for honor is, at
least in some respects, a step forward. Sacrifices undertaken in pursuit
of honor often appear irresponsible, even absurd—Hector’s decision has
foreseeable consequences dreadful for him and for those about whom he
cares. The transition can be viewed as restoring a healthier type of nor-
mative guidance, one closer to the early stages of egalitarian deliberation
about the character and promotion of shared ends, even a correction of
spectacularly destructive altruism failures.12
Many people conceive Christianity as transforming the ethical frame-
work of the Greco-Roman world. What they probably intend is: some
features of the ethical attitudes of most social groups identifying them-
selves as inspired by Jesus were absent from most other groups living
under the aegis of Rome. What might these features be?
Obvious answer: the growth of Christian belief increased compassion
in the ancient world. Jesus enjoined his followers to forgive their enemies
and love their fellows. His influence reformed brutish Roman institu-
tions. According to an eminent Victorian:
No discussions, I conceive, can be more idle than whether slavery,

or the slaughter of prisoners in war, or gladiatorial shows, or poly-
gamy, are essentially wrong. They may be wrong now—they were
not so once—and when an ancient countenanced by his example one
or other of these, he was not committing a crime.13
There was no simple impact of Christianity on the ancient world,

even with respect to the conduct of war, the abolition of slavery, the
12. Here it is worth recalling Mill’s insight that attention to the consequences does not
rule out self-sacrifice but simply demands that the sacrifice be worth something (John Stu-
art Mill, Works, Vol. 10 (Toronto: University of Toronto Press, 1970), 217).
13. W. E. H. Lecky, History of European Morals from Augustus to Charlemagne (New
York: Braziller, 1955), 1:110. For Lecky’s advocacy of the idea of ethical progress, see also
vol. 1, 100–103, 147–50; vol. 2, 8–11, 73–75. The apparent relativism of Lecky’s formulation
is misleading rhetoric: he does not literally think that what is right (or wrong) has changed,
but that what is taken to be right (or wrong) is altered. This is plain from his confidence in
the Victorian values that have emerged.
character of public spectacles, or views about marriage. The decades

and centuries after Constantine were marked by frequent acts of violence
carried out by Christians and commended by their leaders (the savagery
of the Crusades is prefigured in quarrels about orthodoxy erupting from
the fourth century onward); nor did Christianity play any straight-
forward role in replacing chattel slavery with serfdom and villenage.14
Nor is the priority of love and forgiveness consistently upheld in all ca-
nonical Christian writings.15 The evangelists attribute to Jesus such dicta
as “I came not to bring peace but a sword” and also describe the disciple
commanded to forgive others a very large number of times (490), causing
the death of members of the movement who failed to contribute all their
goods.16
On Lecky’s account, Christianity introduced a progressive shift, cen-
tered on adopting a new ideal of altruism. Recall the dimensions of
altruism: intensity, range, scope, recognition of consequences, and em-
pathetic understanding (§5). Even without taking into account further
complexities (those noted in §21), the ideal can be formulated in several
ways. In the extreme version, one commending “golden-rule altruism”
with respect to every person and every context, and demanding complete
accuracy about people’s preferences,17 it would be quite impossible to
follow. Any community trying to adopt it would face difficulties when
individuals’ initial preferences for indivisible goods are incompatible
(and in widespread “Alphonse-Gaston” situations where both parties are
moved by the wish to abnegate their desires in favor of the other). Any
14. For an insightful discussion of issues about slavery, see Moses Finley, Ancient Slav-
ery and Modern Ideology (London: Chatto and Windus, 1980). Historians from Gibbon on
have noted the savagery of the confl icts among early Christian sects. A lucid account of the
Crusades is offered by Steven Runciman, A History of the Crusades, 3 vols. (New York:
Harper and Row, 1964–1967), see esp. his narrative of the massacres that attended the “tri-
umph of the cross” in Jerusalem, 1:286ff.).
15. As §18 noted, this theme is anticipated in Mesopotamian texts predating the Gospels
(see n. 20).
16. See Matthew 10:34–39 (also Luke 12:49–53), Matthew 18:21, and Acts 5:1–11.
17. I characterize the ideal in terms of preferences rather than interests because a Chris-
tian formulation in terms of interests would adopt a very special notion of the interests of
individuals, one that does not obviously translate into a judgment that the ideal would
mark a direction of ethical progress.
viable version requires a system of principles explaining when the needs

of one individual are more urgent than those of another, thus assigning
the roles of altruist and beneficiary. We can conceive the appealing parts
of Jesus’s message as articulating a part of this system, by identifying the
urgent needs of those often excluded from consideration—those beyond
the range of altruistic dispositions.
The popular thesis that Christianity represents an ethical advance is
most plausible when the movement is viewed as promoting altruistic
responses to marginalized people whose most basic desires have previ-
ously not been met. Victorian confidence about the ethical progress
made by Christianizing the Roman Empire supposes that sort of expan-
sion. It should allow for variation across regions and across periods in
the newly Christian world, for different advances made by groups whose
altruism was extended to different targets and to different extents.18 In-
stead of thinking of a defi nite Christian ideal, we would do better to
conceive of a general trend, for which Christianity provides a forceful
expression.
In none of these instances do we have any sources revealing how and
why people made the apparently progressive shifts. For them we need to
move closer to the present.
§24. Second-Sex Citizens

During the last two centuries, in the countries of Western Europe and
North America, there have been important changes in the civil status
of women and in women’s abilities to gain access to positions and privi-
leges, previously viewed as an exclusively masculine domain. This shift
has not proceeded at the same pace across all sectors of the societies in
question, nor has it eradicated earlier attitudes opposing women’s entry
into spheres from which they were previously blocked, nor has the move-
ment finally attained the goals for which many of those involved in it
18. The expansion of altruism in this way is hardly the exclusive province of Christian-
ity. As noted, the injunction to love and forgiveness, even toward enemies, appears in Baby-
lonian literature several centuries before Jesus, and the ancient world contained groups of
non-Christians (for example, Jews and Stoics) whose ethical codes extended the altruism
of prior traditions.
have striven.19 Nevertheless, things have changed: it no longer seems

appropriate for the leading English jurist to strip his daughter naked, to
tie her to a bedpost, and to whip her until she will agree to marry the psy-
chologically disturbed nobleman whom he has selected as her husband;
or for a widow to give up her children to her husband’s family, even
though, by the lights of the surrounding community, the mother, not her
in-laws, belongs to the orthodox church; or for women to be denied any
education; or, when education is grudgingly allowed, for them to be de-
barred from receiving degrees, despite the fact that one of their number
shows herself superior to her male contemporaries on the most presti-
gious mathematical exam of the day.20 The list of horror stories from the
past could be enormously extended. The plausibility of ethical progress
in this domain is signaled by the reactions of many citizens of contempo-
rary democracies, who not only firmly believe these practices were ut-
terly unjustified, but also cannot conceive how reflective people could
ever have permitted them.
The character of the advance is twofold. First, rules preventing women
from playing coveted roles in their societies, from having access to par-
ticular institutions, from possessing things men around them wanted to
acquire, and from exercising certain kinds of choice were rejected as
ethically wrong. Second, the presence of women in roles and institutions
traditionally held as male preserves has led to improvements in those
roles and institutions.21 The first type of change was consolidated earlier,
19. The writings of Catherine MacKinnon serve as important reminders of what a sig-
nificant number of women (as well as some men) think remains to be done; see her Femi-
nism Unmodified (Cambridge, MA: Harvard University Press, 1987) and Towards a Femi-
nist Theory of the State (Cambridge, MA: Harvard University Press, 1989).
20. See J. Morrill, ed., The Oxford Illustrated History of Tudor and Stuart Britain
(New York: Oxford University Press, 2001), 97, for the account of Sir Edward Coke’s coer-
cion of his daughter; George Walker, although a Quaker, was given custody of his children,
despite the fact that his ex-wife, Ann, was Anglican—the law of Virginia ranked patriarchy
ahead of orthodoxy (see Julia Cherry Spruill, Women’s Life and Work in the Southern Colo-
nies [New York: Norton, 1972], 345); in 1889, Charlotte Angas Scott, a student at Girton,
obtained the highest score in the Mathematics Tripos at Cambridge, although she was not
able officially to register for a degree.
21. Many people would view the impact of women on life in the professions from which
they were so long excluded as a good thing. Within academic discussions, for example, in-
and some view the second as more controversial. I take both as instances
of ethical progress.
How were the advances made? In the ancient examples it is impossi-
ble to identify psychological processes through which individuals, or
groups, made ethical discoveries. Here, however, there is material to
which investigators can turn in hopes of picking out the new perception
or new piece of reasoning that fueled ethical evolution. A sequence of texts,
retrospectively inspiring, leads from the writings of Mary Wollstonecraft
at the end of the eighteenth century, to the documents of the nineteenth-
century American feminist movement, to the classic essay co-authored
by John Stuart Mill and Harriet Taylor, to the fiction of Charlotte
Perkins Gilman and the social commentary of Virginia Woolf, to Sim-
one de Beauvoir, to Betty Friedan, Catherine MacKinnon, and their
successors.22
As with scientific revolutions, the triumph of a radically different per-
spective proves far more complex than might have been supposed.23
Once the revolution is over, the confident insistence on male privilege
seems monstrous in its blindness. How could Sir Edward Coke have tied
his daughter to that bedpost, or Sophie Jex-Blake’s father have hampered
clusion of women has sometimes fostered a more cooperative approach to research, and
established this as a rival model for the aggression of male-male competition.
22. Wollstonecraft, A Vindication of the Rights of Woman (New York: Modern Library,
2001); Alice Rossi, ed., The Feminist Papers (New York: Bantam, 1977); John Stuart Mill
(and Harriet Taylor, whom I include as coauthor here), On The Subjection of Women in Mill
On Liberty and Other Essays [Oxford: Oxford University Press (World’s Classics), 1998];
Charlotte Perkins Gilman, Herland (New York: Pantheon, 1979), and The Yellow Wallpa-
per (New York: Routledge, 2004); Virginia Woolf, A Room of One’s Own (New York: Har-
court, 1957); Simone de Beauvoir, The Second Sex (New York: Vintage, 1974); Betty
Friedan, The Feminine Mystique (New York: Norton, 1963); Germaine Greer, The Female
Eunuch (New York: Bantam, 1972); Catherine MacKinnon, Feminism Unmodified (Cam-
bridge, MA: Harvard University Press, 1987), and Towards a Feminist Theory of the State
(Cambridge, MA: Harvard University Press, 1989); Donna Haraway, Simians, Cyborgs
and Women (New York: Routledge, 1991); and bell hooks, Ain’t I a Woman (Boston: South
End Press, 1981). Although I mainly focus on Anglophone texts, there are many other im-
portant sources—for example, the response of Olympe de Gouges to the Declaration of the
Rights of Man and the Citizen, produced by the (male) leaders of the French Revolution.
23. The intricacies Kuhn discerned in major scientific debates (The Structure of Scien-
tific Revolutions [Chicago: University of Chicago Press, 1962]) are even more apparent in
the ethical case.
her energetic wish to engage in medical practice? The autocratic men in

these stories resemble the cartoon figures of old narratives in the history
of science, the simple-minded Aristotelians who do not understand
the brilliance of Galileo’s arguments, and who maybe even refuse to
look through his telescope. Investigating more closely, we find not one
sequence of texts, with a compelling set of insights demolishing unsup-
ported prejudice, but two—and each connects with particular parts of
past ethical practice.
The heart of the feminist insight is a factual claim: the social practices
prevalent in society (that is, the society in which the feminist author
writes) confine the desires of women to a narrower range than would be
achieved under different forms of socialization. Later, that claim is ex-
panded—“and the same goes for men, too.” The claim strikes directly at
the conservative case. In many societies, from the ancient world to the
present, women are assumed not to want certain kinds of possessions
and positions, supposed to be incapable of particu lar kinds of choices.
If they occasionally do, by some quirk, express a desire for the goods or
offices, or want to make the choices, these preferences diverge from their
interests. Conservatives see no sense in which society fails to respond to
the wishes—or to the proper wishes—of its female members. To the ex-
tent that the society is good at socializing young girls, the “deviant”
wishes will not arise with any serious frequency, and the theses about
lack of desire and lack of ability will rarely be challenged. Sometimes the
theses are buttressed further by assertions about the divine will.
Once entrenched, attitudes that certain types of female desires are
deviant, and thus not to be endorsed, are difficult to displace in societies
skilled at socializing the young. The rare girls and women who voice
“deviant” wishes can be dismissed as in need of correction; they rarely
have the chance to challenge common views about their incapacities.
Under such conditions, initially inexplicable attitudes no longer appear
monstrous—although the extreme case of whipping a young woman into
submission is hard to view as anything but pathology. Fathers who dis-
courage their daughters from public life are profoundly wrong about the
aspirations they check, but their society has not only drummed into them
the impropriety of those aspirations but also made it hard for them to
acquire evidence about what would happen if the wishes were fulfilled.
One way to counter paternalism is to show that hopes expressed oc-

casionally by a small number of women would be far more widespread if
society did not so efficiently smother them. How can that be done? Woll-
stonecraft’s Vindication of the Rights of Woman often seems timid to
contemporary readers, who misread its clever rhetorical strategy. Woll-
stonecraft argues for a limited goal—allow the education of women—
precisely because she can connect that goal with improved fulfillment of
roles traditionally assigned to women. Her conservative opponents,
whether they maintain that the tasks of bearing and raising children,
and supporting husbands, are divine provisions for women, or whether,
like Rousseau, her principal foil, they emphasize proper nurturing of the
(male) citizen, are committed to allowing women access to whatever
bests fit them for the roles of wife and mother. Hence, Wollstonecraft can
argue for her proposed change by showing the superior ability of edu-
cated women to discharge “their” roles. She highlights the point in a
passage, often embarrassing to contemporary readers: the fates of un-
educated and educated wives, and of their children, are contrasted; the
comparison culminates in a vision of the educated widow; having suc-
cessfully raised her family (“her work is done”), she ascends to rejoin her
husband in heaven. Because her opponents are committed to take the
conventional role seriously, the rhetorical effect is devastating. How
much does this gain? Surely the educated woman will be confined to the
domestic sphere?24 Opening the door to education, as Wollstonecraft
probably saw, weakens the power of traditional systems of socializing girls
and young women and thus increases the chance women will express
desires for broader roles in society. It is a crucial first step in normalizing
those desires.
On the Subjection of Women goes further, replacing the argument that
women’s education is needed to develop better wives and mothers with
an appeal to individual freedom. Education can be viewed, as it is by Mill
and Taylor, as a crucial device for men and women to formulate what is
24. Late in Vindication, in chap. 9, Wollstonecraft does venture a little further, advert-
ing to the possibility that women may do some kinds of work (“keeping a little shop,” par-
ticipating in medical care). I read this as a clever signal to sympathetic readers (the ones
who have come this far with her), to the effect that the changes for which she officially cam-
paigns are only beginnings.
central to their lives, to “find their own good in their own way.”25 They
call for social experimentation, both as a means for providing the young
with potential models from which they can assemble their individual
conceptions of how to live, and as the proper expression of what people
want, not to be confined unless it does harm to others.26 Later, when the
desires of educated women to participate in public life—and, in some
cases, to change the character of public life—have become even more
widespread, Woolf documents the ways in which those desires continue
to be resisted.27
Why does confinement continue? Woolf’s own quotations from the
oppressive men who rein in their daughters reveal the structure of con-
servative thought. Suppose the step recommended by Wollstonecraft is
taken: a society of educated women contains wives and mothers who
discharge their roles with unprecedented success. Not only would broad-
ening the activities of educated women require new arguments, but, ac-
cording to the case already made for educating women, it would likely be
counterproductive. If the emphasis is firmly on improving the conduct
of wives and mothers, pitching women into the public world appears a
bad idea, one likely to produce weakened marriages and neglected chil-
dren. Conservatives protest the Millian insistence on the primary value of
individual development, on formulating freely one’s own plan of life—by
appealing to higher, divinely prescribed, goals for human existence, by
emphasizing the health and flourishing of individuals or of society—but
they can even adapt their reasoning to the framework their opponents take
for granted. Desires are to be honored only if they do no harm to others.
If women are given access to public life, they will do harm—their hus-
bands and children will suffer. Women’s desires for time-consuming ca-
reers, for prominent positions in society, are viewed as altruism failures.
25. See John Stuart Mill, On Liberty[Oxford: Oxford University Press (World’s Clas-
sics), 1998], chap. 1.
26. John Stuart Mill, On the Subjection; plainly the essay reinforces and is reinforced by
the central ideas of On Liberty.
27. Virginia Woolf, A Room of One’s Own (New York: Harcourt, 1957), and especially
Three Guineas (San Diego: Harcourt Brace Jovanovich, 1966). A crucial move in the latter
is the suggestion that public life be transformed by the pressure of women who form a “so-
ciety of outsiders.”
Married women without children engage less frequently in such failures,

but female pursuit of a career can be fully exempt from them only if the
woman is free of family commitments. Since the desire for a family is
central for women, the Mill-Taylor recipe will produce wrongdoing or
unhappiness.28
So begins a (new) round of catch-22. From the reformers’ perspective,
once the public aspirations of women are no longer viewed as pathologi-
cal, the lack of response to them constitutes a class of altruism failures;
the counterpart desires of husbands and fathers are endorsed at the cost
of suppressing women’s wishes. Principles of fairness, shared by conser-
vatives and reformers, oppose constant sacrifices by one spouse to ben-
efit the other. Conservatives see asymmetries here—women have special
talents and abilities expressed in nurturing the family; their happiness is
centrally bound up with the family’s flourishing. The obvious counter:
this is a product of existing conditions of socialization, of the particular
way the institution of marriage has been framed, and things could be
done differently. We should experiment. But, conservatives insist, ex-
periments are properly canceled if they risk great damage. Framing the
roles of husbands and wives differently would, given “the natural desires
of the sexes,” damage a valued institution (marriage) and cause frus-
tration and unhappiness for parents and children alike. An obvious
charge: conservatives beg the question. The charge is met with a coun-
tercharge: tu quoque. To assume the experiments will not rub against
human nature is already to presuppose the desires of men and women to
be adaptable, and that is equally to beg the question. Each side must ar-
gue from its preferred ideas about the plasticity of human preferences.29
28. Arguments of this kind survive into the present. Their persistence is a product of
the continued inability to solve broader problems about the provision of public goods and
the education of the young. Resolution is hard because the issues are so entangled. See
Chapter 10 (§60).
29. The line of reasoning attributed to conservatives here is still present in many societ-
ies, and in many corners of societies that officially endorse women’s entitlement to seek
demanding careers. Because the problem of combining work and family life remains un-
solved, for a large number of women and for a smaller number of men, issues about refram-
ing the institution of marriage remain. These combine with other questions about the
forms of the division of labor in contemporary societies, about the pressures that division
of labor exerts, about the distribution of resources, and about the provision of public
Brief rehearsal of this dialectic exposes important aspects of the ethi-

cal transition. The Mill-Taylor point, that women’s capacities are un-
known because potentially revealing experiments in socialization have
never been tried, is indisputable; the conclusion that those experiments
should be tried is the locus of controversy, yielding the subsequent im-
passe. How then was this revolution actually resolved?
Through demonstrations and demonstration. The protests of suf-
frage movements, including the willingness of women to sacrifice their
lives to reveal how strongly they wanted full participation in society,
made their aspirations impossible to ignore. Female labor during the
Great War demonstrated women’s capacities, and also the possibility of
combining work with the nurture of children. It is no accident that the
United States and the United Kingdom both granted rights to vote after
the end of the war. Later contributions of women in the Second World
War, and their withdrawal into a more traditional domesticity in the im-
mediate postwar period, provided the background against which the
women’s movements of the 1960s could uncover suppressed and unsatis-
fied yearnings. The centrality of consciousness raising to feminism re-
calls the advance already made by Wollstonecraft: the first step is to re-
veal that certain wishes are widespread, not therefore easily dismissed as
pathological, and that they remain unexpressed because of the smother-
ing effects of social expectations. The women of the 1960s who attended
group meetings, sharing aspirations and experiences, could look both to
the demonstrated capacities of their predecessors at times of national
need, and to the partial fulfillment of desires they recognized in them-
selves. Their voices could not be ignored. The impasse was broken.
Ethical progress was made.
goods. The resultant entanglements make debates among conservatives and reformers
even harder to resolve. Where women’s lives are able to combine demanding work and fam-
ily most smoothly, this is frequently achieved at the cost of deferring the burden to other
women whose choices are far more restricted (women of lower socioeconomic status who
serve as caregivers or housekeepers). Wollstonecraft’s assumption of the presence of ser-
vants in the domestic arrangements she envisages is not quite the anachronism it initially
appears—nor is Woolf’s reliance on the idea that someone else will do the shopping and the
cooking. (I am indebted to Martha Howell for presentations on Wollstonecraft that have
helped me to see these aspects of her problem more clearly.)
What was discovered? Factual knowledge advanced: people learned

that, under different conditions of socialization, women wanted things
traditionally denied to them; that they found satisfaction in attaining
some of these things; that fulfillment of the wishes did not thwart desires
previously seen as central to female nature—public life combined more
or less satisfactorily with family life.30 Increased factual knowledge pro-
liferated desires for access to public life, fostering acceptance of the de-
sires as prevalent and no longer pathological. Recognition of the suppres-
sion or frustration of those desires aroused sympathy, recruiting male as
well as female allies for the reform movement. Like the early elaboration
of normative guidance, in which particular altruism failures cause too
much trouble, the increase of sexual egalitarianism occurred partly be-
cause, in the end, traditionalists wanted a quieter life.
§25. Repudiating Chattel Slavery

If there is one example in which the attribution of progress is almost in-
controvertible, it is the abolition of the “peculiar” institution, chattel
slavery.31 Opposition to slavery intensified in Britain in the late eigh-
teenth century and in America in the first half of the nineteenth, culmi-
nating in the Emancipation Proclamation and the Civil War. Ultimately
the view that slavery is ethically permissible was replaced by the denial
of that claim. How was that advance accomplished?
Seventeenth- and eighteenth-century northern Europeans, whether
resident in the ancestral countries or dispersed among the colonies,
attempted to embed slavery within their ethical codes. Apologists drew
concepts and distinctions formulated centuries earlier in efforts to jus-
tify ownership of human beings. One traditional defense distinguished
between people who are permissibly enslaved and those who are
30. The “more or less” enters here as a reminder of the difficulties already remarked (n.
29), of the fact that special social circumstances are needed for the combination to work
smoothly. See §60.
31. I think it can be opposed only by arguing that the notion of ethical progress cannot
be given a clear sense. I am grateful to Edie Jeff rey for reinforcing my conviction that this
example is indispensable for any account of ethical progress, and for giving me good advice
about how to investigate it.
not—formulated in the Hebrew Bible by differentiating the chosen

people from their captives in war, by Aristotle in terms of variants in
individual nature, and by the medieval church through separating the
faithful from infidels. Colonial Christians added further support from
the scriptures. The Pentateuch testifies that the patriarchs had slaves
(and had sexual relations with female slaves); the letter to Philemon
endorses slavery; further, the particular people enslaved in North Amer-
ica (people of African descent) are descendants of Ham (or, in some ver-
sions, Canaan), inheritors of a biblical curse. Protestant Christianity
also contrasted the liberty of the soul to attain to God’s grace with mere
bodily liberty. On this basis, some claimed, slave traders were doing
their captives a favor. 32
The slave-owning cause constructed its own account of the trade.
The slaves’ native situation in Africa was portrayed as a state of Hobbes-
ian nature, dominated by strife, bestial practices, and utter ignorance.
After transporting the unfortunate people across the Atlantic, kindly
slave owners provided food and shelter (as well as paternal affection) in
exchange for toil. Even more important, slaves were given the opportu-
nity to hear the true religion and gain spiritual salvation.
All this is rubbish, but it is impossible to understand Christian accep-
tance of slavery without recognizing the self-serving interpretation. In
1700, Samuel Sewall published a pamphlet proposing an analogy be-
tween the slavery of the colonies and the (unlawful) servitude of Joseph.
His tract drew a response from John Saffi n, who, in 1701, addressed
Sewall’s suggestion that “we may not do evil that good may come of it”
by writing, “It is no Evil thing to bring them out of their own Heathenish
Country, where they may have the Knowledge of the True God, be Con-
verted and Eternally saved.”33 Five years later, Cotton Mather saw Afri-
can slaves as providing religious opportunities for colonists:
32. In The Problem of Slavery in Western Culture (Ithaca, NY: Cornell University Press,
1966), David Brion Davis provides an illuminating account of all these apologist strategies
and their relationship to ancient and medieval thought.
33. John Saffi n, “A Brief Candid Answer to a Late Printed Sheet Entitled The Selling of
Joseph” (1701), in Against Slavery: An Abolitionist Reader, ed., M. Lowance (New York:
Penguin, 2000), 16.
The State of your Negroes in the World, must be low, and mean, and
abject; a State of servitude. No Great Things in this World, can be
done for them. Something, then, let there be done, towards their
welfare in the World to Come. . . . Every one of us shall give account
of himself to God 34
In the middle eighteenth century, the influence of the argument dimin-

ished because the efforts to propagate the Gospel were so obviously
unsuccessful. Slaves preferred to spend their Sundays dancing, trading,
and resting—and, as David Brion Davis notes, they did not flock to a re-
ligion “which sanctioned their masters’ authority, which enjoined them
to avoid idleness and to toil more diligently, and which promised to deprive
them of their few pleasures and liberties.”35 Colonists concluded, how-
ever, that slaves were incorrigible. The collapse of one line of proslavery
argument buttressed another.
Saffin’s response to Sewall already claimed that Africans and Europe-
ans were distinguished in moral and intellectual temperament: his tract
closes with a piece of doggerel attributing innate vices (cowardice, cru-
elty, libidinousness, etc.) to the black races.36 The judgment survived
into the nineteenth century. As late as 1852, Mary Eastman could write a
response to Uncle Tom’s Cabin in which she assembled all the main parts
of the “difference” argument: Africans are descendants of Ham, cursed
by God, with traits of character requiring firm discipline by wiser (and
benevolent) people of European ancestry; slaves are no more appropri-
ate bearers of freedom and self-government than wayward children. 37
Ideas like these were current, not only among literalist Christians but
also in Enlightenment circles. Although Montesquieu, the most insight-
ful early critic of slavery, punctured the appeal to innate differences,38
34. Cotton Mather, “The Negro Christianized: An Essay to Excite and Assist That Good
Work, Instruction of Negro-Servants in Christianity,” in Lowance, Against Slavery, 19.
35. Davis, Problem of Slavery, 218; see also 211–22.
36. Lowance, Against Slavery, 17.
37. See Mary Eastman, Aunt Phillis’ Cabin: Or Southern Life As It Is, excerpted in Low-
ance, Against Slavery, 296–300.
38. See Charles Montesquieu, Spirit of the Laws (Cambridge, UK: Cambridge Uni-
versity Press, 1989), for Montesquieu’s famous remark that Africans cannot be humans
eighteenth-century speculative anthropology inspired Voltaire, Hume,

Buffon, and their intellectual descendants to advocate African inferior-
ity. Adam Smith was a rare dissenter, but he was soundly rebuked by the
Virginian Arthur Lee, who drew on his extensive experience of black
slaves to set Smith straight.39
General considerations about racial hierarchy were coupled with
claims about the behavior of Africans, both in their native countries and
in their state of servitude. A body of literary attempts to depict the nobil-
ity of enslaved Africans (Aphra Behn’s Orinooko is a prominent repre-
sentative) was countered by a far larger volume of writings from the
people (slave owners) who claimed to know the subject best. Achieve-
ments of individual slaves were systematically undervalued. Expressions
of a conviction that black people are doomed to lesser accomplishments
(also unpleasing, if not disgusting) are even found in the words of two of
America’s most high-minded presidents. Jefferson wrote:
Whether the black of the negro resides in the reticular membrane

between the skin and the scarf-skin, or in the scarf-skin itself;
whether it proceeds from the colour of the blood, the colour of the
bile, or from that of some other secretion, the difference is fi xed in
nature, and is as real as if its seat and cause were better known to us.
And is this difference of no importance? Is it not the foundation of a
greater or less share of beauty in the two races? Are not the fine mix-
tures of red and white, the expressions of every passion by greater or
lesser suff usions of colour in the one, preferable to that eternal
monotony, which reigns in the countenance, that immoveable veil of
black which covers all the emotions of the other race? Add to these,
flowing hair, a more elegant symmetry of form, their own judgment
because if they were it would follow that we (Europeans) are not Christians. His ironies
were unappreciated. For some uncomprehending reactions to Montesquieu, see Davis,
Problem of Slavery, 403. It is also worth noting that one of Montesquieu’s most important
arguments against slavery imagines that the roles of slaves and masters are determined by
lot; for what seems to be an anticipation of Rawlsian appeals to ignorance of social position,
see the addenda to Spirit of the Laws.
39. Arthur Lee, An Essay in Vindication of the Continental Colonies in America, from a
Censure of Mr. Adam Smith in his “Theory of Moral Sentiments” (“Printed for the author”
London, 1764).
in favour of the whites, declared by their preference of them as uni-

formly as is the preference of the Oranootan for the black women
over those of his own species.40
Jefferson goes on to assess the character and accomplishments of the

slaves he knew: their passions are transient and instinctual, they have
scant power of reason, little imagination, virtually no artistic skill; Jef-
ferson allows that their moral lapses (lying, stealing, and so forth) can
sometimes be traced to the difficulties of their situation, but even virtues
are transmuted into defects—the courage of African blacks is seen as
absence of forethought. Although he concludes that “the opinion that
they are inferior in the faculties of reason and imagination, must be haz-
arded with great diffidence,” his preceding catalog reveals Jefferson not
only hazarding it but showing little diffidence about the constituent
claims. Decades later, Lincoln echoed Jefferson’s judgment, averring a
“physical difference between the white and black races which I believe
will forever forbid the two races living together on terms of social and
political equality.” 41
The transition from an ethical practice that permits slavery—either as
unproblematic or as problematic but, on balance, acceptable—to one
identifying it as a patent ethical wrong surely looks progressive. How
was it accomplished? A collection of counterarguments systematically
dismantled the justificatory attempts of apologists. They dissect the
evidence for taking black Africans to have inherited some biblical curse;
they note other ways of bringing the African soul to grace than subject-
ing the African body to the middle passage, the slave auction, unremit-
ting toil, sexual, abuse and the lash; they display the accomplishments of
individual slaves, or ex-slaves, whose words and works refute theses of
innate racial difference.42 The overall abolitionist campaign consisted in
40. Thomas Jefferson, Notes on the State of Virginia, in The Portable Thomas Jefferson,
ed. Merrill D. Peterson (New York: Penguin, 1975), 186ff. It is interesting to ask how the
man who wrote these words conceived his relationship with Sally Hemmings.
41. Cited in the “General Introduction,” to Lowance, Against Slavery, xxiv. Perhaps, as
the editor notes (xxv), Lincoln was simply bowing to political pressure.
42. The writings and speeches of Frederick Douglass are notable examples of this
last strategy. In his “General Introduction” to Against Slavery, Mason Lowance tells an
destroying all the devices deployed to avoid applying to people of Afri-

can descent the same attitudes and principles routinely accepted as gov-
erning peaceful interactions among the civilized—it tore down the dis-
tortions allowing Europeans to view Africans as utensils rather than
people.
Besides the negative side of the campaign, there were also positive
discoveries. A few courageous visitors to the parts of the African interior
from which slaves were drawn were surprised to discover communities
with different customs, but with stable social relations and, above all,
familiar human needs and feelings. As more was learned about slave re-
cruitment and the character of transatlantic voyages, many of the vaunted
benefits conferred by enslavement were disclosed as a farrago of non-
sense. Factual discoveries, integrated with strenuous readings of the
scriptures, allowed slaves at last to be seen and to become targets of sym-
pathy. Audiences eventually responded to the eloquence of Douglass
and others, black and white, who cataloged slave suffering, but their ap-
plause depended on earlier advances, made on a more abstract theologi-
cal basis.
Although they did not consistently condemn slavery, American Quak-
ers were often especially concerned with the problem and sometimes
moved to argue for abolition of the institution. Arguably, the pioneering
abolitionist was John Woolman, whose Journal records how he reached
his position. Woolman’s public campaign culminates in his Some Con-
siderations on the Keeping of Negroes (first published in 1754), in which
he argues that “Negroes are our Fellow Creatures.” Woolman’s defense
linked slave suffering to that experienced by the outcasts who excited
Jesus’s sympathy.43 His spiritual odyssey depended as much on his re-
flections on the New Testament as on his experiences of slavery. There
interesting story of one of Douglass’s presentations: “Once during the opening moments of
a lecture in London, an audience expressed hostile disbelief in his past as a chattel slave
because his oratory and elocution were so powerful. (It was well-known that slaves were
held in illiteracy and ignorance as a means of control.) Without speaking another word,
Douglass promptly stripped off his shirt and turned his fl ayed back to the incredulous
audience to show the scars of his floggings” (Against Slavery, xxx).
43. Excerpts from Some Considerations appear in Lowance, Against Slavery, 22–24; see
also John Woolman’s Journal (New York: Citadel, 1961).
never seems to have been some perception of the ethical standing, the
worth of the slaves; what troubled Woolman was the conflict between
the institution and his Christian duty.
The early pages of the Journal explain how the sixteen-year-old Wool-
man “began to love wanton company,” how a period of self-indulgence
was punctuated by ever-longer intervals of remorse, and how eventually
he “recovered” and came back to “live under the cross”.44 As his own
ability to resist temptations increased, he began to be troubled by the
backslidings of others, and to be “uneasy” when he did not remonstrate
with them; uncharitably, we might describe the twenty-three-year-old as
a bit of a busybody. The first (mentioned) concern about slavery arose
when he was asked to perform a task:
My employer, having a negro woman, sold her, and desired me to

write a bill of sale, the man being waiting who bought her. The thing
was sudden; and though I felt uneasy at the thoughts of writing an
instrument of slavery for one of my fellow-creatures, yet I remem-
bered that I was hired by the year, that it was my master who di-
rected me to do it, and that it was an elderly man, a member of our
Society, who bought her; so through weakness I gave way, and wrote
it; but at the executing of it I was so afflicted in my mind, that I said
before my master and the Friend that I believed slave-keeping to be a
practice inconsistent with the Christian religion.45
Shortly afterward, Woolman refused to sign a similar document for a

young acquaintance, also “of our Society.” 46
The language of this passage is telling. The woman sold remains
anonymous. Perhaps Woolman did not know her—he kept his master’s
shop, and lived there alone, at a distance from his employer’s house. Yet
this bare characterization (“a negro woman”) typifies the entire Journal.
Slaves appear in it only under the most abstract descriptions, never per-
ceived as individuals. Woolman provides no extended portrait of their
44. Woolman, Journal, chap. 1; quotes from 4, 5, 8.

45. Ibid., 14–15.
46. Ibid., 15.
behavior and capacities, one that might persuade others to see them as
people. Similarly, the concern with his own spiritual standing, evident
in the hesitations over the bill of sale, is omnipresent. When he discloses
his discomfort, to his master and the purchaser (and to his readers), he
claims an inconsistency between Christianity and slavery. The nature of
that inconsistency is alluded to by the characterization of the woman as
a “fellow creature.”
What led Woolman to draw up the document? He clearly saw it as an
action commended by his duty of obedience (the Christian servant obeys
his Christian master), and yet he knew slaves often suffer (that was surely
the initial cause of his “trouble”). He temporarily suppressed doubts—
the buyer was elderly and a Friend, qualities likely to prevent sexual and
other forms of abuse. As he reflected, however, he recognized his par-
ticipation in an institution prone to inflict suffering on “fellow-creatures,”
and, although the chances of serious abuse in this case seemed remote,
they were not zero. Once the document had been signed and the woman
“conveyed,” there was no guarantee she would not be maltreated. He
would have been partly responsible.
Perhaps I overinterpret the passage. But this reading accords with
Woolman’s subsequent discussions of his growing opposition to slavery.
He is constantly concerned that he is infected by living among those who
support themselves by slavery—interested, too, in saving them from cor-
ruption.47 At times the spiritual plight of slave owners troubles him, and
his reprimands have the character of the sober young man who inter-
vened to save his acquaintances from “wantonness.” Moreover, more thor-
oughly than his predecessors, he takes seriously the Christian apology for
slavery, quoting scripture to rebut the characterization of slaves as inherit-
ing the curse laid upon Ham, and urging priority for the official aim of
redeeming these “lost people”:
If compassion for the Africans, on account of their domestic troubles,

was the real motive of our purchasing them, that spirit of tenderness
being attended to, would incite us to use them kindly, that, as strang-
ers brought out of affliction, their lives might be happy among us.
47. Ibid., 22, 39, 53.

And as they are human creatures, whose souls are as precious as

ours, and who may receive the same help and comfort from the Holy
Scriptures as we do, we could not omit suitable endeavors to instruct
them therein.48
The case Woolman makes to his slave-owning interlocutors, and to his

readers, lies within the abstract framework of Christian duty. His goal is
to remove the blemishes from the Christian community, whether they
are individual propensities to “wantonness,” the “burdensome stone” of
slave ownership, or the traffic in “impure channels,” which distresses
him during his visit to England.49
Woolman made a large and important progressive step. It is hard not
to admire his rejection of slavery, or the courage and perseverance dis-
played in his many attempts to persuade others. His reasons, however,
are not those of any contemporary secular ethical framework. As in earlier
instances, progress is not achieved through some clear new ethical insight.
To be sure, there are genuine cognitive accomplishments, consisting in
the recognition of previously masked facts; Woolman and his successors
appreciate that the view of the African lineage as cursed is groundless,
that conditions in “savage Africa” are hardly ameliorated by shackling
and confining human beings, separating them from their kin, and beating
and raping them; his successors come to see that slaves (and ex-slaves),
with little opportunity and virtually no motivation, can do remarkable
things. Later abolitionists, building on Woolman’s advance, recognize
how altruistic dispositions, shaped in the prevailing ethical practice, are
confined from any extension to black Africans only because the perti-
nent people are kept out of sight, portrayed from a distance as brutish
and incapable of “superior” feelings. The course of the change in atti-
tude, and the consequent growth of sympathy for slaves, is unsteady and
incomplete, even in the late eighteenth and nineteenth centuries (witness
Jefferson, and perhaps Lincoln). It comes about at all only because pro-
foundly devout men and women wrestle with problems of scriptural in-
terpretation, eventually producing the possibility of seeing the sufferings
48. Ibid., 54; see also 53–56.

49. Ibid., 54, 212.
as inflicted on real people. Finally, the men and women routinely bought
and sold are no longer anonymous, no longer undifferentiated “fellow
creatures,” but fully, individually, and equally, human.
§26. The Withering of Vice

My final example represents an entire class of transitions occurring in
the secularization of (some) ethical codes. In these episodes, conduct
previously regarded as wicked, depraved, or sinful comes to be seen as
ethically permissible, and perhaps appropriate for some people. Ethical
concepts prominent in earlier discussions are abandoned or refined. So
citizens of many affluent societies no longer condemn those who express
sexual desires for people of the same sex. Sometimes the shift is only par-
tial: homosexuality is no longer a vice, but still something regrettable—a
sickness, a defect, an incomplete form of sexual fulfillment. When the
transition is thoroughgoing, same-sex preference simply becomes the way
in which some people give direction to their sexual desires, neither in-
trinsically better nor worse than heterosexuality. Terms previously used
to characterize those drawn to their own sex are rejected as prejudiced,
confused, and uncharitable; even broader notions—“vice,” “sin”—come
to seem askew. A more egalitarian view prevails. Homosexual inter-
course, like its heterosexual counterpart, can be loving or exploitative,
tender or cruel, deeply expressive or a shallow pleasure. Homosexual
relationships can vary along all the dimensions of heterosexual ones.
Ethics is not all about regulating sex. Nonetheless, probably from the
beginning, ethical codes have appraised various sorts of sexual activity,
allowing some, forbidding others. When homosexuality is no longer
characterized as a vice, the framework of appraisal is modified. Instead
of focusing on the sexes of the partners (or on the anatomical organs
brought into contact), actions are judged on other grounds: whether they
are coercive, exploitative, in violation of prior promises, and so forth. In
consequence, people who had fought to curb desires that often arose
with great violence within them, people who were compelled to seek
transient expressions of their sexual passions in clandestine and unsatis-
factory encounters, people who constantly feared exposure of their secret
lives, people whose central love for someone else could never be fully
developed in arrangements that openly expressed it, are succeeded by
similar people for whom all these problems are overcome. It is hard not
to view that as ethical progress.
From an older perspective, one still surviving in some societies and in
some groups even within countries that have made the transition, any
tolerance of “deviant” sexuality is a sign of corruption, a mark of ethical
decay. That perspective relies on two major claims, emphasized differ-
ently in different versions. First, homosexual desires are genuinely devi-
ant, unhealthy eruptions within degenerate people, who should be en-
couraged to suppress them in favor of more salubrious (heterosexual)
inclinations. Second, these desires—or, at least, the expression of them
in homosexual behavior—are forbidden by divine command. Accepting
same-sex preference rests upon establishing facts about the prevalence
of homosexual desires and about the consequences of expressing them,
as well as undermining the thought that satisfaction of these desires is
forbidden by the deity.
As in the case of women’s aspirations, discussed in §24, the normality
of the desires is difficult to recognize in a society where they are seen as
deviant. When homosexual acts count as a form of vice, when those who
engage in them are reviled, mocked, and even prosecuted, the society
will lack reliable statistics about same-sex desires and their behavioral
expression. There will be little public knowledge of the character and
consequences of homosexual relationships. Finally, those relationships
will be profoundly and adversely affected by pressures to keep them hid-
den: not only will men and women struggle to find ways of meeting po-
tential partners, forced to seek love furtively in squalid places, but they
are also likely to absorb the social condemnation of what they do, feeling
shame and guilt even while they achieve some temporary satisfaction.
All this supports a public image of homosexual activity as infrequent,
deviant, insalubrious, and stripped of all positive traits associated with
the expression of love.
In part, this picture was rectified through the scientific study of sex-
ual behavior, from sexologists of the late nineteenth century to psycholo-
gists and sociologists of later decades.50 Whether or not its methods and
50. Even studies of sexual behavior that regard homosexuality as defective can play a
liberating role—just as Wollstonecraft’s apparently limited plea for female education
opened the way to broader expression of women’s aspirations. Freud’s recognition of
data were completely reliable, Kinsey’s famous report played a large part
in undermining the repudiation of homosexuality as deviant.51 If men
and women were engaging in homosexual contact at the rates Kinsey
claimed, the effects of the behavior could hardly be so terrible.
Also important was a related shift in ethical practice, acceptance of
the wrongness of treating private consensual homosexual acts as crimi-
nal offenses. Against the background assumption that the law should
intervene only to prevent conduct causing harm to others, increased
understanding of facts about homosexuality induced many countries
to repeal their (frequently harsh) statutes.52 These legal steps neither
modified the common evaluation of homosexual acts as immoral (as
vice) nor removed the stigma associated with homosexuality. 53 To de-
clare oneself a homosexual was an act of great bravery when same-sex
acts were criminal, and it continued to require courage even after le-
galization, when “only” scorn and derision remained. Coming out was
still hard to do.
Yet, just as consciousness-raising was crucial to full public recogni-
tion of the prevalence and extent of women’s aspirations for public
roles, so acts of coming out presented a different picture of homosexual-
ity. Individuals who had previously seemed “normal” and “respectable”
suddenly exposed the “darkness” and “squalor” of their private lives.
widespread homosexual wishes, even though associating them with incomplete develop-
ment, modified prevailing ideas about their frequency.
51. Alfred C. Kinsey, Wardell B. Pomeroy, and Clyde E. Martin, Sexual Behavior in the
Human Male (Philadelphia: W.B. Saunders, 1948); Staff of the Institute for Sex Research,
Indiana University, Alfred C. Kinsey, Wardell B. Pomeroy, Clyde E. Martin, and Paul H.
Gebhard, Sexual Behavior in the Human Female (Philadelphia: W.B. Saunders, 1953);
E. O. Laumann, John H. Gagnon, Robert T. Michael, and Stuart Michaels, The Social Or-
ganization of Sexuality (Chicago: University of Chicago Press, 1994).
52. This depends on a prior ethical shift, adoption of the Millian conception of law
(classically expressed in On Liberty). During the 1960s and 1970s, that conception com-
bined with increased factual knowledge to produce a cascade of liberalizing reforms in
European countries and in some parts of North America (Canada and some states in the
United States, with Illinois leading the way). Denmark (1933) had taken the step much
earlier, and, interestingly, the focus on the private may have inspired France (which had no
antihomosexuality law) to institute a law against public displays of homosexual affection.
53. As Mill so clearly sees (in On Liberty) the effects of social stigma can be just as con-
fi ning as those of the criminal law.
There were so many of them that standard assumptions about normal-

ity and respectability had to shift. The varieties of homosexual rela-
tionships became visible, and so too did the ways in which social attitudes
blocked attainment of positive features people, whatever their sexual
preference, want in their connections with others. As homosexuals re-
sisted invasion of their lives, at Stonewall and after, the initial reaction
to “deviants” who opposed the forces of “law and order” gave way to
sympathy for people prepared to fight for the right to love whom and
how they chose. 54
Making the realities of homosexual desire and homosexual life visible
was one part of the revolution. The other consisted in weakening the
force of the idea that this is a form of sexual behavior proscribed by God.
As some societies, notably in Eu rope, experienced a large decrease in
the proportion of their citizens who accepted the authority of particu-
lar religious texts (the Hebrew Bible, the Old and New Testaments),
justifying injunctions against behavior by appeal to the authority of the
scriptures became increasingly suspect. Even among the devout, how-
ever, emphasizing a ban on “men lying with one another” came to ap-
pear curiously selective. Socially liberal theologians pointed out that
the prescription occurs in a lengthy cata log of rules, almost all of which
are disregarded by Christians and many of which are neglected by Jews.
They commended the central scriptural doctrines, the ones enunci-
ated again and again, illustrated with famous parables. If the will of
the deity is to be honored, we should focus on what is centrally on his
mind.
As with the examples of the previous two sections, it would be folly to
claim that progress has gone as far as it can. Purging evaluation of sexual
activity from any consideration of the sex of the partners—attending to
the relevant qualities of homosexual and heterosexual relations alike—
remains incomplete. The withering of vice depends on achieving a more
selective, and more sophisticated, view of divine commands. It might
have been accelerated by a deeper skepticism about the whole idea.
54. This transition can be traced in newspaper responses to the nascent Gay Pride
movement; see in particu lar the reports in the New York Times, in the immediate aftermath
of Stonewall and in subsequent years.
§27. The Divine Commander

Several episodes of previous sections would have gone differently—
advancing further or faster—if there had been another revolution. From
Plato on, philosophers have scrutinized the idea of grounding ethical
codes in the commands of a deity. Although the arguments presented
are powerful, they have failed to dislodge the idea, still popular around
the world.55 I shall later consider why this might be. First, however, the
arguments.
Plato offers a dilemma. Either there is an independent standard for
assessing the commands issued by the deity 56, or there is not. If there is,
divine commands can be appraised as good or bad, so we can justify our
following them if they are good; but creating this possibility simultane-
ously displaces the deity as the source of ethics; there is a fundamental
measure of ethical goodness (rightness, virtue) prior to the divine acts of
commandment. If, on the other hand, there is no prior source, we can no
longer appraise the deity as good, nor see the commands as anything but
arbitrary expressions of will; in consequence, the injunctions no longer
have ethical force. Kant recapitulates the point succinctly, claiming there
have to be prior sources for the moral law, because, without them, we
could no longer recognize the “Holy One of the Gospels.”57
After the twentieth century’s spectacular organization of social ma-
chines for the brutalization and massacre of human beings, we should be
sensitive to the ethical status of following orders. In Russia and Rwanda,
Johannesburg and Jerusalem, defendants have sought to excuse them-
selves by claiming they were merely following orders issued by authori-
ties. To judge them guilty, as courts and citizens do, presupposes that
55. I simplify. Some religions suppose transcendent beings are impersonal and lack
wills. In these instances, one should speak of prescriptions to “align” oneself with the tran-
scendent forces. Adopting more circumspect language here would be clumsy and obscure
the lines of argument.
56. In Plato’s Euthyphro, the divine source is represented as plural— ethics is a matter of
what the gods love—and Socrates has a preliminary bout with Euthyphro in which he takes
advantage of the possibility of the gods having divergent tastes. This is a flourish on the
main line of reasoning.
57. Kant, Groundwork of the Metaphysics of Morals, a good English translation is that of
Mary Gregor (New York: Cambridge University Press, 1998) (Akademie pagination 408).
the ethical characteristics of commanders affect the ethical status of

those who obey. You can follow orders issued by a wrongdoer without
yourself doing wrong—your boss, who is, unbeknownst to you, an em-
bezzler, tells you to type an apparently innocuous document; but if you
are in a position to detect the commander’s corruption, you should resist
the command. Often, the order given should cause doubts about the
character of the person who issued it, as when you are told to herd the
prisoners into the gas chamber.
Suppose ethics is really founded on a divine command: there is no
prior source of goodness and badness, rightness and wrongness, except
the will of the deity. You hear orders enunciated by the deity’s represen-
tative, or read them in a sacred text. Should you follow? There is no
independent standard by which you can judge the command. Issuing a
different set of prescriptions would be neither better nor worse—the
actual list reflects an arbitrary choice. You might obey, in the way you
drive on the right in many countries. Equally, you might resist. The de-
ity has commanded obedience—but he might equally have ordered dis-
obedience. Why comply with actual orders, rather than those he might
have given? Nothing demands that, except another order, and following
that order has no higher backing than the command to comply with it.
In fact, however, your situation is worse, for sometimes the deity com-
mands people to harm others. He orders a man to kill his son, declares a
geographical region must be taken by force from those who inhabit it
and most of the residents slaughtered, and insists that we put to death (or
at least expel from the community) any men among us who are found
“lying together”.58 You feel uncomfortable about following orders like
these, but you find them in the sacred text and go along. In obeying are
you so different from the functionaries who did their jobs in the machin-
ery of death?
You have independent evidence about the will of this deity. Appar-
ently he demands complete subordination and ser vice: special places are
to be erected for his worship and adoration; his will is to be carried out
58. Leviticus 18 and 20 suggest different punishments for the “lying together.” Simi-
larly, there are variations in what Canaanites are to expect, although, at best, only the
young women will survive.
in everyday life. Those not prepared to follow the commands are to be

punished, and the punishment will be eternal and infinitely agonizing
on all possible dimensions.59 Knowing this, you might be cowed into
submission, just as some of the underlings of twentieth-century dictators
did what they did out of fear of retribution. The deity is very powerful,
the author of the whole show. Sheer power, however, has no bearing on
whether you ought to follow his caprices.
Consider your predicament more carefully. You recognize the deity
has commanded a large number of things, some of them apparently waste-
ful and expressing his narcissism (demanding elaborate forms of wor-
ship), some of them apparently breathtakingly evil. You do not know if
there is an independent ethical standard by which the commands and
the commander himself can be measured. If there is, your own indepen-
dent judgment suggests some of the actions are radically at odds with
it. If there is not, you are simply being ordered to satisfy a caprice, one
alienating you completely from your human sympathies. Compliance is,
at best, ethically neutral and quite possibly ethically incorrect. Hence,
you should surely not follow the order.
An obvious response: who are you to judge? You are a thoroughly fi-
nite being whose knowledge is puny. But you should be clear on just
what sorts of knowledge are pertinent to your predicament. If there is
no antecedent ethical standard, no sense can be given to the idea that the
deity knows more about what ought to be done than you do. Moreover,
there is no sense in which satisfying his caprices is better than respond-
ing to your own human sympathies. He is more powerful than you are
and knows more facts (perhaps all the facts) about the universe he cre-
ated. Nevertheless, without an independent standard, following the or-
ders of the more powerful and factually knowledgeable cannot count as
better than following the orders of the weaker and more factually igno-
rant. On the other hand, if there is an independent standard, perhaps
the deity has an access to it that his finite creatures do not: he has greater
59. See a posthumous essay of David Lewis, “Divine Evil” (in Philosophers without
Gods, ed. Louise Antony [New York: Oxford University Press, 2007]; I completed this es-
say from an outline left by Lewis at his untimely death). The essay considers various pos-
sible ways for Christians to avoid supposing their God causes suffering on an infi nitely
vaster scale than any of the world’s most celebrated human evildoers.
ethical knowledge and transmits it to us in his decrees. When our own

judgment suggests the commands are hideously evil, we should wonder
if our ethical knowledge is partial, and if the deity sees things more clearly
than we do. To follow the orders, however, requires more than the bare
possibility of the deity’s superior insight. We must either have evidence
for thinking the commander has special access to the standards of ethical
correctness, or we must simply take this on trust. The former option is
not available unless we use our own judgment about what the standards
are, and, if we do so, the fact that the deity commands things that are, by
our lights, horrible tells against the hypothesis of special access. In the
end, then, the suggestion must be that we simply have to have faith in
the deity as a source of ethical insights.
This is the best way to think about the divine commander. According
to it, ethical standards are not created by the deity’s fiat, but the deity has
superior knowledge of those standards and communicates the knowledge
to us (or to a few of us); we should trust that this is so and consequently
obey. We are now exactly in the position of the functionaries who defended
their participation in acts of massacre and genocide. The defendant
speaks: “My job was to follow the orders. Although I felt uneasy about
some of these orders, it was not for me to question them. For I trusted
they were given by a leader who saw the whole situation far more clearly
than I could ever do. I had faith in the leader, faith in the superiority of
his judgment to my own, and faith in the rightness of not letting my own
doubts intrude. That’s why I obeyed.” The defense is no more adequate
in the context of following divine commands than when the one in charge
is a human dictator.
Conceiving an unseen enforcer is a useful technique for socializing
members of the group in the ethical code, and thus valuable in cultural
competition (§17). The intellectual problems of viewing ethics as an ex-
pression of the divine will have been articulated by Plato and his succes-
sors, but the arguments fail to dislodge the thesis that the precepts of the
group articulate the commands of the local deity(ies). Why is that?
The answer returns us to a central question of this chapter: Is the
mere-change view acceptable? Ordinary thought about ethics accepts
the possibility of ethical progress and seeks an independent standard
against which ethical practices can be appraised. What could that be,
other than the will of some greater being? Abstract philosophical sub-
stitutes are hard to grasp, or to fit to prominent examples of ethical
advance.60 So, for all its fl aws, the picture of the divine commander
survives.
Understanding the ethical project, its origins, its evolution, and the
historical episodes supporting a conception of ethical progress can free
us from the choice between unconvincing philosophical abstractions
and problematic religious foundations. Showing that will be the work of
the rest of this book.
60. The next chapter will defend this claim. I suspect that many people have an incho-
ate appreciation of it.
pa rt t wo
A Metaethical Perspective
chapter 5
Troubles with Truth
§28. Taking Stock

The history of Part I aims to provide considerations for freeing us from
unsatisfactory conceptions. We turn now to metaethical liberation,
focused on questions about the possibility of truth and knowledge in
ethics. The first goal is to show how standard accounts of possibilities of
ethical truth and knowledge, explanations put forward to resist the “mere-
change view” and its kin, fit the history poorly. Recognizing that will
prepare for a positive proposal in Chapter 6.
The examples of the last chapter show apparent ethical progress. A
handful of instances do not, however, portray the historical unfolding of
ethical practice as a story of constant advances. Far from being prevalent
in history and prehistory, progressive transitions might be quite rare.
Given the large number of normative traditions, and the long period of
time through which they have evolved, the total number of changes in
ethical practice is vast. Most are unknown, and, for times and places at
which ethical change can be studied, little has been done to understand
the character of ethical practice and the ways in which it has evolved.
We have “histories of ethics,” attending to the theories phi losophers
and religious thinkers have constructed, but few studies systematically
173
exploring the ethical practices of societies and their modifications.1

Consequently, little is known about the prevalence of progress in ethical
practice.
An unsystematic review prompts tentative skepticism about the
steadiness of ethical progress. Abolishing chattel slavery was a progres-
sive step. Yet it was preceded, three centuries earlier, by the reinstitu-
tion of a practice—the buying and selling of people—that had been rare
in Europe for a long time; and it was followed, within a few decades, by
practices of discriminating against the newly freed slaves and their de-
scendants. Any Christian advances in spreading compassion in the
Roman world gave way to sectarian warfare: the Romans who exclaimed
at the love Christians displayed toward one another would have re-
thought if they had witnessed the appalling bloodbaths of the contro-
versy over Arianism. Perhaps each progressive change has a regressive
counterpart?
To resist the mere-change view is not to defend the prevalence of ethi-
cal advances. It is to suggest the possibility of progress. If that possibility
makes sense, there must be conditions marking off the progressive tran-
sitions from others (regressive or simply nonprogressive). What con-
straints govern improvement of ethical practice?
To investigate the viability of a notion of progress, it is helpful to start
with the impulse leading to judgments about progressiveness ( judgments
the examples of the last chapter aimed to elicit). Assume people today
1. Even Lecky’s (Victorian) study of a long period in the history of ethics concentrates
more on theoretical ideas than on the actual practices of groups of people. For the ancient
world, the writings of Kenneth Dover (Greek Popular Morality in the Time of Plato and
Aristotle [Oxford: Blackwell, 1974] and Greek Homosexuality [Cambridge, MA: Harvard
University Press, 1978]); Moses Finley (World of Odysseus [New York: Viking, 1978]); Wal-
ter Donlan (The Aristocratic Ideal in Ancient Greece [Lawrence, KS: Coronado Press,
1980]); and Joseph Bryant (Moral Codes and Social Structure in Ancient Greece [Albany:
State University of New York Press, 1996]) point in appropriate directions, as do the dis-
cussions of Peter Brown, The Body and Society (New York: Columbia University Press,
1988) and Jerome Carcopino, Daily Life in Ancient Rome (New Haven: Yale University
Press, 2003), on the Roman world. I know of no sources for other places and later periods
(collectively) as good in elaborating the ethical lives of ordinary citizens. Many historical
works offer particu lar insights, but, without a focus on transitions in practice, ethical
change is hard to study seriously. Hence my (amateur) efforts in the previous chapter.
Troubles with Truth 175
agree in repudiating chattel slavery. They are repelled by the buying and
selling, the harsh treatment, the division of families. Because they judge
ethical progress to have been made when slavery was abolished, they do
not want to go back.2 When they contemplate the world before the rein-
troduction of slavery and the world after slavery returned, they prefer
the earlier state to the later (in this respect, at least).
A first, very simple, subjective criterion for progress can be based on
these reactions: a change in an ethical code is progressive just in case
those who live after the change prefer life in the later world to life in the
earlier one. Is the subjective criterion adequate as an account of ethical
progress? There are ample reasons to worry. Desires expressed in attri-
butions of ethical progress are purely contingent. Were the individuals
who make these judgments to be placed within a rival tradition, one
making different transitions in ethical practice (typically transitions run-
ning counter to those actually preferred), they would be likely to en-
dorse incompatible judgments of progressiveness.3 Human beings may
be malleable enough to be brought—by the right, or the wrong, systems
of training—to issue radically different verdicts on the same situations.
Even when many different traditions agree in their modifications of ethi-
cal practice and retrospective endorsements, history might easily have
gone differently, so consensus judgment is thoroughly contingent.
The skeptic issues a challenge: “You can call this ‘progress’ if you
like, but this is no more than a way of comforting yourself; you have
socially shaped preferences for the kind of life you now lead, in contrast
to the lives your forebears led; you dignify these preferences with a label,
but it is nothing more than an honorary title, masking the contingencies
of the events and of your retrospective judgments.” If things are as the
challenger says, there are two possibilities. Either nothing more can be
done to elaborate or defend the subjective criterion, and it is useless for
2. Or, at least, without the publicly tolerated institution of the seventeenth and eigh-
teenth centuries. They may also be indignant about contemporary practices (illegal prac-
tices) of human trafficking—and believe that stronger measures, both legal and ethical, are
needed to combat them.
3. In fact, actual people who lived through progressive transitions made incompati-
ble assessments. Not all those who lived through emancipation preferred the world it
produced.
opposing the mere-change view, or the subjective criterion can be devel-

oped to meet the challenge. Start with the latter option.
A point must be conceded to the skeptic. Often in the history of ethi-
cal practice, a modification is made at one stage, only to be reversed
later. Decisions to afford certain groups of people protections and op-
portunities are sometimes unstable: consider attitudes towards Jews in
Western societies, from the Middle Ages to the present; similarly, com-
munities have changed their minds about the fair distribution of wealth
or the responsibilities of the rich to provide for the poor. Pending a revi-
sionary reconstruction of these apparent oscillations, mere postrevolu-
tionary satisfaction looks too weak to suffice for ethical progress. The
constraint allows progress to occur whenever people are immediately
happy with a change. Modifications in contrary directions can count
equally as progressive.
An obvious remedy suggests itself. Not only do progressive changes
require postrevolutionaries to prefer life under the new dispensation to
life under the old, but the preference must also endure. As the ethical proj-
ect evolves, later generations must endorse the preference, looking back
on the revolution as a progressive step. So strengthened, the criterion
avoids the simplest difficulties, posed by rapid oscillations, but it needs
to fi x a period to serve for certifying past progress. How long must the
preference endure? Any prescription seems arbitrary and, furthermore,
subject to the obvious possibility of reversals occurring at a slower pace:
for any chosen span of time, a preference could endure throughout that
span, only to be reversed later.
Very well. Demand something yet stronger: the desires must be not
only stable throughout some period after the revolution, but stable “in
the limit.” 4 Now the subjective criterion answers the worries about os-
cillations besetting weaker versions, but the notion of the “limit” of ethi-
4. Here, there are obvious connections with ideas of classical pragmatism, specifically
to the account of truth often attributed to Peirce in Nathan Houser and Christian Kloesel,
eds., The Essential Peirce, vol. 1 (Bloomington: University of Indiana Press, 109–123) and
apparently also present in James. In the ethical context, the idea surfaces in Jamesian refer-
ences to “the last word of the last man.” William James, “The Moral Phi losophers and the
Moral Life,” in The Will to Believe and Other Essays (Cambridge, MA: Harvard University
Press, 1979).
cal practices needs explanation. Should this notion be tied to the actual
course of human history, or should it be understood in terms of some
idealization of our future? The actual past includes occasions on which
apparently progressive ethical transitions have been reversed—for ex-
ample, episodes in which par tic u lar groups have been protected by
the ethical code, only later to be excluded again. To judge the inclusions
as progressive is to diagnose the reversals as events of terrible social
blindness. At some periods in human history, people—socially blinded
people—have preferred to live in a world produced by a regressive transi-
tion, and there are potential causes that induce social blindness in any
group. What guarantees the absence of these causes in the limit of the
extension of our ethical practices? Understanding the criterion in terms
of the actual course of human history, it cannot be expected to judge
progress correctly.
Again, there is an obvious remedy. Require the preferences to be sta-
ble, in the limit, in a history that proceeds under “ideal” conditions: a
transition within ethical practice is progressive if and only if, in a future
proceeding from the transition and modifying practice under ideal con-
ditions, there is a stage after which the preference for life after the transi-
tion is never reversed. This version is superior to any previously consid-
ered, but it contains, at its core, an unexplained notion, that of the ideal
conditions these futures must satisfy. The ideal conditions depend on
certain kinds of perturbing forces being absent, the variety of causes
producing the forms of blindness of our ethical past. Historians are di-
vided on how to account for socially induced ethical blindness, even in
the most well-studied episodes.5 Because we understand so little of ear-
lier modifications in ethical practices, let alone the potential causes of re-
gressive steps in it, the causal factors diverting the past course of ethical
evolution from the ideal state cannot be specified. Nor can we hope to
identify further factors possibly affecting our descendants. Our only
purchase on the idea of the ideal unfolding of ethical practice is through
5. There is a large and sophisticated literature devoted to questions about how German
citizens could have avoided seeing what was done to the “outcasts” under the Nazi regime,
but it is hard to draw from it any defi nite cata log of all the perturbing forces potentially at
work.
the distortion of human preferences: the perturbing forces are just those
leading people to prefer to live in the world produced by a regressive
transition. That specification of the “perturbing forces” would make the
subjective criterion quite hopeless. Transitions count as progressive just
in case the worlds they produce are preferred by people unperturbed by
forces generating misjudgments about progressiveness.
Even in its best version, the subjective criterion is not good enough. It
reduces to banal circularity. Furthermore, it introduces epistemological
problems. We judge the transitions of the last chapter as progressive. Are
we committed to a prediction about the future course of ethical evolu-
tion, actual or ideal? If progress was made, preferences for life after
slavery or life with greater female equality will be shared by our descen-
dants—so long as they are not “improperly swayed” by unspecifiable
forces. How can we assess that? The judgments of progress are grounded
in aspects of the episodes themselves, not because we anticipate the
eventual stability of human desires (assuming no “perturbations”). Any
confidence we have about the future, if it proceeds “properly,” rests on
thinking there is something about the transition to which people who
come after us will continue to respond (insofar as their vision is not dis-
torted). Their desires, like ours, are secondary, symptoms of the pro-
gressiveness of the transitions, not constitutive of it.
§29. Prima Facie Problems

The most obvious alternative to the mere-change view introduces an
idea of ethical truth. Progress occurs when an ethical practice substi-
tutes truth for falsehood. More generally, you might think of ethics as
a form of inquiry subject to external constraints, constraints beyond the
contingent preferences people have. In the rest of this chapter, that gen-
eral thought will be confronted with the past evolution of the ethical
project.
We have examples of apparently progressive changes during the past
few millennia, and some larger modifications, about which we know
much less, occurring in the more distant past. The examples invite obvi-
ous questions: Where exactly do historical actors bump up against the
external constraints and acknowledge their force? How did the con-
straints play a role in the larger changes of the Paleolithic and early Neo-
lithic? I offer blunt answers, to be defended more carefully in this and
the two following sections: the transitions recorded in the historical
record, as well as those hypothesized for the prehistoric past, are best
conceived as “local adaptations,” not episodes of ethical discovery.6 The
changes appear to be responses to difficulties of the social situations in
which individuals and groups find themselves. “Moments of ethical in-
sight” are elusive. My first aim is to show these answers to be prima facie
plausible.
The historical figures who figure in ethical transitions, the vast ma-
jority of them unidentifiable as individuals, do not start from some situ-
ation in which they lack ethical convictions, follow a process of reason-
ing or observe some facet of reality, and thereby arrive at a well-grounded
belief in an ethical judgment. Actual historical agents (and their prehis-
toric counterparts) were born into societies and socialized from early
childhood. They acquired practices of expressing ethical evaluations,
an extensive repertoire of ethical concepts, and dispositions to accept a
body of ethical statements, most of which they never questioned. For
the revisionary historical actors who stand out relatively clearly—Mary
Wollstonecraft, John Woolman—what occurs is a change in ethical con-
viction: ethical beliefs transmitted within the society and shared by ev-
erybody around are rejected in favor of claims incompatible with them.
The psychological processes these people seem to undergo differ rad-
ically from the forms of evidence conjured in typical philosophical
accounts of ethical justification. There are no special abstract forms of
reasoning, nothing reported as a moment of “perception” or “intuition.”
Reformers take up the ethical project as framed in their culture, making
proposals on the basis of empirical information they find salient.7 In light
of the background ethical views inculcated in the society, women act
well by performing particular tasks as wives and mothers, while black
6. Here I recapitulate themes presented in a pregnant passage from John Dewey, Hu-
man Nature and Conduct (Amherst NY: Prometheus Books, 2002) 103.
7. Their justificatory achievements and the limits of what they can support will become
fully clear only after the account of ethical method in Chapter 9. For the time being,
though, I emphasize only the differences from the views of those who treat ethics in terms
of conformity to external constraints.
slaves need to be brought to godliness. Wollstonecraft recognizes facts

about the world: uneducated women often fail at some of the allegedly
important tasks. So too does Woolman: slaves are not made more God-
fearing by being treated as they usually are, and their owners do not ex-
emplify standards of godliness. Recognizing these facts, the reformers
construct arguments bearing on prevailing ethical assumptions. The
discoveries they make about the world are not at all mysterious. It may
have taken an acute observer, one sensitized by his or her own worries
about personal salvation, to see how socially accepted behavior affected
slaves and slave owners, but ordinary forms of observation can deliver
the kind of information Woolman acquired and used. The forms of rea-
soning are of types well understood from other contexts: if women should
perform particular tasks as wives and mothers, and educated women do
a superior job (by current standards), women should be educated. To
have uncovered the facts, seen their salience, and made the case were sig-
nificant achievements, but they did not turn on some novel apprehension
of “ethical reality.”
Compare episodes in which prevalent scientific beliefs are changed.
Historical studies have brought home the difficulties attending the dis-
closure of radical novelties: seeing the swaying incense burner as a
pendulum requires a shift of perspective hard to achieve.8 Occasionally,
however, the unexpected forces itself upon the observer; Röntgen can-
not overlook the fluorescence on the screen. What analogs can be found
in episodes of ethical change? What psychological processes go on in
innovators, and how are the constraints on ethical progress registered in
their thinking or feeling? How exactly do reason, intuition, and percep-
tion work to generate new ethical insights? Woolman and Wollstonecraft
saw aspects of the world those around them had not brought into focus;
ancient aristocrats probably learned that, without a phalanx of hoplites,
their city-states were indefensible. Once those features were widely
8. Here the loci classici are Thomas S. Kuhn’s Structure of Scientific Revolutions (Chi-
cago: University of Chicago Press, 1962 and 1970) and N. R. Hanson, Patterns of Discovery
(Cambridge, UK: Cambridge University Press, 1958). The example of the pendulum is
Kuhn’s, and seems to me one in which his invocation of a “gestalt switch” is most
convincing.
appreciated, continuing ethical practice in the traditional ways became

socially problematic.
Consider what histories supporting philosophical talk of subjects
“perceiving” or “intuiting” good or bad, right or wrong, would be like.
Woolman might simply have “seen” the injustice of treating an individ-
ual slave in one of the usual abusive ways; Wollstonecraft might have
“intuited” the wrongful oppression to which members of her sex were
subjected. Reflecting on the crucial processes, they could have offered
pointers to help others attain the insights vouchsafed to them. Although
their efforts to persuade are extensive and passionate, neither attempts
anything of the sort. They are radically unlike the scientific figures who
undergo transformative observations. Galileo teaches his readers to see
the swinging censer differently; Röntgen shows the fluorescent screen.
What could encounters with external constraints be like? Some phi-
losophers have supposed we can have contact with something deserving
the name of “ethical reality,” and that people have psychological capaci-
ties enabling them to arrive at well-grounded ethical judgments.9 Imag-
ine you come across some boys dousing a cat with gasoline and igniting
it. You judge the action to be wrong. The judgment is immediate, tempt-
ing you to say you see the wrongness of what the boys do, just as you see
the unfolding event. In the same way we can gain knowledge of cats, gas-
oline, and boys through perception, we can also learn about the good-
ness or badness of states of affairs, the rightness or wrongness of actions.
Or perhaps judgment is mediated by a feeling of repugnance, a violent
antipathy to what occurs, so the negative reaction forms the basis for our
ethical assessment. The perception puts us into an affective state, and
being in that state warrants our ethical judgment.
There is a better explanation. You make the judgment, and if you
make it immediately, you do so because your society has inculcated psy-
chological propensities to apply the vocabulary as you do. If you feel
9. This position has been most clearly and precisely articulated by Nicholas Sturgeon;
particularly valuable is his exchange with Gilbert Harman. See Nicholas Sturgeon, “Moral
Explanations,” in Morality, Reason, and Truth, ed. David Copp and Dean Zimmerman
(Totowa, NJ: Rowman and Littlefield, 1985), and Gilbert Harman, “Moral Explanations
of Natural Facts,” Southern Journal of Philosophy 24 (1986): 57– 68. In the following text,
I use an example extensively discussed in this exchange.
particular emotions—horror, repugnance—it is because your society has

connected affective states with cognitions and volitions and has marked
out certain types of sentiments as “ethical,” rather than merely idiosyn-
cratic responses. In either version, your psychology has been shaped to
generate, on perceiving particular cues (the cat’s agonized squeals and
squirming), an immediate, or relatively immediate, judgment.10 Back-
ground ethical practices of your society, communicated to you in your
early socialization, underlie your judgment. For those who propose mod-
ifications of practice, the requisite preparation for such direct responses
is absent. Without something like a fogged photographic plate to subvert
prior expectations, Woolman or Wollstonecraft would have no ability
for direct assessment.
Socialization plays an analogous role in ordinary observation and in
the refined scientific versions of it. The technician observing the bubble
chamber or the sequencing machine arrives at judgments about types of
particle collisions and characteristics of the DNA. Here too, immediate
responses depend on prior training: this person does not have any pecu-
liar ability to “see” subatomic particles; she differs from others only in
having been taught to respond to particular types of tracks in prescribed
ways. Her training enables her to gain knowledge of events involving
subatomic particles, but it does so only in virtue of historical events
through which these initially inaccessible aspects of reality were detected
and ways discovered to make them manifest. The knowledge gained
through current observation depends on the achievements of a previous
group of observers, who were able to transform a world in which the re-
mote particles were unknown to a world in which they could be detected
by the contemporary devices. Historians and philosophers can recon-
struct the processes of observation and reasoning behind our enhanced
ability to observe nature. Observable phenomena provided a basis for
relying on microscopes, techniques of microscopy were extended to re-
veal Brownian motion, recognition of Brownian motion led to knowl-
edge of atoms and manufacture of devices for fathoming the properties
of the atom, experiments with these devices led to the detection of the
10. In the version where your evaluation depends on appreciating your own emotional
response, there is an intermediate step, but the process from perception to judgment re-
mains relatively direct.
subatomic world, and so on. The history is long and complex, but our
confidence in the quick judgments of the technicians who observe tracks
in bubble chambers would be radically undermined if we thought it
could not be told.11
Back to your horrified witnessing of the tortured cat. Like the techni-
cian at the bubble chamber, you make an immediate judgment—“There’s
a positron!,” “That’s wrong!”—and you do so by exercising socially in-
culcated psychological dispositions. To view you in this way is not yet to
cast doubt on the thesis that you are making contact with some external
constraint and thereby arriving at new ethical knowledge. Sustaining
the thesis, however, requires a story about the historical background to
your belief, and to the dispositions (“techniques”) instilled in you, one
similar to the tale told for the detection of the subatomic world. Recon-
structing history requires tracing a justified route—disclosing processes
of reasoning, perception, intuition, or whatever—that led from the stage
at which people looked on at squirming animals with indifference (mak-
ing no ethical judgments about the actions producing the writhings) to
the stage at which the judgments (“That’s wrong!”) became firmly ac-
cepted and the young became socialized to have propensities for making
such judgments.
This particular ethical change did not figure in the catalog of the last
chapter. Its history is obscure—even the epochs and societies within
which it occurred cannot be identified.12 We can, however, consider
ethical revolutions for which we have historical evidence, and inquire
11. Historians and phi losophers of science have paid great attention to this example, so
it is reasonable to claim confidence here. See, for example, the writings of Mary-Jo Nye
(Molecular Reality [London: McDonald, 1972]), Wesley Salmon (Reality and Rationality
[New York: Oxford University Press, 2005]), Ian Hacking (“Do We See Through a Micro-
scope?” in Representing and Intervening [Cambridge, UK: Cambridge University Press,
1985]), and Peter Achinstein (“Is There a Valid Experimental Argument for Scientific Re-
alism?,” Journal of Philosophy [2002] vol 99 pp. 470–495).
12. I suspect it is very ancient. An obvious conjecture is that it occurred fi rst in connec-
tion with domestic animals—but that is no more than a conjecture. Some thinkers will
maintain, with considerable reason, that the “revolution” of responding to nonhuman
animal suffering is still far from complete. (A classic is Peter Singer, Animal Liberation
[New York: Random House, 1975]). As we shall see, ethical conclusions about nonhuman
animals raise difficulties for my own approach. For discussion, and an attempt at resolu-
tion, see §47.
whether in these instances a parallel to the story about the discovery of

subatomic reality is available. The ways people like Wollstonecraft and
Woolman arrived at and defended their proposed changes in ethical
practice are absolutely critical, for these are the places where the con-
straining power of the external sources will emerge—if it ever does. The
rejection of chattel slavery provides an especially good instance for
these purposes because here our access to the psychological lives of
those who might have made “ethical discoveries” is relatively good (al-
though there is much even about John Woolman’s psychological devel-
opment that remains unknown).13 Nothing in the recorded testimony of
any abolitionist discloses any analog of the critical observation in which
external constraints are apprehended differently (and correctly)—the
analog of Röntgen’s fluorescent screen. Woolman’s Journal, the most
revealing document we have, contains no mention of an occasion when
he saw the plight of slaves in a new way (a sudden revulsion at the bat-
tered bodies before him, that—somehow—transmitted ethical insight).
As I remarked (§25), black slaves do not appear as individuals in Wool-
man’s narrative—nor do individualized slave owners whose “corrup-
tion” Woolman might suddenly “perceive.”
Woolman does feel an emotion—an “unease”—on the pivotal occa-
sion when his master asks him to draw up a bill of sale. As background to
his experiences, the Journal presents his repudiation of his own past
“wantonness,” his renewed discipline, often expressed in reprobation of
others. We are given a portrait of a young man who views the conven-
tions of behavior, even among the Friends (the Quakers), as not always
sufficiently scrupulous, and who minutely scrutinizes his own conduct
to forestall possible lapses. The “unease” takes very explicit forms, ex-
hibited in the reflections leading him to quiet it and perform the task
assigned him: the buyer is elderly and himself a Quaker, thus decreasing
the chances of sexual and physical abuse and increasing those of spiritual
guidance for the slave. In the background are facts Woolman recalls as
13. It is also a good case for present purposes because it is one principal defenders of the
idea of contact with external ethical constraints typically cite. Woolman is preeminent
among early abolitionists for explaining the circumstances leading to his new ethical
stance.
he ponders what to do—slaves are often beaten, female slaves are often
sexually coerced, concern for the spiritual development of slaves is rare.
He is suddenly asked to do something that makes him, if only tangen-
tially, complicit with these common features of the slave-owning insti-
tution, and, for a young man so sensitive to his spiritual temperature,
even that tangential involvement promotes unease. This psychological
explanation fits the record we have far better than any (nebulous) pro-
cess through which some (mysterious) external constraint manifests
itself.
It is instructive to contrast the historical sources available in the ethi-
cal case with those in another instance in which philosophers want to
explore possibilities of new (disputed) knowledge. In The Varieties of
Religious Experience, William James canvasses the reports of those
who acquired a new belief about “the transcendent,” and his discussion is
fully based on particular episodes in which people “saw” something strik-
ingly new. Whatever scruples we may have about the reliability of the
processes James’s subjects underwent, reports of this kind are, in prin-
ciple, the right kind of material from which a defense of new religious
knowledge might be drawn. Historical documents describe the experi-
ences and their apparent power. In the repudiation of chattel slavery,
records of any similar experiences, moments of sudden revelation, are
entirely absent.
If progressive ethical inquiry increases conformity with external con-
straints, the absence of episodes in which such constraints are recog-
nized, even if only dimly, is thoroughly perplexing. When we reflect on
other examples of ethical revolutions considered in the previous chapter,
there is simply no evidence of times and places at which some sense of
these constraints modified ethical practice, allowing for the institution
of reliable techniques for everyday “observation” (ways of cultivating
psychological dispositions enabling people to “see” particular states and
actions as good or bad, right or wrong). For some transitions, the idea is
absurd. Greek replacement of the ideal of heroic courage with that of
solidarity has far more to do with the technology of fighting than it does
with any moment of insight disclosing the ethical flaws of the Homeric
hero or the virtues of the hoplite. Nobody would insist on moments of
ethical insight, analogs to the scientific observations perturbing preva-
lent ideas about nature, if the grip of a background picture of ethics were
not so powerful as to make it appear that such moments have to have
occurred.
So far, a prima facie challenge for the thesis that ethics is a form of in-
quiry responding to external constraints. Defenders of the thesis must
explain how those external constraints play some substantive role in the
evolution of the ethical project. Otherwise invoking them is idle, a piece
of comforting rhetoric easily excised. Yet the issue should not be left
there, with a pointed invitation to tell a different narrative or to interpret
the history of Part I differently from ways so far suggested. Diagnosis of
the flaws of popular philosophical views is intended to fashion a better
approach to ethical progress. Hence, the problems just posed should be
examined more deeply, to reveal how serious the challenge is.
§30. Truth, Realism, and Constructivism

Interpreting ethical progress as consisting in the attainment of new
truths has evident attractions. Although ethical codes are often distin-
guished by their rules and paradigms for behavior, the latter often en-
capsulated in inspiring stories, the codes of most societies employ spe-
cial concepts to reformulate the content of commands as (what appear to
be) descriptive statements: “Do X!” or “Act as Y did!” are accompanied
by “It is right to do X” and “What Y did was good.”14 If such statements
are true or false, the progressiveness of a step in which particular new
commands were introduced (or in which old rules were modified or
dropped) can be understood by considering the descriptive counterparts
of the commands and rules (statements that evaluate the actions required
or commended as right, or good, or virtuous), taking the descriptive coun-
terparts to eliminate falsehood in favor of truth. Returning to episodes
from the last chapter, the modification of the lex talionis is progressive
because the statement “It is wrong to punish someone other than the
14. Anthropologists recognize cases in which the name of the group substitutes for the
vocabulary used to recommend or command actions. For the purposes of this chapter, it
does not matter whether the predicates used are “good”/“bad,” or “right”/“wrong,” or
“virtuous”/“vicious,” or even the names applied to a specific group and to those it counts
as outsiders.
perpetrator” is true; and the abolition of chattel slavery is progressive

because “Slavery is permissible” is false (as is the claim apparently es-
poused by Cotton Mather and others: “Slavery is a good thing”).
The challenge posed in the last section can now be re-posed: how do
you integrate an account of ethical progress as the substitution of ethical
truth for ethical falsehood (or the accumulation of ethical truth) with the
actual evolution of the ethical project? There are two main possibilities.
Either there are episodes of ethical insight, occasions on which new ethi-
cal truths are discerned, or progressive transitions, those attaining new
truth, have to be viewed as fortunate occasions in which blind stumbling
turns out well.
The latter option proposes that ethical practice evolves in ways not
requiring the apprehension of the constraints, but sometimes turning
out to conform to them. When this occurs, the historical actors are un-
aware of the fact. They argue with one another and eventually reach peace
on quite different grounds. For example, they take an important progres-
sive step in extending some of their prescriptions to cover interactions
with a neighboring band because they appreciate the possible benefits
of trade (§19). Only much later thinkers, perhaps only those with a syn-
thetic vision of the ethical project as a whole, can understand how they
have unwittingly responded to the external constraints. Is this a satisfac-
tory amendment of the original idea, one able to acknowledge the his-
tory of the ethical project and also to find a place for the idea of external
constraint?
People arrive at true statements in a variety of ways, including ways
providing them with no justification for their new beliefs. If ethical prog-
ress pervaded the history of our practices, it would be hard to credit
such “sleepwalking.” Where progress is systematic and sure-footed, as it
is in mathematics and in the sciences, there would be genuine problems
in integrating the thought of progress as consisting in the accumulation
of truth with denying episodes of insight: it would be reasonable to won-
der how human beings could, so consistently, be so lucky. With respect
to ethics, however, where progress seems unsteady, no such worry about
a happy series of coincidences arises. The sleepwalkers stumble along,
often, indeed perhaps most of the time, lurching from error to error, but
occasionally lighting upon new ethical truths. After the fact, matters be-
come clearer, as expressed in the confidence of some judgments about

ethical progress. Eventually those latter-day evaluations will have to be
understood, recognized as knowledge—and that will demand some ac-
count of how the knowledge is gained. Providing the account poses dif-
ficulties of its own. For the moment, however, expand the menu of pos-
sibilities for meeting the challenge of the last section by supposing the
external constraints are not apprehended by the participants, but only
by much later thinkers, gifted with special insights into the character of
ethical transitions.
Consequently, the challenge could be met in one of three different
ways: by developing an account of ethical truth, in light of which epi-
sodes of ethical change could be reconstructed to show how individual
participants apprehended these truths; by using the account of ethical
truth to understand the capabilities of the later thinkers who can at last
recognize the external constraints; or by proposing an account of ethical
statements on which they are not seen as having truth-values.15 Those
who suspect the problems of the last section are only prima facie difficul-
ties may maintain that the appearance is generated because the un-
analyzed notion of external constraint is tacitly interpreted in crude ways,
so the recognition of external constraints comes to seem much more
mysterious than it really is. They may expect an account of ethical truth
to clear up the trouble. The next step, then, is to understand the notion
of ethical truth.
The approach to truth most obviously suited to a thesis about exter-
nal constraints is that associated with the sciences and with everyday
descriptions of the world around us. In these domains, progress can be
conceived as the acquisition of significant truth, where truth is under-
stood in familiar terms, as correspondence to reality.16 The conception
15. This last possibility would then need to be elaborated with some alternative account
of the external constraints, supplemented with a proposal as to how these constraints are
recognized either by participants in episodes of change or by the clever folk who come later.
16. My The Advancement of Science (New York: Oxford University Press, 1993), as cor-
rected by Science, Truth, and Democracy (New York: Oxford University Press, 2001), pro-
vides an approach of this type. Other accounts are possible, for example, approaches con-
ceiving progress as problem solving. Those would be closer to the perspective I shall
eventually take in the ethical case.
of correspondence does not presuppose peculiar composite entities,

“facts,” to which true statements correspond. Rather, to put the point
pedantically, simple atomic statements, “Jakie helped Krom,” for exam-
ple, are true in virtue of the referential relations between logically con-
stituent terms and parts of reality, together with the set-theoretical rela-
tions of inclusion. “Jakie helped Krom” is true because “Jakie” and
“Krom” are singular terms, picking out particular objects in the world,
and “help” is a two-place relational term picking out a relation (a set of
ordered pairs), and the pair <Jakie, Krom> belongs to that set.17 The truth
of more complex statements is understood in terms of the conditions laid
down by Tarski for the language of first-order logic.18 This understand-
ing of truth for scientific sentences combines well with an account of
progress in terms of the attainment of truth and with the actual historical
development of the sciences.19 For humbler types of scientific statements,
we find no great mysteries: Frans de Waal was able to recognize the truth
of “Jakie helped Krom” because he could observe Jakie, Krom, and the
helping behavior. When the entities whose properties are recorded in
true statements are more remote, extra philosophical work must be done
to show how access to these entities becomes possible: but that work
can, and has, been done (as in the instance of subatomic particles, con-
sidered in the previous section).
Because this approach to truth lends itself so easily to the acceptance
of external constraints, it is good to begin with it; other conceptions of
17. Here I rely on the Tarskian account of the truth of logically simple sentences of first-
order languages (sentences that do not contain connectives or quantifiers). I abstract from
issues about tense.
18. Alfred Tarski “The Concept of Truth in Formalized Languages,” in Logic, Seman-
tics, Metamathematics (Oxford University Press, 1956), 152–278; in his important essay
“Tarski’s Theory of Truth,” Journal of Philosophy (1972), Hartry Field shows how Tarski
generates the notion of truth from the notion of reference. The correspondence theory I
espouse does not assume, with Field’s article, that it is possible to reduce the notion of ref-
erence to some physicalist basis. See my essay “On the Explanatory Power of Correspon-
dence Truth,” Philosophy and Phenomenological Research 64 (2002): 346– 64.
19. There is a line of thought, descending from Kuhn’s Structure of Scientific Revolu-
tions, that denies this claim. For defense, see my The Advancement of Science and Science,
Truth, and Democracy. See also my “Real Realism: The Galilean Strategy,” Philosophical
Review 110 (2001): 151– 97, and “On the Explanatory Power.”
truth will occupy us later. How would a correspondence account apply

to the ethical case? Recall the simple example: some boys douse a cat
with gasoline and ignite it. Call this event E. “E is wrong” appears to be
a true ethical statement (we might normally use less restrained and ab-
stract vocabulary). On a correspondence account, its truth will stem
from the fact that “E” picks out an event and “wrong” refers to a class of
events, a class in which E is included. No trouble threatens with the first
part of this condition; any anxiety must result from difficulties with under-
standing the reference of the ethical predicate.
How does “wrong” come to refer to a particular class of events? Dis-
tinguish two broad possibilities. On a constructivist approach to ethical
reference, we stipulate that particular events are to be labeled as “wrong”;
the boundaries of this class are matters for our decision, not fi xed ante-
cedently to the decision. On a realist alternative, there is a preexistent
division between actions that are wrong and those that are not, and
the only place for human construction or convention lies in decisions
about which sounds or signs to use in marking the distinction. Within
each of these broad approaches, further important distinctions should
be drawn.
Consider, first, the realist alternative. In attempting to integrate ethi-
cal truth and its apprehension with the history of ethical practice, it can
adopt one of two rival stances concerning property detection. The prop-
erty marked out by a predicate may be simple and irreducible, in the sense
that no elaboration of conditions necessary and sufficient for the presence
of the property is required to explain how people detect that objects
(at least some objects) have it. The property of being an approximately
straight line looks to be of this type. Although geometry texts sometimes
define a straight line as the shortest distance between two points, we do
not acquire the belief that a path is approximately straight by investigat-
ing all the possible alternative routes between its endpoints and satisfy-
ing ourselves that this one is (among) the shortest. Other properties, the
complex and reducible ones, need subsidiary conditions to play a role in
their detection. You attribute the property of being cilantro to the green
stuff in the grocery store by attending to the shape of the leaves, and, if you
worry about possible confusion with the parsley, by sniffing it; detection
proceeds by explicitly noting the conditions. Properties appearing sim-
ple and irreducible may turn out to be complex and reducible, however,
not because subsidiary conditions are consciously noted, but because
the explanation of their detection involves unconscious apprehension
of those conditions. If color properties apply to macroscopic objects in
virtue of the disposition of those objects to scatter light of particu lar
wavelengths, our detection of these color properties will be explained in
terms of our ways of responding to the characteristic patterns of scatter-
ing.20 A realist approach to ethical truth may thus suppose the proper-
ties marked out by ethical predicates (“good,” “wrong,” and so forth) are
simple and irreducible or complex and reducible; with respect to the
latter alternative, the subsidiary conditions may be consciously noted, or
unconscious apprehension may play a role in detection.
Much more could be said about these options, but brief explanations
are sufficient for refining the challenge of §29. Consider next the possi-
bilities for constructivism. The very simplest version of a constructivist
approach to ethical truth takes the reference of ethical predicates (“wrong,”
for example) as simply a matter for conventional determination. Any
group, or maybe even any individual, can decide, on any grounds or on
none, to apply the predicate so it picks out any set of objects, states of af-
fairs, or events. This is the Humpty Dumpty theory of ethical truth—with
respect to the key ethical vocabulary, we, collectively or individually, are
always the masters—and it allows any ethical transition whatsoever to be
“progressive” from the viewpoint of those who make it, or, since all can
be assimilated, abandons any contrast between progress and regress. It
capitulates to the mere-change view.
More interesting versions of constructivism suppose there are condi-
tions on the processes through which the reference of ethical predicates
is determined. The set of events marked out as wrong consists of those
events acquiring a special status if particular procedures were followed.
20. I do not suppose this is the correct account of color properties. It is a popu lar one,
and hence useful for introducing an option a realist about ethical truth may try to exploit.
The example may inspire some to retract the supposition that the property of being an ap-
proximately straight line counts as simple and irreducible: perhaps the only such proper-
ties are geometrically more basic, or perhaps there are no such properties at all. I need not
quarrel with such reactions. Since I shall conclude that realist truth about ethics cannot be
sustained in any of its guises, it is best to be inclusive about the possibilities allowed.
So, for example, there might be some form of reasoning each individual
can undergo that divides actions into a number of types; or there might
be some hypothetical social process, in which groups of people can en-
gage, generating distinctions among events; or there might be some ac-
tual social process either generating or revising such distinctions. Kant
and his successors pursue the first option, claiming there are processes
of reason, available to all rational beings whatever their social and physi-
cal environments, yielding conclusions about the status of actions. Social
contract accounts of the traditional kinds develop the second, suggesting
that ethical distinctions among actions derive from the deliberations
people would make under ideal conditions. The third is developed in the
next chapter, and, since the present focus is on rival perspectives, fur-
ther consideration of it is postponed till then.
Both realists and constructivists have several ways of developing the
thesis that ethics is governed by external constraints. All the alternatives
so far reviewed adopt the framework of truth as correspondence. One
rival approach deserves a mention. Instead of considering the structure
of truth, how the truth of statements arises (what makes truths true), one
may adopt a functional account, seeking to understand what we aim at
in various areas of inquiry.21 For descriptive statements about the physi-
cal world, from common sense to refined science, a correspondence ac-
count seems to deliver both structure and function. By contrast, applying
a correspondence theory of truth to mathematics appears problematic,
precisely because it requires us to suppose the existence of a realm of
abstract objects, whose properties are ever-more precisely and more
completely described by the mathematical statements accepted, and be-
cause the ways in which mathematical statements come to be accepted
make it utterly mysterious how mathematicians are gaining access to that
realm. The great figures in the history of the subject, those to whom we
attribute the most significant advances, respond to earlier problems by in-
troducing new notation, often inspired by attempts to carry out symbolic
manipulations on a broader scale. They seem to be expanding the lan-
21. The distinction is drawn in Michael Dummett’s important essay “Truth” (reprinted
in his Truth and Other Enigmas [Cambridge, MA: Harvard University Press, 1978], 1–24).
Dummett does not use the terms I employ to mark the distinction.
guage, thereby introducing new games for mathematicians to play. Per-

haps a superior account of mathematics would abandon the enterprise of
saying what mathematics is about, and concentrate on characterizing the
kinds of statements at which mathematical practice aims. It would offer a
functionalist account of truth.
We might approach ethical practice in similar fashion, seeking the
functions that ethical prescriptions and other parts of ethical codes
are to serve. In the next chapter, ethical progress and ethical truth are
considered in this light. As we shall see, the result is very different
from the usual ways of viewing ethical progress as conformity to exter-
nal constraints.
We now have an array of possibilities for developing a response to the
prima facie concerns of §29. Is any of them adequate?
§31. The Sources of the Troubles

We can profitably begin by countering one attractive gambit. Section 29
disclosed some important historical actors—Wollstonecraft and Wool-
man—as discovering important factual truths. They use their discover-
ies to argue, against the background ethical practice of their contempo-
raries, for the favored modifications (education for girls, freedom for
slaves). Perhaps, then, novelties occurring within the ethical project are
always of a special kind: ethical progress consists in recognizing true
factual statements—statements, for example, about what women do under
par tic u lar circumstances or about the attitudes of those treated as
disposable property; these statements are integrated with previously ac-
cepted ethical statements to yield new consequences. Appearances to the
contrary, there are no fundamental ethical innovations. No statements
involving ethical vocabulary are ever added, subtracted, or modified,
without making inferences from premises consisting either of previously
recognized ethical truths or of new factual truths; ethical concepts are
never introduced or refined. (These formulations rely on everyday un-
derstanding of the distinction between ethical and factual statements.
Many writers have recognized “mixed terms” (“thick ethical concepts”)
involving both a factual and an ethical component: concepts such as
cruel. In my usage, a concept with any ethical component is ethical.
Debates about borderline cases are irrelevant for present purposes. For
the issues about ethical truth and progress turn on examples where the
distinction between the ethical and the factual is quite uncontroversial:
it is a fact that educated women do certain things; “Slavery is wrong” is
an ethical statement.)
If the denial of fundamental innovations were sustained, realists
could dismiss a large part of the challenge outright, for there would be a
uniform account of progressive ethical transitions. Once the ethical
project was begun, each generation inherited a body of established ethi-
cal beliefs, indeed ethical truths, from which people drew false conse-
quences because the most general and basic ethical truths were con-
joined with incorrect factual statements. Wollstonecraft discerned the
truth about women’s behavior under conditions of education, and Wool-
man recognized the psychological effects (on slaves and slave owners
alike) of chattel slavery. They made ethical progress by replacing false-
hood with truth.
Realists would still have to explain how the entire project was started,
how, in the dim mists of the Paleolithic, small human groups arrived at
all the fundamental concepts and principles required to yield, with dif-
ferent factual premises, the divergent conclusions adopted by rival tradi-
tions at distinct epochs. Although nobody should deny the significance
of new factual knowledge in the progress of our ethical practices, it is
hard to view the proposed story as explaining the entire evolution of
ethics—can we really suppose the origins were ethically rich enough to
allow the explanation envisaged? In denying ethical novelty, the account
assumes a collection of elaborate ethical principles was not only avail-
able for Wollstonecraft and Woolman to work with, but present at the very
beginning. The history of Part I took the initial institution of socially
embedded normative guidance, undertaken in response to the fragile
and tense hominid society it transcended, to be relatively crude and
limited—to comprise principles about sharing and the like. To suppose
ideas about roles, expanded altruism, and the good life were already part
of the Pleistocene package strains credulity. Furthermore, at the dawn
of written history, societies made important transitions not easily assimi-
lated to the story: the refi ned conception of the individual figuring in
the reformulation of the lex talionis, the reshaping of the concept of
courage, the acceptance of the policy of forgiving your enemies.
Realists must allow occasions on which people persuade their con-

temporaries to accept ethical statements not yet adopted by those around
them (even though most efforts at persuasion are rejected). When such
episodes culminate in ethical progress, the thesis of external con-
straints takes the resultant practice to accord more closely with those
constraints. Articulating this in terms of acquiring truth, it is assumed
that true ethical statements replace previously held falsehoods (or, per-
haps, prior lack of commitment). On the more ambitious version of the
appeal to external constraints, the people who initiate the progressive
transition apprehend these truths. Twenty thousand years ago, some
people saw it was good to extend some of the protections accorded to
their bands to the neighbors; three thousand or so years ago, people ap-
prehended that it is wrong to kill the daughter of a man who has mur-
dered another man’s daughter; after Wollstonecraft, Britons accepted the
value of education for women; after Woolman, Americans gradually re-
jected slavery; and, in the last decades, many have learned that the
combination of sexes in loving relationships has no ethical significance.
Here are statements of the form “E is F,” where E is some event, state of
affairs, or pattern of conduct and F is some basic ethical property (good-
ness, for example) that historical agents are supposed, somehow, to have
grasped.
Suppose the basic ethical properties are simple and irreducible. Just
how did those who made the changes apprehend them? We do not know
exactly when the extension of ethical rules to the neighbors occurred,
but we can easily imagine how it might have happened. Members of a
band have ascertained facts about the resources commanded by those
across the river and envisaged opportunities for exchange. Sitting around
the campfire, they agree to make gestures of restraint and nonhostility
to instigate mutually profitable interactions. The practice succeeds and
is eventually accompanied by explicit declaration that, on the pertinent
occasions, peaceful conduct is required. Speculative though it is, this
sociopsychological account is more plausible than the suggestion that,
one day, one band member enjoys some experience of the rightness of a
pattern of behavior the group has not yet tried and that he communicates
his experience in some way, enabling them to share his recognition.
We know a lot more about Woolman’s journey to his judgment against
slavery—and people of my generation can examine their own changing
views about the ethical character of homosexual relationships. Wool-

man’s dedicated efforts to persuade others to share his convictions about
slavery offer all sorts of considerations that integrate factual information
with prevalent ethical precepts, but his writings never indicate how read-
ers might put themselves into a position to apprehend the simple and
irreducible wrongness of owning other people. If he had some episode of
apprehension, he hid the character of his revelation. Unless other people
are different from me and those contemporaries I know well, shifting
ideas about same-sex relationships are products of wider knowledge
about varieties of human love and of extensive conversations with people
who live in different ways. Factual knowledge and social exchange are
prominent in this important ethical advance, but I find no place for ap-
prehension that same-sex love has a simple and irreducible property
with which the conventional wisdom of the 1950s failed to credit it.
Versions of realism holding ethical properties to be complex and re-
ducible look more promising. Consider the suggestion that one impor-
tant constituent of rightness is acting to diminish suffering and that an-
other is promoting unsatisfied desires. Realists adopting this suggestion
can view ethical innovators as appreciating facts about suffering and the
frustration of human aspirations—Woolman recognizes, in perfectly
straightforward ways, that slaves are subjected to all kinds of pain, and
Wollstonecraft understands that her own desires, as well as the wishes
of women she knows, are thwarted. Perhaps those who imagine peaceful
exchanges with neighboring bands foresee extended possibilities of meet-
ing the needs of the groups. How do they use their perceptions to justify
judgments about the wrongness of slavery, the rightness of educating
girls, or the goodness of extending the protections applying within the
local group? Neither acting to relieve suffering nor satisfying others’ de-
sires is sufficient for doing what is right. The ethical codes within which
the innovators operate deny that the sufferings of slaves are always wrongly
inflicted, view women’s desires for education as problematic and mis-
guided, see the wishes of the neighbors across the river as irrelevant to
goodness. Behind these conventional judgments stand other views
about what is gained, achieved in the way of goodness or rightness, by
allowing the sufferings, quashing the desires, or treating the neighbors
with hostility. Woolman, Wollstonecraft, and my hypothetical advocate
of trade question those views, effectively because they do not find any-
thing strong enough to counterbalance the perceived pain relief and de-
sire satisfaction.
Do the pioneers or their interlocutors have any articulated account of
the good and the right, clearly held in view, an account showing the
exact circumstances under which desire satisfaction and relieving
pain contribute to goodness or rightness? Apparently not. Woolman
and Wollstonecraft are aware of contentions about valuable goals achieved
by permitting the sufferings of slavery and not responding to female de-
mands for education—and they actively and specifically oppose those
contentions. They express their own assessment of what is gained and
what is lost, but they do not operate with any independent understand-
ing of goodness and rightness beyond those available in the formulations
of their opponents. They are catalysts of a renewed social exchange, a
conversation of the sort that initiated the ethical project, and the proper-
ties of goodness or rightness can be seen as fixed through such exchanges.
It is even harder to imagine how the early champion of neighborly exchange
could appeal, either in his own thinking or in his efforts to persuade, to
some reduction of goodness or rightness sanctioning his proposed mod-
ification of the code.
The point can be illuminated by focusing on another example. The
modification of the lex talionis relieves one form of pain but substitutes
another of the same kind: in either version, someone’s life will be trun-
cated. Perhaps that fact is itself ethically problematic, but the substitu-
tion of perpetrator for daughter surely looks like ethical progress. We
know nothing about the people who proposed and argued for the change.
Yet, on the realist account, there must be some connection between facts
about suffering and mortality, and ethical properties, that differentiates
the cases in which daughter and perpetrator suffer and die. If the inno-
vators and those whom they convince are to apprehend the rightness of
making the transition, they have to recognize that connection. How? We
can dimly recognize the form of the explanation realists want, but we
have no idea about how to give it substance. By contrast, it is easy to sup-
pose that the ancient societies in which the transition occurred were en-
gaged in frequent debate, that many voices participated, and that eventu-
ally continuing the old practice became socially problematic. Conversation
within a social group replaces the nebulous contact with some external
standard.
The realism just considered supposes some externally fi xed connec-
tion between the sorts of things ethical innovators apprehend—things
the examples reveal them as recognizing in perfectly straightforward
ways—and fundamental ethical properties. The challenge is to say what
the connection is, to explain how it could be apprehended, and to sup-
port the hypothesis of actual (albeit dim) apprehension. I now turn to
what seems the most promising realist position.
According to that position, ethical properties are conceived by anal-
ogy with colors, and colors identified with dispositional properties of
objects: redness, for example, is a disposition to cause us to enter certain
neuropsychological states (triggered by the impinging on our retinas of
light of particular wavelengths).22 Goodness and badness, rightness and
wrongness, apply to actions (say) in virtue of the tendency of those ac-
tions to generate reactive emotions, feelings of approbation and repug-
nance, for example. A great advantage of this account is its license of
justified ethical assessment without any complicated cognition (backed
by some unspecified process). You see the torturing of the cat, you feel
the repugnance, and your reaction both prompts and justifies your ethi-
cal judgment.
The reactive emotions that figure in this proposal are not merely af-
fective states available to be triggered across a wide range of social envi-
ronments. If they were, it would be impossible to account for the phe-
nomena we are attempting to understand, cases of progressive ethical
change (where in similar environments people react quite differently to
the same events). Instead, the reactive emotions—approbation, gratitude,
resentment, repugnance—involve cognitive and volitional states con-
22. Working out the details of any such account of color properties is itself a large task,
but I shall simply suppose that it can be done. Many contemporary thinkers have been at-
tracted to the thesis that ethical properties of things, states of affairs, actions, and patterns
of behavior can be treated analogously. The version offered by John McDowell (“Values
and Secondary Qualities,” in his Mind, Value, and Reality [Cambridge, MA: Harvard
University Press, 1998]) is, in my judgment, the best, precisely because it is sensitive to the
thought that the responses in the individual are shaped by the culture in which he or she
grows.
nected, in ways nobody yet knows how to specify (§4), with affective
states. What reactive emotions are available to an individual, the extent
to which a person is sensitive to such emotions, and the entities trigger-
ing the emotions are all subject to social shaping (§4).
It is important to appreciate this triple dependence on the environ-
mental conditions, including the social environments, in which we find
ourselves. Features of the ambient environment at the time at which we
encounter some occurrence affect whether we have an emotional reaction
to it, and what form that reaction takes. Features of the developmental en-
vironment, the surroundings in which we learn and grow, shape the ways
in which our emotional reactions are directed. Even more fundamentally,
that developmental environment affects our emotional repertoire. For, al-
though there may be some physiological responses relatively invariant
across regimes of socialization, the emotions pertinent to ethical assess-
ment are more complex than any such affective reactions, having socially
shaped cognitive components. No realist approach to ethical properties
that ignores these three modes of environmental influence can be ade-
quate, for the simple reason that certain types of environments can pro-
duce reactions strikingly at variance with one another. Realists need a
distinction between types of environments, proposing that particular re-
actions of people who have been socialized in normal developmental en-
vironments, and who find themselves in normal ambient environments,
signal the goodness (say) of the states of affairs to which they respond.
A full version of this form of realism has to suppose an external stan-
dard fixing some environments as the pertinent class in which specified
reactive emotions suffice for specified ethical properties—the fact that in
some situations (death camps, the near-starvation conditions of the Ik)
people fail to respond to acts of cruelty does not affect the ethical proper-
ties of those acts. Let us concede that the specification can be provided,
and the realist can even explain why the “normal” environments are
privileged.
Consider, now, the ethical innovator. At last we appear to have an ar-
ticulated, and convincing, account of how that innovator responds to
the external constraints on ethics. An ordinary experience produces a
strong reactive emotion, say, a violent feeling of repugnance against
some form of behavior tolerated by the surrounding community. The
innovator is prompted to a judgment of the wrongness of the conduct,

and the reactive emotion provides support for the assessment.
Yet surely we—and the innovator—should hesitate. Reflective people
know they can feel different, quite complex, emotions toward many types
of events and actions; they know that their background moods, their idio-
syncratic affections, and any number of other contextual features can
shape their feelings. Faced with the indifference of those around them,
they should reasonably wonder about their own reactions. Are they ex-
pressing personal idiosyncrasies? Is something awry with the environ-
ment in which they find themselves (is it “abnormal” in whatever sense
the realist has provided for that term)? Is the emotion what they take it to
be, ethical repugnance rather than some variety of nonethical disgust, or
even oversensitive distaste? There is nothing “inner” to which they can
point to settle these doubts. Hence, they must either have some more
explicit knowledge enabling them to convince themselves that the senti-
ment is a genuinely ethical one, or they must enter into dialogue with
their fellows, attempting to explain just how and why they are moved
as they are.
Woolman was a tireless campaigner for his abolitionist views, a man
whose “unease” only deepened as he debated with people who dis-
agreed with him. His initial response, however, was appropriately
modest—he completed the bill of sale. Despite his unhappiness with
the transaction, he surely worried that he was being “over-nice.” Who
was he, after all, to question practices other godly men and women, the
Friends among whom he lived, accepted without demur? Only the persis-
tence of the feelings, together with the extensive confrontation with many
alternative points of view, intensified his conviction that slavery was
wrong. If we view him as justified, it is because he comes to recognize his
emotional reaction is not transitory—it cannot be displaced by the most
severe attempts to revise some of the associated cognitions and volitions.
It is the social exchange, not any awareness of an external standard en-
dorsing his feelings, that gives Woolman whatever ethical insight he has.
There are gaps between unrelieved suffering and wrongness, between
thwarted desires and wrongness, and between negative emotions and
wrongness. The commitment to external constraints requires those gaps
to be filled with a standard, obtaining independently of the individual
and of society, which justifies the move from perception of suffering,

or of frustrated desire, or the feeling of the emotion, to a judgment of
wrongness. Innovators appropriately worry whether their willingness to
make this move is idiosyncratic. If they are justifiably to reach the con-
clusion, on the realist account, they must have some apprehension of the
standard. Not only do latter-day analysts struggle to explain just what
the standard is, but there is not a shred of evidence of historical recogni-
tion of it. As a result, realism is fundamentally flawed.
So, too, are the idealizing forms of constructivism, and for the same
reasons. Consider the individualist proposal attributing capacities for
moral reasoning independent of social environment (in some views, ab-
solutely independent of all experience). Let us (charitably) suppose the
processes giving rise to the distinctions the constructivist claims to draw
are thoroughly and precisely characterized. Not only are we told that
“wrong” applies to actions when the principles on which they rest can-
not be universalized, or that the word marks out actions that would be
rejected by people deliberating under ideal conditions, but we are given
a complete and precise account of universalization or of the ideal condi-
tions.23 Constructivists take the external constraints to which ethics is
to conform to consist in the privileged status of these procedures. When
ethical practice is properly pursued, people apply ethical predicates
in ways matching those sanctioned by the procedures. In the approach
currently under consideration, historical actors are assumed to appre-
hend the accord between their innovations and the deliverances of
the privileged processes.
23. The examples I offer here recapitulate the most famous versions of constructivist
approaches, the first evident in Kant, Groundwork of the Metaphysics of Morals (Mary
Gregor, trans.) (Cambridge: Cambridge University Press, 1988), and the second attribut-
able to Rawls, A Theory of Justice (Cambridge, MA: Harvard University Press, 1971), and
his successors [most obviously T. M. Scanlon, What We Owe to Each Other (Cambridge,
MA: Harvard University Press, 1998)]. I do not think any of these constructivists succeeds
in providing a completely clear and precise account of the processes envisaged—Kant’s
critics, for example, have rightly complained that his appeals to “contradictions” and his
implicit restrictions on the types of principles of action (maxims) permitted are vague and
loose. The challenge I am presenting, however, to integrate the account of ethical truth
with the evolution of the ethical project can be more accommodating and can concede to
constructivists successes so far not achieved.
Given the complexity of the special procedures, constructivists prob-

ably do not want to contend that innovators recognize all the details,
that they can formulate the correct account of the ethical vocabulary and
see clearly how it applies in the case at hand. At best, they have inklings
of the right version of constructivism and can apply some rough ap-
proximation to the tests it supplies—they cannot universalize in the
strict sense of the theory, but they can ask, “What if everyone acted like
that?” Not even this much can be traced in the historical record. We
might attribute some inchoate universalizing thought to Wollstonecraft
or to Woolman, but it does not figure in the works they write to persuade
others. The absence is comprehensible, for relatively imprecise thoughts
about universalization would have little force in the debates engaging
them, debates centered on allegedly important distinctions among kinds
of people (men and women, Africans and people of European descent).
Much more precise constructivist ideas, not inklings, would be required
to make any headway, and Wollstonecraft and Woolman do not have
those tools. For earlier transitions, attributions of incomplete apprehen-
sion of the constructivist’s preferred standard are less plausible, even
ludicrous. The unknown modifiers of the lex talionis and the hypotheti-
cal pioneers who proposed extending the code to allow for trade almost
certainly lack the concepts appeals to universalization require.
Like realists, constructivists face a more fundamental difficulty. Imag-
ine some hypothetical innovator, as insightful as you please, who tries to
follow your favorite constructivist procedure for applying ethical vocab-
ulary. On the basis of the attempt, she announces her proposed change
in ethical practice. It is clear to her, from the beginning, that others,
including those counted as authoritative within her community, dis-
agree with her. What should she make of that fact? One obvious self-
diagnosis is that, for all her efforts, she has failed to carry out the proce-
dure properly. To the extent she cannot articulate the grounds of her
judgment, discussing them with those around her who are reputed to be
wisest in ethical judgment, she should have doubts about her own com-
petence. If she could explain and convince, laying out the procedure
she has followed in detail, explaining its privilege, and challenging
others to conclude differently, she might have justified confidence in
what she has done. No historical actor shows even the slightest signs of
these abilities.
Constructivists often suppose individual “reason” can generate justi-

fied applications of ethical predicates, without attending to potential
worries that what has been done fails to get “reason” quite right (per-
haps even bungles badly). In circumstances of innovation, those worries
are especially apt. How could they properly be suppressed? The question
fails to arise because constructivists often attribute a special self-certifying
property to “reason” (it is a priori, luminous, or whatever). With respect
to mathematics, where people often, but by no means always, reach
agreement, that ascription might enjoy some small initial plausibility—
although even there it is, I believe, profoundly misguided.24 In the con-
text of ethics, where disagreements are rife, it has no plausibility, and in
situations where someone is challenging features of the prevalent code
it should surely be discarded.
To sum up the discussion so far: neither realist nor standard con-
structivist approaches to ethical truth can provide a satisfactory account
of progressive transitions in the evolution of ethics that will attribute
to innovators some apprehension of external constraints. Before we
take up the less ambitious idea of viewing historical actors as “sleep-
walkers,” whose fortuitous conformity to the constraints is appreciated
by their more enlightened successors, it is worth exploring another
position distinguished above. This position supposes ethical statements
are, properly speaking, neither true nor false. To attribute wrongness
to an action is not to ascribe some property to it, but to express an at-
titude. Some phi losophers, disconcerted by the “strangeness” of the
idea of preexistent divisions among acts and states of affairs whose
contours ethical truths limn, have claimed ethical statements lack
truth-values, taking them instead to express the emotional reactions of
those making them, coupled with an injunction to share the emotional
reactions.25
24. I defend this unpopu lar assessment in The Nature of Mathematical Knowledge (New
York: Oxford University Press, 1983), chaps. 1–4.
25. The very simplest versions emerge in the early writings of A. J. Ayer and Charles
Stevenson, although there are previous sources of inspiration in Hume. Both Ayer and
Stevenson articulate the position with increasing sophistication, and there are intricate
debates about the adequacy of the proposed semantic treatment for all contexts. A well-
developed version of the view is provided by Allan Gibbard, Wise Choices, Apt Feelings
(Cambridge, MA: Harvard University Press, 1990).
As characterized so far, noncognitivism (to use the standard name)

provides no account of ethical progress. Reactive emotions of different
sorts could arise in different societies, being directed toward differ-
ent actions and states of affairs, without there being any standard against
which the rival types of emotions, or the deeds triggering them, could be
assessed. Some versions of noncognitivism might abandon any search
for ethical progress. In exploring the challenge for the thesis of external
constraints, however, we should focus on positions seeking a distinction
among transitions. Those views take the ethical project to have made prog-
ress when specific types of emotions became available to us (we learned
to feel shame, guilt, or resentment) and when such emotions were gener-
ated by particu lar types of situations. The notion of ethical truth is
abandoned, but constraints remain. Some capacities for emotion and
some emotional reactions are apt; others are not.26
Any version of noncognitivism hoping to go in this direction, and to
avoid viewing ethical evolution as a matter of mere change, has further
work to do. It must specify what is meant by taking emotional responses
to acts and states of affairs to be improved; it needs a conception of
“emotional progress.”27 However any such conception is developed,
if it is to buttress the idea of historical actors apprehending external
constraints—presumably by feeling things they can recognize as “more
apt” than those of their fellows—it will face problems exactly parallel to
those discerned above. Appeals to emotional reactions work no better in
the context of noncognitivism than they did when ethical properties were
seen as dispositions to induce particular kinds of emotional responses.
The challenge defeats the thesis that ethics is a form of inquiry in
which innovators recognize and respond to external constraints. Re-
treat, then, to viewing the attainment of ethical novelty as sleepwalking.
Human beings, individually and collectively, stumble along, sometimes
responding to the difficulties of their social lives, sometimes feeling
26. Here I echo the language used by Gibbard, whose version of noncognitivism is the
best I know.
27. Many noncognitivist writings accept the relativist conclusion (or are uninterested in
resisting it). There are hints of a more progressive view in Gibbard’s Wise Choices, but, for
all its attention to many details of noncognitivism, I have not been able to draw from that
work any clear account of ethical progress.
confined by the ethical codes they inherit, and consequently modifying

those codes. Some innovations replace ethical falsehood with ethical
truth (or less apt emotional responses with reactions that are more apt).
Although the processes through which such novelties become incorpo-
rated in ethical practice are various and complex, typically having little
to do with any apprehension of ethical divisions among actions and among
states of affairs, there are occasions of progress, as communities align
their codes with external constraints.
How could anyone know? Some who come later, after the changes,
must be able to pronounce on what has occurred. Looking back, the
enlightened recognize ethical progress, perceiving new ethical truth in-
corporated into an ethical code. How do they make their judgments? If
their judgments are formed like those of the historical predecessors,
there is no reason to believe they are anything other than further ventures
in sleepwalking, providing no new knowledge. To avoid skepticism, the
defender of external constraints must show how the latecomers do better
than their predecessors, who, it is admitted, had no knowledge of novel
ethical truths. Analysts like us are in a better position to judge.
Many philosophers hold a parallel view of mathematics. Committed
to foundationalism, they propose basic principles, axioms, somehow ap-
prehended, from which mathematics proceeds by deductive arguments,
codified in proofs. One problem with foundationalism stems from the
elusiveness of the ways in which the basic truths are justifiably appre-
hended: most authors content themselves with a phrase or a label.28 A
second difficulty lies in the schism made between the “sleepwalkers” of
history and the enlightened folk of the present. Since the basic axioms
come late in the history of mathematics—the route to them is full of er-
roneous formulations—the great mathematicians of the past did not
apprehend them. Euclid, Descartes, Newton, Euler, Gauss, Weierstrass,
and company formulated mathematical truths but did not know them;
only in the twentieth century did mathematical knowledge become pos-
sible. Pragmatic naturalism rejects the odd combination of elusive
28. Most popu lar is “intuition,” a term Frege rightly saw as overused. It stands to Kant’s
credit that he attempted to give the notion of intuition some clear content. See Kitcher,
Nature of Mathematical Knowledge, chap. 3.
epistemology with a peculiarly self-confident attitude toward the present

and the recent past.
Mathematics, however, has the advantage of consensus. Scholars who
reflect on how mathematical knowledge is attained offer divergent ac-
counts but concur in supposing a body of knowledge, shared among math-
ematicians. Ethical theory resembles mathematics in lacking any convinc-
ing proposal for how anyone, past or present, has access to any external
constraints. At the same time, ethical practice lacks the power to achieve
consensus that supports the attribution of mathematical knowledge.
If, in the mathematical case, we should regard the practitioners of the
present as akin to the mathematicians who preceded them—rejecting a
division into “sleepwalkers” and “enlightened”—stronger reasons favor
a comparable assimilation with respect to ethics. What exactly has been
learned about the external constraints on ethical inquiry, constraints the
unwitting innovators could not apprehend? Ethical theorists often make
confident assertions about fundamental constraints. Moore declares it to
be evident there are only two types of intrinsic goods, human relation-
ships and beautiful things. Kant locates the moral law within us, identi-
fying it with Reason, which generates the Categorical Imperative. These
radically dissimilar suggestions would be defended along similar lines.
There are supposedly fundamental processes giving rise to the conclu-
sions defended, intuition of the good (Moore), and deliverances of a pri-
ori reason (Kant). If you worry about nebulous processes, incompletely
characterized, supposedly giving rise to knowledge of basic mathemati-
cal axioms, in a domain where people largely agree, you should be
profoundly skeptical about equally mysterious invocations generating
assertions of incompatible claims.
Was Moore a novel type of Homo ethicus, equipped for the first time
with an ethical analog of the microscope enabling him to discern previ-
ously unrecognizable constraints? Psychosocial explanation appears
more plausible. Moore’s judgments about beautiful things and human
relationships arise from a long psychological journey. The development
of his ideas began in the Victorian socialization of his childhood, wind-
ing through his school experiences and interactions at Cambridge, the
social and sexual entanglements of the elite university community of which
he was a prominent member, the chafing of the late Victorian ethos and
the various revolts against it. Detailed biographical research would ex-
plain Moore’s forthright dicta about value—and the attractions his views
held for his contemporaries.
Equally for Kant. His brilliant account of ethics in terms of Pure Prac-
tical Reason can be understood as a reaction to the moral law as com-
manded from on high. Attracted by ideals of autonomy and repelled by
the thought of subjection to a divine commander, Kant solves his con-
flict with a perennially popular account of ethics by placing the lawgiver
within, elaborating his revisionary theory in a framework of powerful—
but purely hypothetical—faculties. Kant can be assimilated to the great
innovators in the history of the ethical project. He responded to the state
of ethical practice he found in the Prussia of his day in an especially in-
genious way, but there is no reason to think he transcends the predica-
ment of his fellow pioneers.
Theorizing about the ethical project has been hampered by assuming
there must be some authority in ethics, some point of view from which
truth can be reliably discerned. Philosophers have cast themselves as
enlightened replacements for the religious teachers who previously pre-
tended to insight. But why credit any individual participant in the project
with such special authority? Ethics may simply be something we work
out together (Chapters 6 and 9). If latecomers improve on previous efforts,
their success might be based, not in occupying some privileged epis-
temic vantage point, but on study of the history of the project as it has led
up to the practices of their day.29
We can now see why attacks on the idea of the divine commander fail
to dislodge the idea (§27). The alternatives seem so mysterious. Ironically,
supposing ethical truths are claims a deity endorses and reveals does
29. Dewey is optimistic that ethics can fi nd its own analog of scientific method [see, for
example, chaps. 10 and 11 of Dewey, The Quest for Certainty, vol. 4 of John Dewey: The
Later Works (Carbondale: Southern Illinois University Press, 1984)]. I share the hope that
a properly informed understanding of the ethical project and its evolution might lead our
successors to pursue it more sure-footedly. That will depend, however, on more systematic
studies of changes in ethical practice (whose absence I lamented—§28). It would be foolish
to suppose the historical cases inadequately sketched in Part II can do anything more than
reorient discussions of ethics so as to prepare eventually for synthetic understanding and
improved ethical practice.
better at integrating the thesis of external constraints with the evolution

of the ethical project than any sophisticated secular rival. Adherents of
divine-command theories suppose that much, if not all, of the correct
ethical system has been disclosed at some point in the past, and even
that many apparent innovators receive special messages from on high.
They can resist the mere-change view.
Unfortunately, these theories collapse under Plato’s objection (§27)
and, more fundamentally, because of the demise of the principal charac-
ter. Is there any remaining option for resisting the evolution of ethics
as mere change?
chapter 6
Possibilities of Progress
§32. The Centrality of Ethical Progress

It is tempting to think resisting the mere-change view requires making
sense of the concept of ethical truth. Confronted with the deviant prac-
tices of others, including people who came before us, we want to distin-
guish our ethical principles from theirs: so we must claim ours are true,
theirs false. Hence the troubles elaborated in the last chapter appear to
leave no options: even though the mere-change view conflicts with spon-
taneous judgments about historical episodes (Chapter 4), we have to ac-
quiesce in it.
Truth is readily seen as prior to other notions used to explain the ob-
jectivity of our practices, concepts like progress, justification, and
knowledge. To make progress is to accumulate truth, to be justified is to
proceed in ways reliably generating true beliefs, to know is to have a
true belief generated by a reliable process.1 In this way of relating the
1. Here I follow the reliabilism pioneered within epistemology by Alvin Goldman, Epis-
temology and Cognition (Cambridge, MA: Harvard University Press, 1987). There are well-
known problems of saying exactly what kinds of reliability are at stake in securing justifica-
tion or knowledge; these are lucidly presented by Robert Brandom, Articulating Reasons
(Cambridge, MA: Harvard University Press, 2000, chap. 3); solving these problems
209
concepts, an adequate response to the mere-change view has to start

with ethical truth—and we are returned to the array of options of the last
chapter. We could escape by permuting some elements in the picture.
Here is a proposal: ethical progress is prior to ethical truth, and truth is
what you get by making progressive steps (truth is attained in the limit of
progressive transitions; truth “happens to an idea”).2 Pragmatic natural-
ism retains a notion of ethical truth for expository purposes, but it starts
from the concept of ethical progress.
The demand for some idea of objectivity—for talking of ethical truth,
ethical progress, or ethical knowledge—arises when we consider alter-
natives to our own ethical codes. Some of these are codes adopted in the
past, others are codes different people follow in the present, yet others
are codes we envisage ourselves coming to follow. In all instances, the
practical question concerns the possibility of amendment. The mere-
change view finds no basis for choice: there is no sense in which substi-
tuting one code for another would be objectively better or worse. An ex-
treme version of objectivism would declare that, for any two genuinely
distinct codes, either one is objectively better than the other, or there is a
third code, constructible from elements of the two rivals, objectively bet-
ter than both. Pluralism, as I conceive it, proposes there are (1) some
pairs of codes for which one is objectively better than the other; (2)
some pairs of codes for which neither is objectively better than the other,
but there is a third code constructible from elements of both, objectively
better than either; and (3) some pairs of codes for which neither (1) nor
(2) obtains. To salvage the notion of “objectively better than” that occurs in
these claims and counterclaims, we do not need any concept of ethical truth.
It is enough to recognize which kinds of changes would be progressive or
regressive. Opposition to the mere-change view can be formulated by
requires pragmatic judgments, dependent on the values endorsed. Those problems will
not be addressed here, since the task is to understand how appeals to truth and justification
in ethics might be reconceived in terms of the notion of ethical progress.
2. The parenthetical characterizations connect the permuted picture with the pragma-
tist tradition, the first with Peirce, C.S. Peirce “The Fixation of Belief” in Nathan Houser
and Christian Kloesel, eds., The Essential Peirce, vol. 1 (Bloomington: University of Indi-
ana Press, 2009) 109–123 the second with James, William James The Meaning of Truth
(Cambridge, MA: Harvard University Press, 1978), 169.
Possibilities of Progress 211
saying there are pairs of codes such that the transition in one direction is
progressive (and, in the opposite direction, correspondingly regressive).
Looking at the codes of the past, we are concerned to evaluate them
because of the possibility of going back. Fundamental to our assessment
is the question whether a particular transition, one reversing the actual
course of history, would be progressive. Faced with the codes of actual
rival people, we are interested in possibilities of shifting to their commit-
ments, or of constructing from the elements of their current code and the
elements of ours a new amalgam. We ask whether the change to their code,
or to the amalgam, would be progressive. Similarly for codes we envis-
age. What concerns us is the possibility of a progressive transition.
In fact, it is better to approach these decisions by starting with the no-
tion of progress rather than that of truth. Thinking in terms of truth
narrows the focus. For truth applies to statements, so we are led to con-
ceive the decision as one about descriptive counterparts of rules of the
alternative code. There are other components of ethical codes—concepts,
exemplars, habits, emotions, modes of inducing compliance—and im-
provements to our own practice could occur in each of these respects. A
rival code whose rules agree with ours might do substantially better
at preventing relevant forms of blindness. Thinking in terms of progress
responds more directly to the practical choices we face.
Attending to contexts of comparison, in which issues of progress
arise, poses an obvious question. Is it right to take the overall progres-
siveness of an ethical transition as fundamental, or should we think in-
stead of changes as progressive in particular respects? Historical changes
often seem to involve losses as well as gains. Rejecting the glorification
of the Homeric hero in favor of social solidarity impresses us overall as
progressive, but we may want to concede something to Nietzsche’s nos-
talgia for lost creativity, daring, and freedom. On balance we can talk of a
progressive transition, but that is because we can weigh the relative impor-
tance of the aspects with respect to which progress occurs. So, it might
be suggested, the fundamental notion is that of a transition progressive
in certain respects: judgments of overall progress are made by adding the
weights of the progressive respects and subtracting the weights of the re-
gressive respects. Sometimes the sum can be done, and we can talk con-
fidently of a progressive (or a regressive) transition, but there will be
occasions on which no common scale for comparison can be found,

when no determinate weights can be assigned, and no overall judgment
can be reached.
We should reject this universal atomism. Although it may sometimes
be a useful strategy to decompose an overall judgment of progressive-
ness by identifying advantages and disadvantages of an ethical change
(and showing the former outweigh the latter), the thought that all overall
judgments are built up out of atomic judgments involving no commit-
ment to weighing gains and losses is suspect. With respect to any change
whatsoever, are there really atoms, possessing only advantages or disad-
vantages? Human behavior is so multifaceted that even if we focus on
one aspect of the change—the bonding together of the citizens of the po-
lis, or the increased freedom felt by women who see new doors open for
them—are these facets uncontroversially positive in all respects? Per-
haps they bring relatively modest problems in their train, or at least in-
creased risks of small disadvantages (the anxieties of newfound freedom,
comrades’ blindness to outsiders), slight, to be sure, in contrast to the
large positive gains. If so, further decomposition is needed to find the
genuine atoms, whose weights will figure in the eventual sum. However
far we dissect, new instances of the same phenomenon are likely to arise.
Overall assessment of progressiveness is essential to ethical judgment.
The basic concept is that of a progressive transition (period). Yet we can
allow for dissection of transitions to proceed where, and to the extent, it
is useful—for there will be cases of overall judgment in which a consider-
ation of respects, without commitment to the idea that they are “atomic,”
clarifies the basis for judgment.
So far, I have tried to show how appraisals of progressiveness could
do the work often achieved by appeal to truth. Issues relating to justifica-
tion have been ignored. That is because the need for objectivity is felt pri-
marily in contexts of live choice, when we wonder whether an envisaged
transition would be worth making. On the picture of truth as funda-
mental we ask whether we would be gaining new ethical truth and elimi-
nating old ethical falsehood; in my replacement view, we are concerned
with the progressiveness of the transition. If and when we need a notion
of ethical justification, it is easily found: people are justified when their
decisions are generated by processes likely to yield progressive changes.
Reliability in the production of ethical truth gives way to reliability in

the genesis of progressive transitions.3
§33. Generalizations from History

Part I portrayed the evolution of ethics as driven by forces of selection.
Darwinian considerations figured in the prehistory of ethics (the emer-
gence of the preconditions for normative guidance and experiments of
living), possibly retaining a role in subsequent stages of cultural compe-
tition (as when experiments lead groups who practice them to wither
and go extinct); they were clearly present indirectly when ethical codes
became attractive to other groups because of biological benefits they ap-
peared to confer (greater survivorship of young members, for example).
Forces of cultural selection, dependent on the attractiveness of particu-
lar ethical ideas and thus answering to human desires, desires possibly
independent of Darwinian advantages (§16), have also shaped the evolu-
tion of the ethical project. Friends of ethical truth find no reason to think
forces like these are likely to generate true ethical beliefs; by the same
token, nobody should suppose them conducive to ethical progress. That
means, however, only that progressive transitions are not to be identified
with those promoting Darwinian or cultural success, a point simply
recapitulating the unsteadiness of ethical progress—as well as the failure
of crude evolutionary reductionism.
What could ethical progress be, if not accumulation of ethical truth?
An obvious (empiricist) strategy for finding an answer is to start with the
motivating examples of Chapter 4 and examine what, if anything, they
have in common.4
Suppose we were to have a comprehensive vision of the progressive
transitions occurring in the evolution of ethics, from the beginnings of
socially embedded normative guidance to the present. Merely listing all
3. Issues about justification and knowledge, and how these notions apply to historical
actors, will occupy us briefly at the end of this chapter.
4. This is the empiricist strategy pursued by Hume in his second Enquiry [Enquiry
Concerning the Principles of Morals (Indianapolis, IN: Hackett, 1986)], a strategy in accor-
dance with his understanding of Newtonian method. (On this topic, I have been helped by
Matthew Jones.)
these transitions and then declaring any ethical change progressive just
in case it occurs on the list would plainly be unsatisfactory. That would
fail to allow for future progressive transitions. More important, the skep-
tical challenge (§28) would arise again, in the pointed form of asking
why anyone should care about ethical progress, if it is just a matter of
making one of the officially designated transitions. We want an account
of ethical progress isolating common features of the favored transitions,
features revealing why we might be concerned to make ethical changes
of this particular kind.
Chapters 2 and 3 offer only an infinitesimal fragment of the envisaged
list. Yet their constituent examples are often in line with the most popu-
lar idea about ethical progress: insofar as they have attended to ethical
progress, historians and philosophers have singled out a particular kind
of movement as constitutive of advances. They have talked of “the circle
expanding” or “the expanding circle.”5
The proposal is motivated by a striking feature of some progressive
transitions in the history (and prehistory) of ethical practices. In the rel-
atively recent examples of the abolition of slavery and the recognition of
women as equal participants in public life, individuals who had previ-
ously not been brought within the purview and protections of ethical
precepts routinely applied to others (white men) became recognized as
full people, as proper subjects for the application of principles hitherto
applied on a restricted scale. Slave owners were committed to precepts
forbidding certain kinds of behavior toward full people: full people were
not to be permanently separated from their families; they were not to be
bought and sold. Before the ethical change, black men and women did
not count as full people; after it they did, and old proscriptions now
applied to them too.
Earlier examples reinforce the idea. A prominent Gospel message de-
nies that boundaries of aid and compassion embrace only the local group.
Even in prehistory, interaction with other groups, temporary fusion of
5. The former phrase is used by W. E. H. Lecky in his History of European Morals from
Augustus to Charlemagne (New York: Braziller, 1955); the latter is the title of Peter Singer’s
book-length attempt to identify ethical progress, The Expanding Circle (New York: Farrar,
Straus, Giroux, 1981).
smaller bands into larger units, negotiation, and trade become possible
when some rules were extended to cover outsiders.
Reflections of this kind inspire a straightforward theory about ethical
progress. Ethical codes begin by increasing the scope of altruism within
a group (enlarging the set of contexts across which members are pre-
pared to respond altruistically; §5); they progress by further expansions
of scope and by expansions of range (the set of individuals toward whom
altruistic responses are made; §5). Progressive transitions occur exactly
when the modified code contains precepts enjoining altruism of wider
scope or greater range than the code it replaces.6
There are difficulties. Not all our paradigms of progressive ethical
change fall under the envisaged rubric. Consider the modification of the
lex talionis. Here there is a shift from inflicting suffering on individuals
specially related to the perpetrator to punishing the person who did the
deed. The transition does not begin with a class of people initially pro-
tected by an ethical precept and another class of people not so protected,
and transfer some of the latter individuals to the protected class. Initially,
anyone is vulnerable to harm, provided he or she stands in a particular
relationship to the crime—being the son or daughter of someone who
killed the son or daughter of another person; after the transition, anyone
is vulnerable to harm if he or she stands in a different particular relation
to the crime (being the doer of the deed). No circle is expanded; one
circle is replaced by another.
Sometimes, what changes is not the group attitude toward a particu-
lar individual or class of individuals, but the group attitude toward the
desires those individuals have. Consider the withering of vice (§26). With
some strain, you could say there are two classes of people, those with pref-
erences solely for sexual activity with members of the opposite sex and
those who sometimes want to engage in intercourse with a member of the
same sex; before the change, there is toleration (an altruistic response?)
of the desires of the former class, but toleration is not extended to similar
desires of the latter class; after the transition, toleration extends to both.
6. As I read Singer, he is attracted by a theory of this sort (although possibly not one
so extremely rudimentary); at an early stage of my own thinking about ethical progress, I
was too.
Quite apart from the issue of whether talk of altruism is really illuminat-
ing (or appropriate) here, the crucial step consists not in gathering per-
sons under a protective umbrella from which they were previously ex-
cluded, but in recognizing their desires as worthy of expression.
Inclusion and exclusion apply to desires, not people.7
Attention to some changes that must have occurred in prehistory re-
inforces these conclusions. During the Paleolithic, our human ancestors
modified their ethical codes to require individuals to use their talents for
the benefit of the group, they introduced roles and role-specific prescrip-
tions, they added requirements for development of talent, they came to
appreciate the higher forms of altruism that play a role in human rela-
tionships, and, in doing all these things, they developed a richer concep-
tion of human life and of the human individual.8 Among the members of
the earliest societies to have left written records are at least some who
have a conception of what it is to live well, far more elaborate than any
present in the groups that began the ethical project: for well-born Baby-
lonians and Egyptians, a life satisfying the basic desires felt by their dis-
tant ancestors would not be good enough.9 To be sure, aspects of their
understanding of what is good constitute no advance at all (their eager
acquisition of gaudy luxuries); others are simply regressive (desires to be
7. Defenders of the “expanding-circle” proposal might reply that they allow for exten-
sion of scope as well as of range. Yet the example is importantly different from central in-
stances in which an ethical advance consists of increasing the collection of contexts in
which people are willing to help one another (for example, by commanding the response,
even when the costs for agents would previously have inhibited addressing the other’s
plight). Failure of response arises not from the burdens of altruism but from the character
of the desire. Precisely because of what the other person wants—sexual relations with an-
other man or another woman—there is condemnation, assault, punishment, even murder.
Progress does not consist in expanding any circle but in recognizing facts that normalize
desires.
8. In citing these changes I do not presuppose my “how possibly” explanations of them,
but simply that these transitions occurred. That is a simple consequence of the difference
between the initial stages of normative guidance and the practices present at the dawn of
history.
9. Whether it would be enough for the distant ancestors themselves is a question to be
considered. One might well attribute to them desires for cooperation and for making con-
tributions to joint projects. If so, the seeds of a richer notion of living well are already
present.
viewed as superior). Important parts of their conceptions, however—the

wish for relationships marked by higher-order altruism, the wish for a
life of ser vice to the community—mark genuine ethical progress. In
these respects, what has evolved is an advanced appreciation of human
possibilities.
The experiments of living of our prehistoric ancestors generated
an important mode of progress that fits the “expanding circle” account,
in extending their prescriptions beyond the small bands of the fi rst
human beings, as well as increasing the scope of (behavioral) altruism.
Besides that, they improved techniques of securing compliance (in large
measure by inventing the unseen enforcer; §17). They also changed radi-
cally the framework in which altruism, and its failures, could occur, by
enlarging the repertoire of human desires. Part of the story of ethical
progress must consist in understanding how acquiring new desires, not
merely satisfying them, counts as progressive.
My formulations contain important modifying words: some people in
Near Eastern societies had more elaborate views of the good human life,
and some desires in the enlarged repertoire represent real advances.
These words signal the fact that the lineages extending from the first small
bands of ethicists to the far larger and more complex cities of Egypt and
Mesopotamia made ethical losses as well as ethical gains. One obvious
loss consists in replacing an initially egalitarian society with something
highly pyramidal. That shift renders many members of the society in-
visible to others—and dramatically expands the class of altruism fail-
ures. For many of those acquiring the richer conception of the good life,
the invisible people no longer figure as potential targets of altruistic re-
sponse. Another loss, interconnected with the development of larger,
hierarchical societies, is the shift from a situation in which all are in-
volved in socially embedded normative guidance to one in which rules
are set by particular figures (typically representatives of the unseen en-
forcer). These are the principal loci, at which it becomes attractive to
talk of ethical progress in particular respects, not overall progress. The
significance of a mixture of gains and losses will be apparent in later
chapters.
There are other modes of ethical progress besides “expanding the
circle.” The next sections offer a more inclusive account.
§34. Problems, Functions, and Progress

One way to break the stranglehold of the idea that progress consists in
accumulation of truth is to consider an area in which advances are under-
stood differently. Technology serves as a paradigm. Our world is full of
instruments, machines, and devices that improve on previous efforts.
The chair in which I sit, the light illuminating my desk, and the com-
puter on which I type these words are all refinements of similar things I
used a decade ago, and spectacular advances on things my ancestors
employed to similar ends.
Progress with respect to these artifacts, and in the domain of technol-
ogy generally, is readily understood as functional refinement. We start
with a function to be fulfilled and an initial device that does the job. From
first success descends a sequence of improvements, things performing the
task more reliably, more quickly, more cheaply, and with less demands on
the user. Decreasing the error rates is important because those who use
the devices usually want the job done correctly on a series of occasions.
Speed, cost, and ease of use are worth having because we have other jobs
to do, and these compete for our time, our resources, and our energy.
More reliable, faster, cheaper, and less effortful performances are better,
contributing to functional refinement and to progress.10
Behind the original device stands a person and a problem, or a person
with a problem—usually a class of people with one or more problems.
The people want to solve the problem, and that wish grounds the func-
tion of the successful device. A particular thing, introduced into a con-
text, has the function F in that context if and only if that thing is present
because someone wanted something to do F and the thing was introduced
into that context to satisfy the wish.11 Functional refinement consists in
10. In many cases, people have to make decisions about what factors are most important
to them, trading reliability against cost, or speed against ease of use. Here, as in the ethical
case, we can recognize progress in particu lar respects. Sometimes it is possible to talk of
overall progress, but sometimes not.
11. Here I adapt ideas from a seminal discussion by Larry Wright (“Functions,” Philo-
sophical Review 82 [1973]: 139–68). There are long and intricate discussions of how to elabo-
rate Wright’s “etiological” approach to functions, but details are not pertinent to present
purposes. The version of the account I favor, and that underlies my general claims about
satisfying the wish more reliably or more completely, and doing so in

ways that generate fewer problems for potential users (that is, allowing
satisfaction of their other desires to proceed more freely).
If a device is complex, its function frequently generates functions for
its parts. One function of my desk chair is to enable me to sit at my desk
for long periods without back pain; the slide letting the surface against
which I lean move up and down has the function of enabling me to di-
rect the support where I most need it. In general, functions of parts are
determined in virtue of their causal contributions to the function of
the whole.12
Attribution of functions is straightforward where there are clear-
sighted potential users who can express their desires and identify their
problems. On other occasions, we can talk of functions even though
cognitive beings with desires and needs are not involved. Biologists and
physicians routinely discuss the functions of organs, bodily systems, cells,
and molecules. They identify causal processes contributing to the sur-
vival and reproduction of the organisms in which the organic parts are
found. An animal has teeth of a certain shape, and we recognize them as
having the function of grinding tough woody matter into digestible form.
That is to say, teeth of this shape have a causal power to break down the
exterior walls of the plants in the animal’s environment, modifying this
material so it can serve, given the background properties of the animal’s
digestive system, as nutrition, and hence contribute to the organism’s
persistence. Although there is no cognitive being whose wishes are ful-
filled by having animals survive—no benevolent creative deity, no
Mother Nature—there is an analog of the problem background figuring
in the technological case. Here the problem background is constituted
by the Darwinian forces shaping the characteristics of life and its his-
tory: at the basis of these forces lies the general problem of reproduc-
ing, from which results, at a slightly less general level, the problems of
functions here, is given in my essay “Function and Design,” Midwest Studies in Philosophy
18 (1993): 379–397.
12. This is an important insight of Robert Cummins (“Functional Analysis,” Journal of
Philosophy 72 [1975]: 741– 65). I amend Cummins’s original account by requiring that the
causal contribution relate to some function of the whole system (and thus there must be a
prior notion of overall function). For defense, see my “Function and Design.”
surviving, being fertile, and finding mates (for sexually reproducing

organisms).
Functional refinement comes with modifications of organic parts so
they are able to discharge the function more thoroughly, more reliably,
more quickly, or with less strain on other bodily systems. Teeth making
plant casings more digestible, less likely to leave indigestible lumps, de-
manding less chewing time, or diminishing the effort required to chew are
refined versions of teeth that came before. As in the original case of tech-
nological devices, some improvements are directly aimed at the job to be
done, while others respond to the ancillary conditions of user or organism.
Speed and cheapness are good because people have other things to do
with their time and money; speed and diminished effort are worth having
because organisms have other demands on their time and energy.
Physiologists routinely identify the functions of molecules without
knowing very much about the details of the evolutionary processes that
led to the presence of the molecules in the cells carrying them. They can
neglect history because they see the problem background for the organ-
isms in question and can pick out the problem-solving contributions the
molecules make. The problem background results from the most general
Darwinian pressures (the need to reproduce, the need to survive long
enough, the need to be fertile) in light of the organism’s prior constitu-
tion. Animals require food; so, if they are already committed to living—
locked in by their anatomy and physiology—in a particular environment
where only a certain food source is available, they need capacities to pro-
cess that food; if the digestive system is hard to modify, they will need
ways of transforming the raw material into something digestible, given
the present constitution of the stomach; hence the problem of breaking
down the casings, solved by the teeth; from this come attributions of
functions to molecules deployed in building teeth with the required prop-
erties. Reconstructing that problem structure, with its cascade of ever-
more specific difficulties to be overcome, is compatible with any number
of historical scenarios, most of which physiologists cannot formulate
and among which they do not have to choose.13
13. For more detail about the differences between this approach to functions and that
taken by etiological theories, see my “Function and Design.”
Between clear-headed recognition of problem structures and the bio-

logical cases, in which no cognitive subject who sees the problems and
designs the solutions is present, stand intermediate cases. On occasion,
people recognize difficulties—they know not all is well—even though
they cannot frame the troubles exactly. You feel twinges of discomfort
and sometimes pain when you perform particular motions; your doctor
formulates the problem precisely and prescribes a supportive device or a
program of exercises. In examples like this, we can distinguish the dif-
ferent perspectives of the problem facer and the problem solver, the for-
mer encountering the problem background and the latter recognizing its
structure. From one pole, at which problem facers know everything a
problem solver might grasp, there is a sequence of cases in which prob-
lem facers are increasingly incapable of seeing what is wrong, leading to
the opposite pole, biological examples in which problem facers lack any
cognitive grasp or are not even cognitive subjects at all.
Given this general view of functions and functional refinement, I pro-
pose that socially embedded normative guidance is a social technology
responding to the problem background confronting our first full human
ancestors. None of them had a clear understanding of that problem back-
ground. Moved by a sense of the fragilities and tensions of their social
life, they first guided their behavior by regularities to help them avoid
trouble and later discussed with one another rules to govern conduct, to
be applied in increasingly explicit systems of punishment. Crucially, the
problems arise not for a single individual, but for the social group. Each
member of this band is committed to a particular environment, in this
case, a social one: they have to hang together (§§9, 14). Each of them feels
the difficulties the circumstances of their shared life impose, the frequent
tensions, the long episodes of peacemaking. The problems are felt by all.
Ethical codes serve the function of solving the original difficulties, dimly
understood by these ancestors. Initially, they offer only partial amelio-
ration. Ethical progress consists in functional refinement, first aimed at
solving the original problems more thoroughly, more reliably, and with
less costly effort (substitutes for hours of peacemaking; §10). In the course
of progress, however, the problem background itself changes, generating
new functions for ethics to serve, and hence new modes of functional
refinement.
So far the schema of an account of ethical progress: one needing to be

filled in by specifying the original function ethics is to serve, by explain-
ing the ways in which functional refinement takes place, articulating the
idea of generation of new functions, and understanding those further
functions. Pragmatic naturalism aims to analyze the situation in ways the
original participants could not, to bring into focus the dimly perceived
problems our ancestors felt. The historical narrative of Part I provides the
materials for doing that. The remainder of this section explores the ques-
tion of original function.
The tensions and fragilities of hominid (and chimpanzee) social life
arise from the limited altruism of the participants. Altruism failures lead
to confl ict, to pain infl icted, to rough discipline, and lengthy peace-
making. To the extent altruism failures can be avoided, life goes more
smoothly, with increased opportunities for cooperation and, conse-
quently, greater mutual benefits. Group members satisfy more of their
desires and protest less. The first ethicists did not recognize themselves
as responding to the problem of altruism failure—they simply wanted
relief from social instability. All of them felt fear and dislike of the epi-
sodes of conflict; all wanted their interactions to go more smoothly.
From the analyst’s point of view, the members of these small societies
were aware only of the symptoms of their predicament. The underlying
disease was the prevalence of altruism failures. Remedying altruism fail-
ure is the original function of ethics.
To specify the function more exactly, we must examine cases of fail-
ure to respond to the desires of others. Suppose A fails to respond to the
desire he attributes to B; his solitary desire is unmodified. A potential rule
would direct him to act on the basis of a desire incorporating, at some
level of intensity, B’s preference. Imagine, for concreteness, the rule pre-
scribes golden-rule altruism (§5) for this context. Will this rule ameliorate
the difficulties our early humans experience? That depends on whether
promoting B’s desire in this context clears up B’s dissatisfaction, with-
out simultaneously generating trouble for other members of the group. If
B’s desire is to do something that would block the satisfaction of the de-
sires felt by B*, the rule may simply substitute one altruism failure (to-
ward B*) for another (toward B). Perhaps there is a cascade of effects on
the desires of other group members. The bump in the rug may be shifted
and even made larger through the effort.
In discussing the early phases of the ethical project (§19), I introduced

the concept of endorsing others’ desires. Desires are endorsable just in
case there are possible environments in which they could be satisfied for
all our fellows. So, for example, to remain close to the predicament of our
early human ancestors, the desire to have adequate food is endorsable,
whereas the desire to monopolize reproduction is not. With respect to
many desires, it makes sense to think of the extent to which, in the cur-
rent environment, the desire goes unsatisfied. Two members of the
group may each be hungry, even though the food available to one is less
than that available to the other. A desire will be said to be more urgent
just in case the shortfall in satisfying it is greater. One way of analyzing
the problem background faced by our ancestors is to see the function
of ethics as allowing a smoother, more peaceful, and more cooperative
social life, through remedying altruism failures, and, more specifically,
through clearing up those altruism failures involving the most urgent
endorsable desires on the part of the potential beneficiaries.14
Although early ventures in socially embedded normative guidance—
particularly the most successful ones—may have framed principles ac-
cording with this conception of the function of ethics, that was not our
ancestors’ perspective. They surely began with paradigms: desires for
food, for shelter, for protection against attacks. In responding to the most
urgent of these desires, they remedied a class of altruism failures and
reaped the benefits of decreased conflict. Retrospectively, we later ana-
lysts can view attention to cases of altruism failure where the desires are
endorsable and urgent as a preliminary way to specify the problem back-
ground, but it cannot be the final story. The original function of ethics is
to remedy those altruism failures provoking social conflict, and the prob-
lematic class is only partially specified by deploying notions of endors-
ability and urgency.
Consider another aspect of the lives of our early human ancestors, ag-
gressive behavior inhibiting another group member’s attempts to satisfy
desires. B attacks B*, while A looks on. As things stand, two altruism
14. To conceive the problem background in this way is to introduce familiar ideas from
ethical theories. The emphasis on endorsability has obvious connections to universalizing
principles and can be viewed as echoing a Kantian idea; the focus on the most urgent de-
sires has kinship with Rawls’s different principle; see also T. M. Scanlon, “Preference and
Urgency,” Journal of Philosophy 82 [1985]: 655– 69.
failures have occurred. One way for A to “remedy” the failures of altru-
ism would be for her to form a desire that responds positively to B’s
manifest wish—to join the attack on B*. Often that would be to intensify
social conflict. Some forms of positive responses to the desires of others
are contaminated because what the other wants to do is to initiate inter-
ference with the expression of the desires of a third party. Societies with
rules supporting such interference are likely to fan the flames of social
conflict; societies directing altruism away from contaminated desires, and
toward the wishes of the targets of the interference, will probably dimin-
ish the tensions. They will develop a system of punishment directing
retribution toward those who initiate aggression.
So we can indicate another class of altruism failures to be remedied,
those in which the desires of victims to be protected go unsupported. As
with the earlier effort at more exact characterization in terms of urgent,
endorsable desires, this too is incomplete. Societies that fasten on actions
initiating aggression, take altruistic responses to the desires those actions
express to be contaminated, and institute rules for remedying altruism
failures toward the targets of aggression make a large first step toward
solving the problem of social confl ict. There will, however, be occa-
sions for further refi nement. Excluding contaminated desires is not the
whole story.
The large hope of ethical systems is to formulate an overarching ac-
count of fundamental ethical properties (goodness, rightness) or a collec-
tion of basic principles sufficient to adjudicate all cases. Those who share
the hope will believe that the function of ethics can be specified exactly
by latecomers in the evolution of our ethical practices. To the extent they
accept the framework in which I have posed the issue, they will seek an
exact representation of the problem background and a precise character-
ization of the altruism failures to be remedied. I view the situation differ-
ently. We latecomers know more about the problem background than
the pioneers who developed ethical systems partially responsive to it: we
can see how rules directing altruistic responses toward urgent endors-
able desires and away from contaminated desires are likely to dimin-
ish social tensions and confl icts. Codes incorporating those rules have
taken a first step, but they leave room for functional refinement. Ex-
trapolating from whatever progress can be discerned in the past, we
envisage an indefinite sequence of adjustments dealing more thoroughly

with the original problem. Neither we nor our descendants are likely to
achieve a complete solution, one that will correspond to a complete and
exact characterization of the altruism failures to be remedied—or, corre-
spondingly, to a complete system of ethical principles. At its most progres-
sive, the evolution of ethics is (at least as far as its original function goes)
a series of responses to the most powerful sources of residual social
conflict.
Pragmatic naturalism makes the following claims about the original
function of ethics. (1) The problem background consists in social insta-
bility and conflict caused by altruism failures. (2) The original function
of ethics is to promote social harmony through the remedying of altru-
ism failures. (3) Our ethically pioneering ancestors had only a dim ap-
preciation of the problem background, responding to the difficulties and
discomforts of a tense and fragile social life. (4) We know more about the
problem background than they did and can offer partial and incomplete
diagnoses of the types of altruism failures to be remedied; we can under-
stand the success of certain kinds of ethical systems in terms of these
partial characterizations. (5) Even with respect to the original function,
the project of refining the codes we have continues.
I anticipate a family of obvious objections to this approach to the origi-
nal function of ethics, and the rest of this section is devoted to re-
sponding to them. The common strategy of all the worries is to specify
ways of discharging the function, ways that would elicit from many—
perhaps almost all—people the judgment that they are ethically repug-
nant or, at best, ethically indifferent. The simplest and crudest version
remarks that greater social harmony can be produced by brutal means.
We can envisage a society in which a powerful dictator physically pre-
vents conflict from ever occurring. As it stands, this fantasy does not
expose a real difficulty with the approach, since, although it may be con-
ceded that the envisaged society is unsatisfactory, it does not fulfi ll the
function identified in thesis 2, the remedying of altruism failures. To
use an obvious analogy, it palliates the symptoms without attending to
the underlying cause, as if the doctor were to offer to cure your aching knee
by amputating your leg. In situations where confl ict would previously
have arisen, nothing is done to modify the desires of the participants;
instead, they are simply prevented from expressing their—altruistically

failing—desires.
Let the dictator work differently. He lays down rules compelling
people to take the desires of others into account in their own conduct.
How does he achieve that? He makes the penalties for not responding to
the needs of others truly severe. If the modification of desire occurs
across the board—everyone responds to the perceived wishes of those
with whom he or she is in contact—it is not obvious that the dictator’s
creation of the system fails as an ethically progressive step. Behavioral
altruism replaces previously selfish conduct. The resultant system is a
variant on the sorts of successful ethical practices envisaged in the early
period of experimentation (§§11, 17): a system of rules produces behav-
ioral altruism in all members of the society, and following these rules
diminishes social tension; the main difference comes in the way that com-
pliance is achieved; the unseen enforcer is replaced by a looming secular
presence. Further progress would be made if a more intricate mix of
emotions—awe, respect, solidarity, and genuine sympathy with others—
superseded fear.
To generate trouble the dictator must proceed at cross-purposes with
the kinds of altruism failures we later analysts see as the ones to be rem-
edied. Perhaps he does that by instituting rules we take to be unfair:
some members of the society are designated as servants, others as mas-
ters; subject to the usual severe threats, the servants must respond to the
wishes (even the whims) of the masters. Assume the wishes of servants
go unsatisfied. If the class of servants is large, only continued repressive
actions will keep their frustrations in check: social conflict erupts in new
ways. Perhaps it is tightly restrained. Yet that requires coercive action,
and, once again, the dictator has failed to fulfi ll the function of reme-
dying altruism failures. Indeed, his procedures constitute a drastic modi-
fication of human life as lived by the original participants in the ethical
project, for he has suppressed some of the conditions under which the
admittedly limited human capacities for psychological altruism can be
expressed.
If the class of servants is small, however, the dictator has not remedied
altruism failures so much as created a vast class of them, to wit the fail-
ures of masters to respond to the desires of servants. Suppose, on the
other hand, the wishes of servants are satisfied, and this is done through
rules encouraging the masters to respond to those wishes. Now the dis-
tinction between masters and servants lacks the significance it appeared
to possess, and, again, we have a recognizable system of ethics with an
odd means of ensuring compliance.
Consider yet another scenario. Social harmony is achieved because,
in the cases that would have generated conflict, some parties just do not
want whatever it was that caused a quarrel. If we focus the idea by imag-
ining the original experiments of living, the previously contested items
will be basic necessities, and the fantasy consists in supposing that some
members of the group do not want the resources they need to keep them
going. The other members respond to their abstemious wishes, and they
wither and die. In this instance, however, although a certain class of
altruism failures may be remedied, an extreme type of altruism failure
remains—for those who profit from the resources their unfortunate com-
rades forego fail to take any steps in the direction of paternalism. As the
plight of the ascetics becomes ever more dire, identification with their
actual wishes, rather than with the wishes they would have had if they
clearly perceived the consequences for themselves, ceases to be a form
of altruism.
Imagine a last variation on the theme. Social tension within the group
is decreased by expelling some of the members.15 There are two main
possibilities. Perhaps those expelled are the primary causes of social trou-
ble: they sometimes benefit from the altruism of other group members
but consistently fail to make altruistic responses themselves. Ending the
practice of interacting with them genuinely would remedy the altruism
failures by which the band has been beset and so would fulfill what has
been identified as the original function. If these are the circumstances,
however, the account of the original function is not challenged, since
there is no basis for describing the envisaged expulsion as regressive—it
is of a piece with prevalent societal practices of sequestering trouble-
makers. On the other hand, if those driven out are only marginally more
likely than their fellows to lapse from psychological altruism, excluding
15. I am grateful to Christian Nimtz and Boris Hennig for forceful presentation of this
possibility.
them from further social interaction will achieve correspondingly little

in remedying altruism failures. In this case, expulsion is not ethically
progressive, but we no longer have an instance in which the original
function is fulfilled.
Under the original conditions of the ethical project, diminishing the
size of the group was, where even possible, a costly option: recall that
the coalition game culminates with bands of roughly equal strength par-
titioning the environment (§9). Further, the members of these bands are
bound to one another by dispositions to (limited) psychological altru-
ism. Fragmentation of the group is both socially dangerous and psycho-
logically difficult. Appreciating this point enables us to develop an im-
portant perspective on the purported counterexamples to the account of
original function in terms of the remedying of altruism failure. Not only
do the supposedly problematic scenarios fail to undermine the account,
for the reasons given—when they are scrutinized, either they fail to meet
the stipulated condition (remedying altruism failure) or they involve gen-
uinely progressive transitions—but they would not be available in the con-
texts in which our ancestors began the ethical project. Those ancestors
could not have used alternative methods of attending to the symptoms of
altruism failure, the social tensions.
Dictators cannot institute practices remedying altruism failures unless
those dictators are so wise and benign they deserve less pejorative la-
bels. What dictators can apparently do is to relieve social tension by co-
ercive measures. At the dawn of the ethical project, however, dictator-
ship was not a realistic possibility. Dominance of a chimpanzee group
is hard enough (§10), but for language-using, tool-wielding animals with
less sexual dimorphism, repressive rule would be much harder. Failure
to accommodate the wishes of each band member runs serious risks of
fragmentation, and a reversion to the conditions of the coalition game—
made more difficult because of the psychologically altruistic disposi-
tions thwarted in the group’s splintering. The scenarios that might sub-
vert my account of original function not only fail as contrary instances
but also are unrealistic as ways in which, in the early phases of the ethi-
cal project, social tensions could be ameliorated.
The evolution of the ethical project envisaged (§§19–20) created pos-
sibilities for hierarchical societies, within which dictatorial impositions
become real options. These societies can address social tensions in the
brutal ways the scenarios propose. When they do so, they cannot be
conceived as fulfilling the original function of ethics and thus making
ethical progress—for, to recapitulate, they do not remedy altruism fail-
ures. Nevertheless, these societies do raise an important question for the
approach to ethical progress I am developing. Does the original function
of ethics continue to bind those who come later? The question brings home
the important fact that there are two sides to questions about ethical
progress, only one of which has as yet been considered. The facet under
study, here, as well as in Chapter 5, concerns whether one can make
sense of some notion of ethical progress integrable with the narrative of
Part I. Even given success in that venture, however, it would not follow
that progress, understood in the preferred way, would have any force
on later participants in the ethical project. Is “progress” of this sort
something we should aspire to make? Do the functions our ancestors
attempted to fulfill matter to us?
These important issues will occupy us in Chapters 7–9. So far, they
are merely registered—but their turn will come. For the moment, it is
enough to note the circumstances in which the problem background orig-
inally arose, the situations of our fi rst human ancestors with their
uneasy social lives; these are very different from those in which many
latter-day participants in the ethical project have lived (during the past
ten thousand years). The project began in small, egalitarian societies,
in which people with limited tendencies to psychological altruism lived
together. Feeling the tensions of their social lives, they had no success-
ful options except to address the (unrecognized) cause—and ethics was
born with the function of remedying altruism failures. That original
function is refined and gives rise to further functions, in ways we shall
now explore.
§35. Modes of Refinement

Because ethical codes are multidimensional, not all forms of ethical
progress consist in amending rules so the functions of ethical practice
are better discharged. There can be improvements both in socializing
the young and in integrating the precepts with an effective system of
punishment. I consider three modes of ethical refinement, beginning

with advances in techniques of socialization.
The idea of an unseen enforcer serves as a prominent example of
functional refinement. In the earliest experiments with socially embed-
ded normative guidance, the agreed-upon prescriptions are connected
with the system of punishment in an obvious way: those recognized as
breaking the rules will be disciplined. At this early stage, there is little
besides fear of retribution to motivate compliance, and hence, when
people think recognition of their conduct is highly unlikely or impossi-
ble, they may follow egoistic desires. Altruism failures addressed by the
rules persist even after those rules have been publicly accepted; some-
times other members of the group know a breach of the ethical code has
occurred, even though the perpetrator cannot be identified. Mutual sus-
picion produces social conflict. Adding more rules does not help, for
existing rules are directed at the troublesome behavior. The group needs
a device for making the rules more effective when people are deliberat-
ing what to do.
The unseen enforcer is just the right sort of device, for, once the idea
is accepted, deliberators never escape recognition. Using the existing
mechanism for bringing group members into line, the fear of bad conse-
quences from deeds the code prohibits, the thought of perpetual surveil-
lance shrinks the class of altruism failures and diminishes social conflict.
So far, so good. On other grounds, however, this step is problematic.
From the perspective of assessing the cognitive state of the group, it is
not progressive. More importantly, the unseen enforcer sets the stage for
later abuses. When the system of normative guidance has become more
elaborate, when instituting changes to the system is no longer a matter of
a small band deliberating together and trying to arrive at prescriptions
to which all can subscribe, the task of specifying the rules can be left to
a small class of privileged people. Conceived as the enforcer’s represen-
tatives, these social leaders easily inscribe their own dislikes, wishes, and
fears as expressions of the divine will. In the course of subsequent his-
tory, that can render some people’s desires invisible, prohibiting their
expression with such severity that they are only rarely manifest to others
who do not share them, and thus fostering the common judgment that
they are abnormal. Because of its power to secure greater compliance,
the idea of the enforcer counts as a major progressive step along one
dimension, even though it creates obstacles to progress of other types.
This is an important general point: advances in the social technology
that is ethics can shape the way in which further ethical progress be-
comes possible or impossible.
Introducing myths about unseen powers is not the only way in which
our ancestors have improved the techniques of promoting compliance.
As previously noted (§§17, 21), the thought of the enforcer can serve as a
conduit to association of a far richer set of emotions with the ethical code.
Initially, rules are followed out of fear. Later, once the belief that the com-
mandments express the divine will is entrenched, fear of the unseen
power can become awe, respect, even love and gratitude, so, still later,
following the code appears an act of reverent obedience, a gesture of re-
spect and love. The imagined wrathful countenance of a punitive deity
gives way to joy in the prospect of conforming action to shared ethical
rules. People may even describe themselves as acting out of respect for
the moral law, a law they suppose the deity to have inscribed on their
own hearts.16 In my account, these steps count as ethically progressive,
not because they eventually take us to some wonderful place—the privi-
leged “ethical point of view”—but because the multiplication of poten-
tial motivations increases the reliability of compliance to the ethical code.
They are refinements of the original function of promoting social har-
mony through eliminating altruism failures.
Exactly similar points apply to the extension of human sympathy. So-
cially embedded normative guidance first replaces altruism failures with
behavioral altruism, but, as remarked (§§11, 21), continued interactions
16. Kantians will probably not approve of this way of viewing the stance they identify
with a special ethical point of view. They are likely to suppose a decisive break between the
crude and inadequate ethical systems relying on divine commands and the supposedly a
priori unfolding of the moral law. It is, however, completely opaque to me how one could
arrive at correct ethical principles a priori, even if they were (in what sense?) “within,” and
it is worth considering whether Kant’s ethical innovations can be—and should be—viewed
as the recombination of elements from earlier traditions: the idea of divine command ap-
proaches that ethical principles are prescriptions not grounded in good consequences for
human beings, the suggestion of natural law theories (e.g., Aquinas) that the law is written
within our hearts, a long-standing philosophical celebration of reason and a corresponding
suspicion of emotion.
with people, each responding to the other’s desires, can generate sympa-
thetic emotions. You begin the regular work of helping because that is
prescribed, but, as you continue, and as others react to your help, the
responses occur from different causes. This, too, is a progressive shift—
and again not because it attains a special “ethical point of view” in which
mutual sympathy holds sway, but because an increase in the potential
sources of compliance makes fulfilling the original function proceed
more reliably.
People with enlarged sympathies, with respect for the ethical code and
for those who have made conspicuous sacrifices to follow it, are more
likely to comply than those deficient in one of these respects. For them,
failure to respond to the socially approved desires of others opposes
several distinct psychological dispositions. They are not perfect, and,
on occasion, all the mechanisms for leading them to behavioral altruism
fail. Failure occurs more readily as societies grow, as the ethical project
introduces more distinctions and divisions, assigns status differentially,
and makes some members of the large society invisible to others. Even as
some developments in the project buttress old mechanisms and introduce
new ones, they simultaneously make psychological altruism of certain
types more difficult—sympathies expand, but only as far as the limits of
a group with which the higher-order altruist identifies him- or herself.
Transitions progressive in some respects are regressive in others
and even block off further ethical advances. In changed social cir-
cumstances, inherited modes of socialization encourage some forms of
altruism failure.
A second mode of ethical progress consists in integrating the code
with the system of punishment. Normative guidance emerges in a con-
text in which a rough system of punishment is already present (§12), and
it builds on psychological capacities for avoiding behavior that provokes
punishment. In part, the introduction of ethics substitutes prevention for
punishment. It does so by combining specification of commands with
the threat of punishment if they are broken. The systems of punishment
found in chimpanzees inflict harm on individuals who have done certain
things. An ethical code must approach such raw reactions to behavior
more self-consciously. Implicit in any endorsement of punishment is a
suspension of altruistic response: at least some of the desires of the indi-
vidual who has performed the rule-breaking action do not count; they
are not to be accommodated in dealing with him. The perpetrator wants
to avoid all the harm others in the group intend to visit upon him, but
this is not an occasion for responding to his desire. Integrating the ethi-
cal code with a system of punishment requires some things resembling
altruism failures not to be remedied, because, as we would ordinarily
say, the rule breaker has made some of his desires forfeit.
The ethical code connects with the system of punishment through
three questions: When? Who? How? The first issue is addressed by ap-
pealing to the prescriptions of the code: retribution is to occur just when
one of these rules is broken. The second is the translation of the earlier
raw reaction into the social practice of normative guidance: we look to
the individual(s) who did the action that broke the rule. For the third,
the ethical code must itself pronounce, declaring which kinds of harms
are to be inflicted. It must decide what accommodations with the desires
of the perpetrator are to be made. As noted, his desire to avoid all harm
is not to be the basis of an altruistic response; in carrying out the punish-
ment, ignoring that desire is not an altruism failure to be remedied. That
does not mean, however, that the code must be indifferent to everything
the rule breaker wants.
The original function of the code is to remedy altruism failures caus-
ing social discord, and that is achieved if punishment and the threat of
punishment decrease the frequency with which failures occur. Setting
some of the perpetrator’s desires outside the scope of altruistic responses
expresses a commitment to this function: if those desires are ignored, he
and others will be less likely to engage in future rule-breaking behavior.17
In many situations of rule breaking, other members of the group are
compelled to choose the target of their altruistic response, forced to de-
cide whether to align their desires with one person or another. The ap-
paratus of punishment makes this decision for them by distinguishing
between perpetrator and victim. In ignoring the desires of the perpetra-
tor and attending to the desires of the victim, one prominent—and readily
17. Here my functional approach to ethical progress aligns itself with views of punish-
ment that are forward-looking, that view punishment as a means to prevent future harm.
The raw reaction of a desire for revenge is transmuted into a piece of social technology.
understandable—thought is to focus on the exact way in which the rule

breaker has overridden the victim’s desires and to make exactly the
adjustment that will restore the status quo. Hence arise punishment
rules demanding precise reciprocity: as in the lex talionis.
The attraction of such rules is evident from the early Mesopotamian
law codes, and it is highly probable that previous generations of ethical
experimenters had used similar prescriptions. Yet, as we saw (§23), the
lex talionis was elaborated in a peculiar—and disturbing—way, through
the literal idea of compensating for human loss through the death of a
“corresponding” person. Someone who has murdered the child of an-
other is to be punished by having one of his or her own children (of the
same sex as the victim) put to death; wives of rapists are to be raped.
Behind this law lies an apparent dyadic calculus of violated desire: the
code focuses on only two members of the society, perpetrator and vic-
tim; each desires enduring relationships with other individuals; the per-
petrator has violated one of these desires, and the victim is compensated
through violation of the perpetrator’s corresponding desire. One altru-
ism failure is canceled by its precise counterpart.
Societies made progress when they rejected this appalling law in favor
of one harming the perpetrator him- or herself. That can be viewed as
a refinement of the function of the ethical code. Central to the ethically
mediated practice of punishment is, as we have seen, delineating some
desires of the perpetrator as outside the scope of altruistic response.
Yet punishments striking at the perpetrator “through” his or her rela-
tives—in the prominent instances, his children and wife—introduce a
vast new class of altruism failures, violations of the wishes of those who
are going to be the immediate targets of retributive action. Precisely be-
cause individuals, not groups or clans, are the bearers of desires, efforts
to remedy altruism failures are better directed by considering the wishes
of all the members, not simply those of the family, the clan, or some other
collective (or even some single person taken to “represent” this collective),
and the change in focus counts as a progressive step. Acceptance of the
ethical code and its enforcement through punishment requires that
members of the society commit themselves to noninterference in the
inflicting of harm on perpetrators, but they have made no similar commit-
ment to allowing suffering to be imposed upon others. The articulation
of the lex talionis using the “corresponding relative” as a vehicle for ret-
ribution is likely to inspire protests and reactions from those who sym-
pathize with the people targeted. In its early form, the law introduces
altruism failures instead of remedying them and, in doing so, destroys
rather than promotes social harmony. The modified version confining
discipline to the perpetrator can thus be seen as a refinement of the origi-
nal function of ethics.
Another important aspect of ethically mediated punishment deserves
consideration. The initial stages of ethical experimentation had to oper-
ate with a conception of rule breakers, and it is overwhelmingly probable
that they introduced the idea in the most straightforward fashion, think-
ing of perpetrators as people whose behavior violated rules. The func-
tion of punishment is to decrease the future occasions on which similar
altruism failures occur, and realizing that function depends on delibera-
tors’ ability to recognize in advance when a potential action is at odds
with the code. As prescriptions and the concepts involved in them become
more complex, it is apparent that there are episodes in which people
unwittingly break rules. Inflicting harm on those who tried to honor the
code is likely to sap the motivation to comply, since dedicated efforts still
lead to trouble. Acknowledging human limitations, incomplete knowl-
edge, and life histories yielding a distorted perspective makes progress
in fulfilling the original function. Breaking the rules is forgiven when
violators convince others they were doing their best to comply.
The extent to which an ethical code should take account of our limita-
tions of knowledge and the environmentally caused distortion of our
sympathies is something on which the code itself must pronounce. The
strategy of conceding something to human frailty is a progressive step in
the evolution of ethics. Our understanding of how to make such conces-
sions remains incomplete. Too rigid an insistence on the fact of rule vio-
lation proliferates altruism failures by weakening the motivation to try;
too easygoing an approach multiplies altruism failures by undermining
the rules themselves. The challenge, for our predecessors and for us, is to
operate in the space between these extremes and to sharpen the concept
of an excusable offense. Because of the variety of ways in which human
effort can fall short, our progress remains incomplete—we continue to
work out the extent to which we excuse human imperfection.
For the last of my three modes of functional refinement, return to the

attractive, but oversimple, account of ethical progress in terms of “the
expanding circle” (§33). It is relatively easy to understand how many
examples of ethical progress are functional refinements achieved by wid-
ening the scope of existing ethical precepts. A Gospel message exhorts
us to love the downtrodden and despised among us. Incorporating the
prescription in our ethical code identifies members of our society who
have previously been outside the scope of altruism for many other
people—we are directed to remedy a class of altruism failures, allowed
by the rules previously articulated (at least under the common under-
standing), and doing so enhances social harmony. Imagine we come to
view slaves as “full people,” or recognize women’s desires for participa-
tion in public life, or understand how some people can find love with
members of the same sex.18 In all these instances, prior altruism failures
are to be remedied: we now try to act in ways that take into account the
desires of those whose wishes have been overridden, neglected, con-
demned, or simply not seen; the result is a diminution of social conflict.
An especially important instance of “the circle expanding” came in
the lives of our Paleolithic ancestors when they extended certain of their
ethical principles to cover outsiders in neighboring bands. According to
my “how possibly” explanation (§19), the impetus to partial recognition
of people outside the local band came from the perceived possibility of
trade. If so, the result would have been something that was neither a full
society, of the kind already constituted in the small groups of cooperat-
ing human beings, nor a confrontation of unremittingly hostile bands. We
can think of it as a limited society, introduced to promote a particular
purpose (exchange of surplus goods). Because of the desirability of such
exchange, a restricted analog of social harmony was also desired in this
limited society, to be achieved through the extension of some parts of the
ethical code. The problem background is not that of decreasing social
conflict within an already existing society, but of creating a limited soci-
ety that requires decreased confl ict at its founding. To obtain the level
18. With respect to this last example, as §32 noted, the altruism failure consists of an
attitude toward certain types of desires (desires for fulfi llment of same-sex love) that the
progressive transition modifies.
of social harmony enabling this limited society to emerge, classes of

previously occurring altruism failures (in aggressive interactions among
neighboring bands) had to be remedied. Expanding the circle does
just that.
Major instances of ethical progress can thus be seen in terms of func-
tional refinement, where the function in question is that identified as the
original function of ethics (§34). Some of the difficulties recognized in
§33 are now resolved. Not all, however. Progressive changes of a particu-
lar type—those offering new ideas of the human being and of human life
remain problematic. The next task is to show how the functional ap-
proach to ethical progress can be extended to cope with these cases.
§36. Functional Generation

In the evolution of ethics there is an original function that grounds the
concept of ethical progress, but there are also secondary functions, gen-
erated from the ways in which that original function is discharged. The
possibility is evident from the domains serving as analogs: in Darwinian
evolution and in the development of technology, responses to older prob-
lems generate new problems and new functions.
Consider our hominid ancestors. To simplify considerably: they moved
from forest to savannah, taking on an increasingly upright posture; for
reasons still much debated, they also grew bigger brains. The two trends
combined to cause a new difficulty. Females, whose pelvic structures are
constrained by the requirements of walking upright, give birth to young
with larger heads. Adaptive responses to old problems generate a new
problem of giving birth. The hominid/human solution is to truncate the
pregnancy so the baby’s head does not grow too big. That generates a
further problem, one of nurturing an infant that cannot do much for it-
self. The original functions of upright gait and big brains (whatever ex-
actly they were) generate the function of curtailing pregnancy, and the
latter function generates the function of extensive parental care.
When new biological functions emerge, much is held constant through-
out the process. Organisms continue to be subjected to the basic pres-
sure of natural selection, competing for reproductive success. Because
of solutions achieved by their ancestors, they are locked into particu lar
structures and modes of behavior. Their inheritance frames the new

problems generated for them by selection. Successively solving prob-
lems, they make local progress. Whether they also make global progress is
another matter. Is there any Darwinian basis for comparing the primates
in the trees and the hominids on the ground?
Or consider the development of systems of mechanized transporta-
tion. The initial problem background is one in which people feel spa-
tially confined: they want ways of going more speedily over longer dis-
tances and of reaching previously inaccessible places. Mechanized
vehicles fulfill these functions, but, once they have been provided, new
problems must be solved. Appropriate channels (roads, railways, canals)
have to be built and maintained. Traffic must be coordinated, signals
designed, regulations for responding to them enacted and enforced. All
sorts of auxiliary systems arise, as responses to the problems generated
along the way—driving schools, traffic courts, fuel-distribution networks,
insurance requirements, inspection stations, the familiar paraphernalia
of the car-infested world. Local progress is easy to discern. Early auto-
mobiles made more destinations accessible. They generated new prob-
lems of road building. Technology supplied new surface materials: a
progressive step. Standing back from the individual transitions, how-
ever, questions of progress become harder. Did the automobile improve
the world? Was there global progress?
Does a notion of functional generation apply in the ethical case? If so,
does it complicate judgments of global progressiveness? I shall argue for
affirmative answers. Begin with an example figuring in the transition from
the first ventures in normative guidance to the ethical approaches present
at the dawn of history: the emergence of a conception of a good life involv-
ing some type of individual self-development (§§20–21). The initial prob-
lem is to remedy the altruism failures causing social conflict. In the earliest
phases, our ancestors focus on the basic desires of group members—
desires for food, for shelter, for protection. Some desires are endorsed,
others seen as contaminated. Preliminary development of these notions is
a first ethical step—desires for food, shelter, and protection are endorsed;
desires to engage in unprovoked violence are taken to be contaminated;
simple rules express these judgments. The society now attempts to find
ways of enabling all its members to satisfy the endorsable desires.
That is the starting point for a process leading, a very large number of
generations later, to societies with a far richer conception of living well.
My “how possibly” explanation links the two by starting with divisions
of labor (§19). Dividing the labor helps satisfy endorsable desires; speci-
fication of roles and assigning people to roles on the basis of talent refine
the social technology. In the process, new desires emerge: people find
some roles more attractive than others. The conception of the good life
expands, as members of the society view living well not just as meeting
the needs identified at the beginning—it is no longer enough to have a
full belly and a warm, safe place to sleep.19 Additionally, living well in-
volves attaining roles one wants and having the chance to choose one’s
roles.
The original problem remains constant throughout this process. Eth-
ics must continue to promote social harmony through remedying altru-
ism failure. Now it must do so on an expanded field of desires. If the
means to satisfy the endorsable desires present at earlier stages have be-
come available, it is in principle possible to solve the older problems (al-
though, in practice, that may not be done because of violations of the
code). The residual problem is to address social tensions caused by al-
truism failures with respect to desires that the (in principle) successful
solution has generated. Ethical principles are also required to respond
to conflicts within the individual’s expanded repertoire of desires, and
in this sphere, prescriptions for character development emerge.
This is a story about the generation of functions, where the notion of
generation is that discerned in the biological and technological examples.20
To make the structure explicit, consider the following sequence:
19. So simple a menu of basic desires was probably never sufficient to the idea of the
good life, even at the beginning. But this elementary formulation suggests the appropriate
contrast.
20. The account is a story—a “how possibly” explanation. It was originally provided to
show how ethical evolution could have led by gradual steps from the original form of the
ethical project to the complex practices recognizable in the ancient world. It is used here to
illustrate the notion of functional generation with respect to ethics, and also to show how a
challenging example—the emergence of richer notions of the good life—can be seen as
proceeding by locally progressive steps. Neither of these purposes requires my story to be
an account of actual history.
1. An initial problem (P0 ) requires the existence of something to

fulfill an initial function (F0 ).
2. Something (E0 ) is introduced to fulfill F0.
3. For E0 successfully to fulfill F0, a new problem (P1 ) must be
solved.
4. P1 requires something to fulfill a new function F1.
In cases that exemplify this sequence, I shall say that P0 generates P1,
and that F0 generates F1. A function for ethics is anything resulting from
the original function (the remedying of altruism failures) by a finite num-
ber of steps of functional generation, that is, from a finite number of it-
erations of the sequence 1–4.21 The narrative of the last paragraph identi-
fies responding to the endorsed (elementary) desires of all group members
as a function of ethics, introducing and elaborating the division of labor
as a function of ethics, specifying roles as a function of ethics, and ulti-
mately endorsing an expanded set of desires as a function of ethics.
(Those identifications would be preserved, even if my “how possibly”
story is not historically accurate, given the minimal assumption that the
steps leading to the richer conception of the good life involved group
responses to antecedent problems.)
People need not always be aware of their problems. Proliferation of
mechanized vehicles might give rise to a period during which traffic was
unregulated. Some of the population, people unaffected by the accidents
befalling others, might see no difficulty in the status quo. The same can
occur with the ethical project. The evolution of ethical practice can give
rise to codes whose shortcomings and burdens are felt by only a few.
When that occurs, the first task of would-be reformers is to make the
problem apparent to all members of the society. That may occur through
verbal presentations—as with Woolman and Wollstonecraft—or through
the exerting of pressure from people whose voices have not previously
been heard. In the earliest phases of the ethical project, however, when
groups are small and everyone can make his or her particular perspective
known, the problems generated are visible. Stratified societies create
21. At risk of being unnecessarily explicit: this specification counts the original function
as a function, since it is obtainable from itself by zero steps of functional generation.
conditions under which it is harder to recognize when the original func-

tion of ethics is not being fulfilled.
The significance of this point will occupy us later (§§38, 46). The im-
mediate task is to understand the difficulties with global progress. It will
help to have a notion of functional conflict. External functional conflict
occurs when there are two different domains of human practice such
that the optimal available ways of discharging some of the functions of
one of the practices are incompatible with the optimal available ways of
discharging some of the functions of the other. Internal functional con-
flict occurs when there are two subsets of the functions of a practice such
that the optimal available ways of discharging the functions in one sub-
set are incompatible with the optimal available ways of discharging the
functions in the other. It is easy to understand how there can be external
functional conflicts between economic practices and ethical practices,
or between ethical practices and sexual strategies, but my interest will
be in internal functional conflict within ethics.
The story of the emergence of a richer conception of the good life re-
veals conflict of this sort. In the transition to hierarchical societies with
pronounced division of labor and of status, fulfillment of the generated
functions (supplying enough to satisfy the previously endorsed desires
of all) is obtained at the cost of compromising the satisfaction of the origi-
nal function. Similarly, in the invocation of the unseen enforcer, there is
functional conflict between the original function of remedying altruism
failure and improving compliance to the rules of the code. The societies
who emerge from these changes have the choice of whether to take steps
to improve the fulfillment of the original function at cost to the solutions
they have achieved with respect to the generated problems. (That is not
to say they will recognize this choice unless it is forced upon them by some
militant reformer or through the collective pressure exerted by some large
group.)
Armed with notions of functional generation and functional conflict,
we can now take up questions about local and global progress. To what
extent are global comparisons available in appraising the evolution of
the ethical project?
§37. Local and Global Progress

In the evolutionary and technological cases certain judgments about
progress are easily made: we compare the first organisms who invade a
new environment with their later descendants and view the latter as hav-
ing made progress through the refinement of organs or forms of behavior
that were initially able to provide only clumsy solutions to the problems
the environment poses. Similarly, later surface materials for roads pro-
vide a smoother ride and thus constitute a progressive step. Judgments
of local progress are facilitated because it is possible to focus on a single
function, or concordant set of functions, and examine how completely
(thoroughly, speedily, etc.) they are discharged. Assessments of global
progress are difficult because multiple functions come into play. With
respect to technology, we recognize external conflicts between practices:
our ancestors who lacked mechanized transport clearly had less func-
tional forms of transportation, but in other domains of human practice—
their communal interactions with one another and their responses to the
physical environment, perhaps—they may have been better able to ful-
fi ll particular functions.22 So, too, in the ethical case.
Some examples of ethical change, like the easy cases for the evolution
of life and of technology, focus on single functions or on concordant sets
of functions. They can be treated as progressive by concentrating on the
appropriate function and applying the ideas of previous sections. A more
ambitious account would attempt to extend talk of ethical progress to
comparisons of temporally distant forms of ethical practice, with func-
tions that conflict. Can an account of this sort be given—and, if not, can
one manage without it?
Imagine two societies. The Primitive live under conditions in which
it is hard to satisfy the most basic desires: food is in short supply, shelter
is hard to find, danger constantly looms. Their ethical codes respond to
the altruism failures that arise under these conditions and are tightly fo-
cused on responses to the basic desires. The Civilized, by contrast, have
22. Whether this is so requires deeper scrutiny of their lives than I can offer here. I do
not endorse the conclusion that they were better at discharging some functions, but simply
recognize this as a question.
developed means of acquiring resources sufficient to satisfy all the group

members’ basic desires. They have done this by introducing divisions of
labor and distinguishing social roles. Their ethical code enjoins sharing
resources so everyone’s basic desires are met, and let us suppose, opti-
mistically, that the Civilized live up to their code in this respect. Because
they know an efficient division of labor is needed if they are to provide
for the basic desires of all, they have introduced rules, assigning roles on
the basis of perceived talent and enjoining all members of their society to
develop their talents and to contribute to the social effort. Each of these
members has wishes about the form of his or her social contribution,
and, because the roles are differentially attractive, many of these wishes
are frustrated. Amending the system, however, allowing rotation among
roles, would cause shortages of basic resources.
Can we talk of progress across the Primitive/Civilized divide? The
Civilized view themselves as having made progress, talking of their im-
proved conception of the individual and of the good life—they make
witty remarks about disgruntled sages and contented pigs—but that only
means that, having a wider repertoire of desires, they cannot be happy
with a way of life that satisfies only some of those desires. Progress, as we
saw earlier, cannot be grounded in these subjective criteria (§28). We who
come late in the sequence of “experiments of living” may be appalled at
the thought of going back to life as the Primitives lead it. We may think of
ourselves as unable to do that, incapable of simply wiping away the kinds
of wants and aspirations our social environment has equipped us with;
we may imagine descendants of ourselves for whom those desires and
aspirations did not arise as truncated and bereft. So far, however, there is
no basis for claiming that what has occurred between our Primitive an-
cestors and ourselves is progress, rather than simply a historical devel-
opment resulting from people’s ingenious attempts to satisfy their basic
desires.
To support any claim of global progress, it is necessary to resolve a
functional conflict. The Civilized have solved the original problem of
altruism failures, they have satisfied the endorsable basic wishes of all
members of their society, and they have expanded the repertoire of
desires. This latter expansion has also introduced occasions—even long
periods of time—when, for many people, many of the newer desires go
unsatisfied. No response to these frustrated desires is permitted. That

seems to constitute regression through the creation of a large class of
altruism failures (failures provoking social tensions).
One might classify the unsatisfied desires as not endorsable: members
of Civilized society would not want a situation in which the participa-
tory wishes of all were satisfied (say, through rotation of roles). That,
however, is simply to announce a preference for fulfilling some functions
rather than others. Given that participatory desires are as central to the
psychological lives of the Civilized as basic desires are to the lives of
the Primitives, there is pressure to endorse them. If those desires were
endorsed, the Civilized would have made progress over the Primitive
by fulfi lling some of the generated functions of the ethical project but
have compromised the original function. Can this count as an overall
advance?
The concept of progress is central because judgments about progress
figure in our evaluations of past transitions and our deliberations about
what further modifications to make (§32). While the conception of ethi-
cal progress so far elaborated cannot pronounce on questions about
some large historical or prehistorical trends (“Was the shift from Primi-
tive life to Civilized life progressive?”), that does not leave us uncertain
about how to go on from where we are. Faced with the question “Should
we modify our way of life to become Primitive again?,” we cannot an-
swer by arguing that the reversal would be regressive (or that it would
be progressive!). But, from the standpoint of pragmatic naturalism, the
question is idle. Whether the historical processes that have led us to
where we are embodied a fundamental mistake or not, we simply can-
not go back. We are equipped with wants that cannot be wished away:
there is no genuine possibility of arranging a new Primitive future for
ourselves. The lack of a concept of global progress does not make a
difference in the situations where a notion of ethical progress is most
needed—and that is enough for a pragmatist not to be moved to settle the
question.
The contemporary human situation is hardly that of the Civilized—
they are already better than we are at responding to altruism failures. Yet
it is not hard to envisage how they might make further progress, in ways
responding to the function they currently slight. They could introduce
new conceptions of social roles, so people who adapt to forms of social

contribution they did not initially prefer are especially highly valued, or
so what is hailed as most important is energetic ser vice to the society
rather than the form it takes. Even without a principled stand on the func-
tional conflict involved in comparing them with the Primitive, a notion
of local progress suffices for (their) practical purposes.
In striving to improve our practices, we cannot foresee everything.
A messianic23 perspective evaluating their evolution in the indefinitely
long run is impossible for us. Yet issues of functional conflict cannot al-
ways be avoided. As we shall see (§§39, 43) functional conflict lies at the
core of the most important challenge to my version of ethical naturalism
and is the source of the normative questions addressed in Part III (Chap-
ters 8 and 9). Although the account of ethical progress so far presented
answers to many of the uses for which a notion of ethical objectivity is
required, it will eventually need further articulation on the basis of a
proposal about method in ethical practice.
§38. Ethical Truth Revisited

My conception of ethical progress does not rely on any prior notion of
ethical truth. Sections 30 and 32 already recognized the possibility of
approaching truth constructively: true statements are those arrived at by
following particular procedures. The familiar versions of constructiv-
ism are ahistorical. They presuppose a set of privileged procedures
individuals can follow in whatever epoch they fi nd themselves. Kant,
for example, offers us a supposedly a priori method and understands
the truths of ethics to be those generated by following that method.24
23. I borrow the term from Isaac Levi (The Enterprise of Knowledge [Cambridge, MA:
The MIT Press, 1982]), who deploys it to make a kindred point about the importance of
local (“myopic”) standards.
24. The method is specified by the first formulation of the Categorical Imperative, in
which potential maxims are to be scrutinized to see if they can be willed without contradic-
tion. Ethical truths are just statements that record the results of the process—thus, if
“Make a promise not intending to keep it” fails the test, then “Promise breaking is imper-
missible” constitutes an ethical truth. I shall not try to work out the intricacies of this
theory, since its procedure seems so indeterminate. A more plausible constructivist approach
can be derived from Rawls, who can be read as viewing ethical truths as expressing the
Pragmatic naturalism articulates the constructive approach differently,

attending to the historical processes out of which statements come to be
accepted and following the approach to truth suggested by Peirce and
James.25
Descriptive counterparts of ethical rules count as true just in case
those rules would be adopted in ethical codes as the result of progressive
transitions and would be retained through an indefinite sequence of
further progressive transitions. There is no prior conception of ethical
truth, so that people make ethical progress when they discover (or stum-
ble on) independently constituted ethical truths. Progress is the prior
notion, and descriptive counterparts of rules come to count as true in
virtue of the fact that they enter and remain in ethical codes that unfold
in a progressive sequence (“truth happens to an idea”). Derivatively,
Tarskian machinery of correspondence truth allows the extensions of
ethical predicates (“wrong,” “good,” and so forth) to be fi xed to make
the counterparts of rules accepted under indefinitely proceeding pro-
gressive transitions true (§30), but there are no prior independent prop-
erties to which those who formulate and preserve the rules respond.
Given this understanding of ethical truth, can we make claims about
the truth or falsehood of ethical statements? We are never at the end of
the ethical project, never at some hypothetical limit of any progressive
sequence of ethical practices. Yet we confidently assert some, relatively
imprecise and vague, ethical statements, declaring that honesty is typi-
cally right and that murdering people who have done no harm is typically
wrong.26 How can we be confident these statements will endure as our
ethical practices progress?
decisions arrived at in the original position. There are well-known difficulties in specifying
which version of the original position is privileged (and why).
25. See Peirce, “The Fixation of Belief,” and James, Pragmatism (Cambridge, MA:
Harvard University Press, 1978). In connection with ordinary statements about the physi-
cal world, as well as scientific statements, I regard James’s development of the thought as
compatible with a correspondence theory [see my “Scientific Realism: The Truth in Prag-
matism” forthcoming in Wenceslao Gonzalez (ed), Scientific Realism and Democratic Soci-
ety: The Philosophy of Philip Kitcher, (Rodopi, Poznan Studies in the Philosophy of Sci-
ence) and “The Road Not Taken,” to appear in German in a volume edited by Marcus
Willaschek, celebrating the centennial of William James’ Pragmatism (Suhrkamp].
26. As with many generalizations in other areas, it is hard to state these with any great
precision. This has sometimes inspired authors to think of ethical precepts as analogous to
We cannot. There is no good reason to think the concepts we use to

formulate such generalizations will be endorsed by our successors. They
may introduce quite different vocabulary for organizing their ethical
practices, and appreciation of that possibility should incline us to be
tentative and vague in generalizing. The confidence we have is that some-
thing very like what our imprecise statements express will be found in
later progressive extensions of our own practice, and our assurance rests
on two related thoughts. First, despite the variation we find in the world’s
ethical experiments, some themes are discovered again and again, not
easily captured in any exact formulation, but universally present. Tradi-
tions lacking them would make progress by adopting them. Second, as
we think about possibilities of jettisoning our own versions of the rules
expressing these themes, we find it hard to conceive how any society that
abandoned them, that lacked anything like our prohibitions against
murder and lying, would have an ethical system able to reduce social ten-
sion through remedying altruism failures. We make a well-based conjec-
ture about which rules are “destined” to endure, using our own, quite
possibly crude and inapt, language to gesture toward a theme that will
be present in the “last words of the last man.”
Can we be sure? The recognition of functional conflict internal to ethi-
cal practice prompts a genuine skeptical question. The considerations of
the last paragraph turned critically on supposing that the original func-
tion of ethics must continue to be fulfilled, that it cannot be overridden
by other functions—at least, not to the extent of tolerating the sweeping
altruism failures expressed in widespread lying and indifference to hu-
man life. To rely on the continued commitment to the original function
of ethics in so minimal a form appears no less reasonable than suppos-
ing that something akin to a predictively successful scientific generaliza-
tion is likely to figure in future scientific practice. The point is buttressed
further (Chapter 9), by insisting on the priority of the original function
of ethics.
“Aristotelian categoricals”—generalizations about murder or stealing are akin to general-

izations about the properties of types of plants and animals; see, for example, Philippa
Foot, Natural Goodness (Oxford: Oxford University Press, 2001), and Michael Thompson,
Life and Action (Cambridge, MA: Harvard University Press, 2009). I think the connection
with biology is a fruitful one, often spoiled by elaborating the biology as if Aristotle had
written the last word on the subject.
There is a core set of imprecise precepts we can hail as ethical truths.

There are also areas in which the convergence of progressive traditions
is genuinely in doubt. Rival traditions offer different elaborations of the
ethical project, alternative cultural lineages, and societies belonging to
those lineages may by no means find the same problems salient. As they
progressively modify their ethical codes to meet distinct challenges, their
ethical concepts and the rules, roles, and institutions they introduce and
develop may be quite various. Perhaps the solutions achieved are readily
combinable, so that, faced with a diversity of codes, there is a way of assem-
bling ingredients from them to produce something progressive with respect
to all. But perhaps not. We can imagine two different ethical traditions pro-
ceeding indefinitely, making a series of progressive transitions, without its
ever being possible to integrate their differing accomplishments.
We saw this already in the contrast between the Primitive and the Civi-
lized. They take different stances with respect to a functional conflict. We
can envisage the two societies meeting, and each recognizing the func-
tions the other has learned to fulfill most completely. Inspired by the
comparison, they might continue to evolve their ethical practices—just
as the Civilized respond to the problem of altruism failures, so the Prim-
itive absorb some of the devices for increasing the supply of basic re-
sources, perhaps even introducing versions of those roles that generate
the richer conception of human life. So, as time goes by, more and more
roles come to be held worthy among the Civilized, and the society be-
comes more egalitarian; cooperation increases and altruism expands, but
no practice in the lineage that descends from the Civilized ever frees it-
self from competition, status differences, and inequality. By contrast, the
cultural offspring of the Primitive emphasize, from the beginning, the
maintenance and intensification of cooperation on equal terms; the result
is a certain inefficiency; surpluses are much smaller, roles are not so finely
divided; with time, there is some proliferation of different forms of social
contribution, but it is always constrained by an emphasis on cooperation
and the preservation of equality; the descendants of the Primitive are of-
fered fewer options about how to live, but they live on terms of greater
equality than their Civilized cousins.
The societies in these lineages do not overlook the functions the other
fulfi lls better. They judge it good to have what the alternative tradition
achieves. In addressing functional conflict, however, they set priorities.

Discharging one function comes first—one value takes precedence. Dur-
ing the course of time, they may come closer, even though, however long
their unfolding of the ethical project continues, significant differences
between them remain. The lineages have exhibited different variet-
ies of ethical progress, and this points to a real incommensurability of
practices.
This is the possibility of pluralism (§32).27 Where that possibility is
realized there will be no determinate ethical truth. The central thrust of
this chapter has been that we can make enough sense of ethical progress,
prior to any conception of truth, to combat the thought that the evolu-
tion of ethics is a history of mere changes. Once ethical progress is under-
stood, the concept allows for a constructive development of ethical truth,
one useful in enabling us to characterize inexact core statements we take
to be shared among ethically progressive traditions. Notions of truth
and falsehood do not always apply in the ethical domain, for the core of
ethical truth is surrounded by a periphery of pluralism.
§39. Residual Concerns

The inquiry of this chapter began with a pair of difficulties: (1) resisting
the mere-change view requires making sense of the concept of ethical
progress; (2) if progress is the accumulation of (prior) truth, then the con-
cept does not fit the history of Part I. The solution is to revise the
concept of progress so it is prior to the notion of truth. The previous
sections attempt to carry out this revision in some detail.
Important concerns remain. In closing, I consider three. First, it has
not been shown how to connect ethical progress with the psychological
processes moving the great innovators, despite the prominence of the
demand for a similar connection in the arguments of the previous chapter.
Second, grounding the concept of progress in the problems experienced
27. The pluralism envisaged here is close to that defended by Isaiah Berlin. One impor-
tant feature of it is the recognition by each of the rival traditions of the values taken as fun-
damental by the other— each feels the tug of what the other does so well. (I am grateful to
Chris Peacocke for emphasizing to me this crucial facet of Berlin’s pluralism, and thereby
prompting me to make my acceptance of it explicit.)
by the ethical pioneers threatens to make the account vulnerable to the

arguments against subjectivism.28 Third, recognizing that something
has (or, worse, had) a function lacks normative force: if the original func-
tion of the ethical project was to remedy altruism failures, why should
that matter to us, who come so late in it?29
To address the first concern, we need a notion of justification of the
kind sketched in §32: an innovator justifiably makes a revisionary pro-
posal if the underlying processes are likely to lead to ethical progress.
Empathetic understanding, especially when sensitive to (endorsable,
uncontaminated) desires that are typically suppressed can foster societal
discussion to remedy existing altruism failures. People who call atten-
tion to the frustrated desires, beginning a broader deliberation about
them, are likely to move the ethical project toward refinement of its origi-
nal function. Wollstonecraft could not frame the issue in this (analyst’s)
way, but her sensitivity to women’s suppressed aspirations would be ex-
pected to refi ne the original function of ethics. Woolman is a more
ambiguous figure. His account of his route to rejecting slavery reveals
the role his own anxieties about personal salvation played. Insofar as
these were dominant, we cannot view the psychological processes that
generated his proposed reform as reliable. Nevertheless, some consider-
ations underlying his hesitation about writing the bill of sale—concerns
for the pains and abuses visited upon slaves, worries that official interest
in their well-being (their salvation) is a sham—indicate the operation of
an empathetic capacity. To the extent he resembles Wollstonecraft, we
can view him, too, as a justified agent of ethical change.
The figure who emerges from the Gospels shows many inconsistent
tendencies. Yet, if you focus on some passages (and ignore others), Jesus
appears as the most striking example of empathetic identification in
recorded history. The sweeping exhortations to forgive, to cherish and
nurture those whom society neglects or despises, are inadequately de-
scribed by noting that they point out potential altruism failures on an
28. I am indebted to Sharon Street for a forceful presentation of this worry.

29. Several people have expressed this concern, offering it in different (sharp) formula-
tions. I am indebted to Mario Branhorst, Laura Franklin-Hall, Kirsten Meyer, and an
anonymous reader.
extraordinary scale. Because of the scope of his sensitivity to others, you

might credit Jesus with a capacity for empathetic understanding whose
exercise reliably generates progressive proposals for reform. That would
be a misdescription. Jesus does not make proposals for criticism, debate,
endorsement from others with different points of view. He presents com-
mandments. He is portrayed as an ethical authority who recognizes his
role. In my view of the ethical project, there are no “moral teachers”—it
would be wrong to hail him as one. There are justified reformers who
make proposals, and we can count him as one of the greatest—even if
that does not fit his own self-image or the image of his followers. 30
The second worry stems from the connection made between func-
tions and their problem backgrounds. Something is only a problem, it
seems, if it is felt as interfering with the satisfaction of desires; hence the
account of progress offered here is based on subjective considerations of
the sorts found wanting in §28. The partial justice of this charge is ex-
pressed in some of my formulations, for example, in characterizations of
people as wanting relief from the tensions of chimp/hominid social life.
With equal justice, however, the problems can be regarded as objective
features of the social situation to which, in our evolutionary past, our an-
cestors were led. Desires for relief are in no way idiosyncratic—they
would be felt by virtually all members of our species. Indeed, the discom-
forts of chimp/hominid social life would affect human beings and
hominids alike (and possibly chimps as well).
The subjective/objective dichotomy is too blunt to apply to the con-
cepts of function and problem background that lie at the basis of my ac-
count of progress. Forcing the dichotomy into the discussion is bound to
mislead. Dewey saw the point clearly, remarking that “moral conceptions
and processes grow naturally out of the very conditions of human life.”31
30. Few people in the history of the ethical project have offered progressive proposals,
and few have had the insight into possibilities of altruism failure achieved by the three fig-
ures whose status I have (briefly, unsympathetically, and coldly) evaluated. One of the wise
sayings attributed to the most wide-ranging of these reformers is that judgment is danger-
ous. Those whose contributions to the ethical project are infinitesimal in comparison (as
mine are) should acknowledge that.
31. John Dewey and James Tufts, Ethics, 2nd ed. (New York: Henry Holt, 1932). In the
original, the sentence is italicized.
The approach to ethical progress developed here articulates that

thought.
The third concern cuts deepest, and it will require extension of the
account given in this chapter (including the adoption of a substantive
normative stance). As the worry clearly recognizes, people are not bound
to use a device to do the job for which it was originally introduced. Why,
then, should the “original function” of ethics have any purchase on us?
Why should we suppose “progressive” transitions—those contributing
to the original function—are to be valued? This chapter has recognized
functional generation and functional confl ict. An extreme case of the
latter occurs when earlier functions are simply dropped in favor of
the functions they have generated. Why would the ethical project be im-
mune from that?
The next three chapters build resources for answering these questions
(and others besides). We start with some metaethical issues often seen as
devastating for any form of naturalism.
chapter 7
Naturalistic Fallacies?
§40. Hume’s Challenge

Naturalists, so the story goes, inevitably commit “the naturalistic fal-
lacy.” There is, however, no such thing as the naturalistic fallacy. While
critics agree that naturalistic ventures are inevitably embroiled in some
error to which they attach this name, they offer very different diagnoses
of what mistake is made.1 This chapter considers several distinct objec-
tions, attempting to show that, whatever their merits against other natu-
ralistic accounts, they hold no terrors for pragmatic naturalism. Yet as we
shall discover (§§43, 56), substantive ethical work is needed to absolve
pragmatic naturalism completely.
The most venerable challenge derives, ironically enough, from one of
the greatest naturalistically inclined philosophers. In a famous passage,
Hume writes:
1. This is very clearly argued by Richard Joyce in The Evolution of Morality (Cam-
bridge, MA: The MIT Press, 2006). Some of the following discussions present points akin
to those made in this informed and well-argued book, but, since my pragmatic naturalism
diverges from Joyce’s approach in important respects, the formulations are sometimes
different.
253
In every system of morality, which I have hitherto met with, I have

always remarked, that the author proceeds for some time in the ordi-
nary way of reasoning, and establishes the being of a God, or makes
observations concerning human affairs; when of a sudden I am sur-
prized to find, that instead of the usual copulations of propositions,
is, and is not, I meet with no proposition that is not connected with
an ought, or an ought not. This change is imperceptible; but is, how-
ever, of the last consequence. For as this ought, or ought not, ex-
presses some new relation or affirmation, it is necessary that it should
be observed and explained; and at the same time that a reason should
be given, for what seems altogether inconceivable, how this new re-
lation can be a deduction from others, which are entirely different
from it.2
Hume is commonly (and with good reason) interpreted as making a claim

about inference, one formulated as the denial that you can derive “ought
statements” from “is statements.” This is taken to threaten naturalistic
approaches to ethics because these are allegedly committed to claiming
that normative conclusions can be based upon factual premises.
It is not easy to state Hume’s supposed claim about the impossibility
of certain kinds of warranted inference in a form in which it is both pre-
cise and true. Plainly some statements in which ethical vocabulary oc-
curs are logical consequences of others in which no such vocabulary
occurs—for there are logical truths involving ethical terms, and logical
truths are derivable from any set of premises (including the empty set).
You can evade this trickery by requiring the ethical vocabulary to oc-
cur “essentially,” but cases can be devised in which, although the crite-
ria for essential occurrence are met, a conclusion involving ethical vo-
cabulary is derivable from purely factual premises.3 An example:
premise: some Jews (those who won the Iron Cross in World War I) have
2. David Hume, Treatise of Human Nature, 3.1.1. (Oxford: Oxford University Press,
1978).
3. The notion of “essential occurrence” is developed by writers on logic from Bolzano
on and is especially clearly worked out by W. V. Quine (Philosophy of Logic [Englewood
Cliffs, NJ: Prentice Hall, 1970], 80–81). Counterexamples to the thesis that statements es-
sentially involving ethical vocabulary cannot be derived from factual statements were of-
Naturalistic Fallacies? 255
served Germany with honor; conclusion: if we ought to respect and

protect those who have served Germany with honor, there are some
Jews we ought to respect and protect.4
Moreover, some ethical statements have factual presuppositions, and,
in consequence, statements involving ethical vocabulary (“essentially”)
can be derived from the denial of these propositions. Thus, if “X ought
to do A” implies “X can do A,” then “It is not the case that X ought to do
A” would follow from the factual premise “X cannot do A.”5 More gener-
ally, if E is a statement involving ethical vocabulary (“essentially”), and
E implies some factual statement F, the inference of not-E from not-F
will be logically valid.
It would be sophistry to take these points to address Hume’s chal-
lenge. Whatever the difficulty involved in specifying just which types
of inferences are unwarranted, something is uncontroversially correct
about the Humean worry. It is no victory for naturalism to contend that
some negative ethical statements can be derived from factual premises,
for that fails to touch the fundamental question of whether a naturalistic
approach to ethics can justify conclusions offering positive guidance for
our actions. Hume was reflecting on the “systems of morality” offered by
books he had read, and he observed that the kinds of conclusions they
wanted to reach about proper conduct received only a problematic jus-
tification from the factual premises invoked. The problems are well
illustrated by Darwinian ventures in ethics that postdate Hume’s chal-
lenge. When late-nineteenth-century social Darwinists appealed to the
struggle for existence to justify allowing, or intensifying, competition in
human societies, their inferences were unjustified in exactly the way
Hume diagnosed. A century later, when human sociobiologists and fel-
low travelers proposed that prohibitions against incest can be derived
fered by Arthur Prior (“The Autonomy of Ethics,” Australasian Journal of Philosophy 38

[1960]: 197–206).
4. Troubled by this inference, Nazi leaders hit on the idea of sending Jews who had won
the Iron Cross to Theresienstadt, not as a way station to Auschwitz but as a terminal
destination.
5. The awkward formulation here is important—it would be wrong to suppose that “X
ought not to do A” follows from the factual premise. The negation has to be “external,”
applying to the whole sentence that ascribes the obligation.
from the alleged fact that inbreeding is evolutionarily deleterious, they

were repeating the mistake.6
For the moment, I shall take naturalism to be committed to the thesis
that in some sense ethical conclusions are susceptible of justification.7
Hume’s challenge applies to a simple, historically prominent, way of elab-
orating this commitment. If you think there are external constraints on
ethics, constraints whose sources human beings are struggling to
fathom, and also think we have straightforward epistemic access to (and
only to) factual truths about the world, the justificatory problem will
arise in a way that recalls Hume’s objection. For you will suppose that
people can justifiably arrive at a set of factual claims and that justifica-
tion of ethical statements must be through inferences from these claims.
In our own personal development, or at some point in the history of the
cultural lineage in which we stand, the only justified beliefs are factual;
later we have justified ethical beliefs; these must be based on inferences
from the earlier corpus of facts.
In envisaging some “ethics-free” state, this formulation overlooks the
possibility elaborated in Part I. Pragmatic naturalism’s favored history
views human beings as always having been committed to ethical claims:
they have governed their lives together according to agreed-upon rules
and have believed the descriptive counterparts of those rules (when they
have accepted an injunction to do some deed, they have also believed the
deed to be right, or good, or virtuous). On occasion, modifications of ethi-
cal practice have been effected through inferences, but the clearest cases
6. For social Darwinism, see the classic study by Richard Hofstadter (Social Darwin-
ism in American Thought [Boston: Beacon, 1955]). The most prominent versions of the
sociobiological inferences come in E. O. Wilson, On Human Nature (Cambridge, MA:
Harvard University Press, 1978), and in Michael Ruse and E. O. Wilson, “Moral Philoso-
phy as Applied Science,” Philosophy 61 (1986): 173– 92. For extended critique, see the fi nal
chapter of my Vaulting Ambition: Sociobiology and the Quest for Human Nature (Cam-
bridge, MA: The MIT Press, 1985) and “Four Ways of ‘Biologicizing’ Ethics,” in Concep-
tual Issues in Evolutionary Biology, ed. Elliott Sober (Cambridge, MA: Bradford Books,
The MIT Press, 1994).
7. As the discussion of method in substantive ethics, to be presented in Chapter 9, will
reveal, the sense in which ethical proposals are justified is rather special and involves a col-
lective procedure. For the moment, the complications of this method are ignored, and I shall
operate with a blanket notion of justification that figures in naturalism’s commitment.
from history reveal the premises as partly normative and partly factual.
Wollstonecraft’s premises are (1) women ought to be capable of wifely
and maternal behavior (in line with the prevalent ethical code), and (2)
educated women are more likely to have these capacities. She infers that
women ought to be educated (§24). Her inferences are not touched by
Hume’s challenge.
Merely making this observation is no more satisfactory than appeal-
ing to the fact that the inferences Hume criticizes are hard to circum-
scribe. To say that “from the beginning” human beings (human beings)
have been committed to normative judgments would only release natu-
ralism from Humean suspicions if all subsequent inferences could be
assimilated to the pattern discerned in some of my examples (for exam-
ple, those that figure in the reasoning of Wollstonecraft and Woolman).
Unfortunately, if you suppose all modifications of ethical practice in-
volve logical inferences from previously adopted ethical judgments and
factual statements (typically newly discovered), ethical innovation will be
viewed as an illusion. Clear-eyed champions of the Humean challenge
would wonder how, if there were never any novelty in ethics, it was pos-
sible for human ethical practice to achieve, at the beginning, a strong
enough collection of normative premises, and to do so in a justified way
(§31). If pragmatic naturalism is to scotch concerns about its reliance on
illicit inferences, more will have to be said.
One thing that can, and should, be said immediately is that not all
cogent inference is deductive. From Hume himself on, no sophisticated
devotee of the challenge has believed that the inferences leading from
factual premises to normative conclusions have to be cast in a form re-
vealing them as deductively valid. Naturalists can avail themselves of
nondeductive modes of inference. Pragmatic naturalism can plausibly
suppose that some modifications of ethical practice proceed through a
search for reflective equilibrium: given the normative judgments currently
accepted, one looks for principles subsuming them and perhaps jetti-
sons particular normative judgments that do not accord with candidate
subsumptive principles.8 Recognizing this possibility (and others akin
8. Rawls’s conception of reflective equilibrium, apparent in his “Outline of a Deci-

sion Procedure for Ethics” Chapter 1 of John Rawls Collected Papers (Sam Freeman, ed.)
to it) enables pragmatic naturalism to account for some instances of ethi-

cal novelty, and thus to avoid attributing to the first pioneers possession
of the full complement of ethical resources. Unless it could be shown,
however, how all later transitions can be reconstructed by appeal to re-
flective equilibrium, pragmatic naturalism would still face the task of
explaining the remainder—and, in any event, it would have to show how
the enterprise began in a justified way.
Given these preliminary reflections, the Humean challenge can be
tackled head-on. The worry concerns the justification of the elements of
ethical practice. Naturalists must elaborate an account of the justified
beginning and growth of the ethical project, showing it to be free of the
sorts of inferences Hume questioned, the types of inferences appearing,
for example, in social Darwinism and human sociobiology. Without the
narrative of Part I, and the conception of ethical progress presented in
Chapter 6, any serious response would be impossible—which is why
consideration of those worries has been postponed.
The framework developed takes ethical progress, not ethical truth, to
be the fundamental notion. So, instead of posing the challenge as a ques-
tion about whether inferences from factual statements to normative state-
ments would be likely to yield correct conclusions from true premises,
the issue has to be reformulated. Is it possible to understand how our an-
cestors made progressive transitions, and did so on the basis of processes
(observations, emotional responses, modes of reasoning) likely to pro-
mote progressive transitions?9 An affirmative answer would remove the
sting from the challenge. For it would demonstrate how any inferences
made accord with the fundamental criteria for good inference, and thus
are exempt from the mysteries Hume rightly queried.
(Cambridge, MA: Harvard University Press, 1999), and fully explicit in A Theory of Justice
(Cambridge, MA: Harvard University Press, 1971), is well suited to a dynamic understand-
ing of ethical justification. Interestingly, Rawls’s source for the appeal to reflective equilib-
rium, Nelson Goodman, deployed the idea as part of a naturalistic solution to his own
“new riddle of induction” (see the fi nal chapter of Fact, Fiction, and Forecast [Indianapo-
lis, IN: Bobbs-Merrill, 1956]). I shall discuss the credentials of appeals to reflective equi-
librium at greater length in Chapter 9.
9. Here I draw on my reconceptualization of a reliabilist approach to justification. See
§§32, 39.
Divide the problem into parts. Start with the pioneers, those who be-
gan ethical practice. They belonged to a tense society, in which altruism
failures constantly produced social conflict. If their newly introduced
rules address the problem underlying the original function of ethics (they
fulfill the function of remedying altruism failures), they take a progres-
sive step. Despite my emphasis on the sporadic character of ethical prog-
ress (episodes of “sleepwalking” might occur with significant frequency),
the codes introduced by groups successful in cultural competition dis-
charged this function. It is implausible that their success was entirely
accidental. The first ethical deliberators surely perceived clearly some
sources of trouble—failures to share, unprovoked aggression, and so
forth. Almost certainly, the rules they formulated to address their prob-
lems were imperfect, but they were an advance over the unregulated
state in which they suffered social tension. Imagine, for example, early
ventures in regulating alliances and mating, rough-and-ready delinea-
tions of “the elementary structures of kinship,” adjusting the conduct of
group members so it became more frequently behaviorally altruistic and
less likely to provoke trouble. We cannot suppose that deliberators who
recommended progressive transitions were entirely clear about what
they were doing—they were not in the position of later analysts who can
identify functions and show how they would be promoted by proposed
changes—but the successes of these early ethicists did not emerge from
blind guesswork either. In the beginning, the successful pioneers made
ethical progress through processes (diagnosis of prevalent social prob-
lems, joint deliberation) likely to generate progressive transitions.
Their successors expanded the conceptual framework to introduce
ideas (of good, of right, of virtue—or of what it is to be “one of us”) en-
abling them to express descriptive counterparts of the rules previously
adopted. They have a rule enjoining or forbidding a class of actions, and,
within the expanded framework, they declare that actions in this class
are right or are wrong (what members of the group do, or avoid). If adopt-
ing the rule was based on deliberation likely to generate progressive
transitions, promulgating the rule is justified—and so derivatively is ac-
ceptance of its descriptive counterpart.
Turn now to the subsequent modifications. As emphasized (§§36–37),
the evolution of ethics brings new functions beyond the original one.
Hence, at later stages the deliberations, arguments, and observations

prompting new transitions should be evaluated by seeing how well they
serve the overall set of functions ethical practice is to serve. In the simplest
cases, the processes through which transitions are effected refine some
functions without compromising others and can be vindicated as of types
likely to do so. Thus, the extension of some rules, already protecting
group members, to cover trade with others, produces satisfaction of en-
dorsable desires for group members, as well as diminishing conflict with
neighbors; the extension can result from perception of likely advances in
fulfilling these functions, so that processes bringing about the transition
are well adapted to generate ethical progress. As with the original case,
we can recognize the justification of the modification—and derivatively
of the descriptive counterparts of the extended rules.
More difficult are cases in which one function is served at cost to oth-
ers, or a modification of ethical practice threatens future regressions.
Consider, for example, the introduction of the unseen enforcer as a de-
vice for promoting compliance. Insofar as the modification increases the
frequency with which members of the group obey the prescriptions, this
is a progressive step, but it paves the way for damage to the project of joint
deliberation (some group members have direct access to the divine will—
§35). In such instances there is a propensity both for progress and for
regress. The transition cannot be forthrightly classified as justified. In
general, when there are gains and losses with respect to different func-
tions, three possibilities arise: (1) because the balance is significantly
greater on one side (the gains are much larger than the losses), the modi-
fication is overall progressive (or regressive); (2) although there is no
overall verdict, the modification can be partitioned, and some newly in-
troduced elements make progress, while the rest are regressive; (3) the
situation is so thoroughly mixed that neither an overall judgment nor a
recognition of progressive and regressive aspects is possible. Instances
of the first two types can be assimilated to instances in which there is no
functional conflict, by recognizing that the change in practice is justified
as a whole (case 1), or by separating out justified parts of the change (case
2). With respect to transitions of type 3, the change is not justified.
Let us differentiate standards of progressiveness and of justification.
So far, in responding to Hume, relatively weak, short-term criteria of
progress and of justification have been used. We can, however, mimic

the distinctions between local and global progress (§37). Weak concep-
tions of progress and justification can be formulated explicitly:
(WP) A change is progressive just in case it leads to greater fulfill-

ment of those functions of ethical practice that have emerged
at the pertinent stage.
(WJ) A change is justified just in case it is generated through pro-
cesses likely to lead to progressive changes (as specified by WP).
These criteria make allowances for the circumstances in which those

who modify their ethical practices find themselves, focusing on the func-
tions ethics serves for them and on the immediate effects of the transi-
tions. It is possible to demand more.
(SP) A change is progressive just in case it introduces elements that

will be preserved in any indefinitely extended sequence of
WP-progressive changes.
(SJ) A change is justified just in case it is generated through pro-
cesses likely to lead to SP-progressive changes.
Parts of the history of ethics proceeded according to WJ-justified transi-

tions. Although we cannot be sure our ancestors always saw how the rules
they proposed addressed the functions ethical practice served for them,
in some instances the difficulties they faced would have been obvious
(lessening conflict within the group, dealing with hostile neighbors) and
their deliberations would have been overwhelmingly likely to follow reli-
able procedures for dealing with those difficulties: we can expect them
to recognize how rules of alliance would work, how offers to trade peace-
ably with neighbors would satisfy desires endorsed by their existing rules.
The standards for weak justification were probably met quite frequently
in the progressive transitions our ancestors made.
Can a stronger conclusion be defended? Perhaps. Consider our pre-
dicament, if, as pragmatic naturalists, we analyze what has occurred.
Honesty exemplifies the “vague core themes” that serve as the best can-
didates for ethical truth (§38; see also §46). Analysts can offer lines of
reasoning for thinking some elements of our current practice—like rules

enjoining honesty—would be preserved in any indefinitely proceeding
WP-progressive sequence of transitions. The injunction to truthfulness
remedies a class of altruism failures that will continue to be important
to us as long as human beings communicate with one another. By iden-
tifying permanent human needs, analysts can conclude that some ele-
ments in ethical practice will probably figure in any progressive devel-
opment of what we now have. When they do that, their judgments are
SJ-justified.10
Because of the complications just noted, it is worth stating explicitly
how Hume’s challenge has been met. Hume worries that any naturalist
account of ethics is committed to forms of inference incapable of justify-
ing the ethical conclusions obtained. Pragmatic naturalism understands
notions of ethical truth and justification in terms of the fundamental no-
tion of progress, conceived as functional fulfillment and refinement. Intro-
ducing ethical novelties, whether at the beginning of ethical practice or
in subsequent modification, is justified when those who make the change
do so by following processes likely to lead to better functional fulfillment.
Any inferences they make are thus evaluated by appeal to fundamental
criteria, which either may make allowances for their historical circum-
stances (WJ) or may consider the indefinite evolution of the ethical proj-
ect (SJ). Many of the processes through which historical actors made the
changes they did are likely to satisfy the less demanding criterion, and
later analysts can meet the more exacting standard (SJ) for some (perhaps
all) core themes in contemporary ethics.
The key to this response is quite simple. Once ethics is viewed as a
social technology, directed at particu lar functions, recognizable facts
about how those functions can better be served can be adduced in in-
ferences justifying ethical novelties. The mystery that worried Hume
disappears.
10. In making the particu lar claim about honesty, I offer no judgment about whether the
line of reasoning that counts as SJ-justified would have been available to our predecessors.
It is possible (but not obviously correct) that SJ-justification requires explicit articulation
of a pragmatic naturalist perspective. If so, it does not detract from the WJ-justification I
take to be present throughout the history of ethics.
§41. Authority Undermined?

Replying to Hume in this way may exacerbate other concerns. If ethics
is a human construction, devised to discharge particular functions, can
it have the authority usually attributed to it? The question expresses
another version of the charge that naturalism commits a fallacy. The
first task is to understand what the challenge might mean.
Pragmatic naturalism offers a simple explanation of why people feel
the authority of ethical prescriptions. Each of us is born into a society
that inculcates a body of lore, and, because each of us has a capacity for
normative guidance, more or less cleverly stimulated by processes of
ethical education, we find ourselves, as children and as adults, feeling a
tug to align our will in a particular direction. Accounting for the author-
ity of ethics might proceed not by adducing some metaphysical claim
about the status of ethics, but by pointing to sociological and psycho-
logical facts. A sense of what is required of us (“the internal sanction of
duty”) stems from a cluster of capacities and dispositions with which we
have been equipped (“Conscience,” or a “mass of feeling which must be
broken through in order to do what violates our standard of right”).11
Perhaps parts of this subjective basis derive from some innate feeling,
while others are “implanted.”12
11. The quoted phrases are from Mill, who favors explaining the “sanction of ethics” in
sociopsychological terms. See Mill, Works (Toronto: University of Toronto Press, 1970)
10:229, 228.
12. Mill is inclined to the latter explanation: see Works 10:230. In accordance with
the discussion of §14, I recommend framing the questions in terms of interactions be-
tween genotype and environment, and remaining open about the precise contributions
of each.
Mill offers an eloquent formulation of the picture he views as most plausible. Our coop-
erative social existence arouses this feeling, and it is given direction by the community
around us:
The smallest germs of the feeling are laid hold of and nourished by the contagion of
sympathy and the influences of education; and a complete web of corroborative as-
sociation is woven round it, by the powerful agency of the external sanctions.
(Works 10:232)
As I read him, he tries to address the question of the authority of ethics (the “sanction of
morality”) by appealing to something like the account of normative guidance and of con-
science offered in §§11–13.
Does a sociopsychological approach “evacuate [ethics] of all author-

ity”? 13 The concern is clear. Explaining felt authority is no problem
for naturalism (specifically for pragmatic naturalism), for the (philo-
sophical) history it offers attends to the psychological processes un-
derlying our obedience to the ethical precepts we learn. That is not
enough: felt authority is different from real authority. Naturalists shift
the question and fail to provide any explanation for why ethical pre-
scriptions have real authority over us. Instead of recognizing the au-
thority of ethics over our social arrangements, naturalists focus on
the authority of our social arrangements over individual conduct and
judgments.14
What exactly does the objector understand by “the authority of ethics”?
At first glance, it would appear that the authority of anything has to
be mediated by psychosocial processes, for how could one’s will be
forced in a particular direction unless there were some such causal rela-
13. The quoted phrase is from Dewey, who extends Mill’s sociopsychological approach.
He confronts what is supposed to be a great difficulty:
It is said that to derive moral standards from social customs is to evacuate the latter
of all authority. Morals, it is said, imply the subordination of fact to ideal consider-
ations, while the view presented makes morals secondary to bare fact, which is
equal to depriving them of dignity and jurisdiction. (Human Nature and Conduct,
79)
Dewey claims that the worry rests on a “false separation,” the supposition that cultural
practices are merely “accidental by-products” (ibid.). Here he recapitulates the point made
in his summary of his ethical perspective, that “moral conceptions and processes grow out
of the very conditions of human life” ( John Dewey and James Tufts, Ethics, 2nd ed. [New
York: Henry Holt, 1932], 343). This prepares the way for his answer to the challenge. After
arguing that, strictly speaking, no view of the “origin and sanction” of moral obligation can
provide what some people seek for the authority of ethics, he suggests that, in “an empirical
sense” there is a simple answer: “The authority is that of life” [Human Nature and Con-
duct, (Amherst NY: Prometheus Books, 2002), 80–81; see also 98, 232, 326]. My approach
to the objection that naturalism loses the authority of ethics will endeavor to articulate
what I take Dewey to have had in mind.
14. Some of Mill’s formulations make him vulnerable to the charge that he has changed
the topic in this way. Yet that accusation is a little unfair, since the discussion in Utilitari-
anism, in Mill On Liberty and other Essays [Oxford: Oxford University Press (The World’s
Classics), 1998], does make some efforts to show how the conditions of community life will
orient our subjective feelings (conscience) in particu lar directions—specifically in leading
us to view our own preference and perspective as one among many. It is easy to see, how-
ever, how my emphasis on processes of socialization provokes the complaint.
tionship? Lurking in the background, however, is the view of ethics as

a form of inquiry subject to external constraints. Ethics is seen as
answerable to something outside (or beyond?) us, as individuals and
as groups, something to which our conduct and our social practices
ought to conform. Even if we honor the authority we feel, complying
with the directives laid down for us in the society in which we live, that
society may not be aligned with the external constraints: mistakes may
have been made.
An analogy with other parts of human culture helps. Consider the
formation of belief, in everyday deliberations and in scientific investiga-
tions. As individuals, or collectively, our inquiries can go astray, failing
to measure up against an external standard: facts, one might say, are
authoritative over beliefs. This slogan is easily overinterpreted. Con-
sider the following readings:
(a) For any true statement p one ought to believe p.

(b) For any true statement p one ought not believe any statement q
that is incompatible with p.
Plausible as they may initially appear, both a and b are false.

There are at least three sources of trouble. First, there are far too many
true statements for finite animals like ourselves to believe all of them.
Our world could be described in any of some nondenumerable infinity
of potential languages, and within each of these languages there are
infinitely many statements describing the course of events within some
small spatial region (a room, say) during some short temporal period (for
example, an hour). Only a truly tiny minority of such statements is
worth bothering about, only an infinitesimal fraction of them has any
significance for any human being—where the standards of significance
are set by human needs, interests, and projects.15 We can rightly criti-
cize people for investigating the wrong things (or posing the wrong ques-
tions). Even true answers to irrelevant questions do not merit belief,
contradicting a.
15. For discussions of this theme, see chaps. 5 and 6 of my Science, Truth, and Democ-
racy (New York: Oxford University Press, 2001).
Second, it is sometimes impossible for us to achieve our purposes by

trying to discover and work with the exact truth. Much of our practical
use of “knowledge” is grounded in accepting simplified versions of the
world, approximations to the truth rather than the genuine article. When
the phenomena are too complex for us to render accurately, and when an
accurate rendering would only interfere with our aspirations to predict,
control, and intervene, we are better served by believing something that
is not, strictly speaking, true. On such occasions, both a and b are false:
we ought not believe the exact truth and we are entitled to believe state-
ments incompatible with the exact truth.
Third are considerations about the limits of honesty (§46). Seriously
ill patients are sometimes better off believing false statements. Not only
are they not obliged to believe what is true about their condition, but
they can permissibly believe things incompatible with the true descrip-
tion. Once again, both a and b fail.
The common trouble lies in human purposes ill served by true belief.
Proponents of the “authority of facts” take attaining truth as an ultimate
standard. It is not. More fundamental is the role of inquiry and belief in
advancing human concerns, and the cases briefly surveyed show how,
occasionally, promoting such concerns can come apart from acquiring
true belief. When it does so, the concerns set the more basic standard
and thus release people from any obligation to believe the truth. Of course,
much of the time, even almost all the time, our human purposes are best
served by discovering and believing what is true—that is why a and b are
plausible and why we easily talk of the “authority of facts.” From a natu-
ralistic perspective, however, one should probe more deeply, exposing
the purposes having true beliefs typically serves, and viewing those
purposes as constituting the fundamental standard from which author-
ity arises.16
The analogy prepares for a similar reconnection in understanding the
“authority of ethics.” We have reviewed the idea of external constraints
that “prescribe” to us in a fashion similar to that in which the world is
supposed to “prescribe” to our beliefs (§§29–31). Critics might take the
16. This is to overcome the “separation” of which Dewey was suspicious, and to recon-
nect the standards of inquiry with human life.
pragmatic naturalist mistake to lie in abandoning any external standard.

Or they might believe pragmatic naturalism cannot find a counterpart
for the “constraints on practical reason” by which we are bound.17
As with Hume’s challenge, the key is to recognize the effect of substi-
tuting progress for truth as the fundamental notion. The objection can be
cast as follows. When naturalists explain the authority of ethics, they
point to contingent ways in which societies formulate rules for their mem-
bers and the contingent processes through which they socialize people
into subordinating their behavior to such rules. The real authority of eth-
ics consists in the fact that certain prescriptions are binding on our con-
duct, whether or not they figure in an ethical code and whether or not
ethical practice includes effective ways of inducing conformity to its pre-
cepts. If naturalism is to succeed, it must differentiate felt authority from
the genuine article and show why particular rules are authoritative.
By appealing to the functions of ethical practice, pragmatic natural-
ism can directly answer the first part of this demand. There is no diffi-
culty in showing how people who follow the ethical code of their com-
munity, in cases where the critic would view the code as diverging from
ethical truth, are mistaken. Consider the following possibilities:
(i) Someone brought up in a society conforms to its ethical max-

ims, maxims that would be replaced in a progressive modifica-
tion of the ethical practice.
(ii) Someone brought up in a society conforms to ethical maxims
that would be incorporated in a progressive modification of the
ethical practice and does so through processes likely to promote
progressive alterations of practice.
(iii) Someone brought up in a society conforms to ethical maxims
that would be incorporated in a progressive modification of the
ethical practice but does so through processes not likely to
promote progressive alterations of practice.
17. Kantians will incline to the latter formulation, resisting any suggestion of an inde-
pendent ethical reality to which our judgments should conform. Their approach to objec-
tivity emphasizes constructions taken to have a priori validity, and they will argue that
pragmatic pluralism can supply only contingent—second-rate—constraints.
Possibility i focuses on subjects who conform to what they have been

taught, even though progressive change would replace the precepts on
which they rely—here there is a clear distinction between “felt author-
ity” and a standard to which the subjects (and their society) are not
living up. Possibility ii, by contrast, reveals those who comply with
the divergent standard and do so in a justified fashion—those whom we
might see as appreciating the “real authority” of ethics and conse-
quently as overriding the felt authority current in their community.
That type of reformer is to be distinguished from the characters who
emerge in possibility iii, who base their conduct on a progressive shift,
even though they have no justification for taking it to be so; like the
subjects of possibility i, they exhibit shortcomings, and it would be
appropriate to view them as failing to appreciate the “real authority”
of ethics.18
Pragmatic naturalism can represent a difference between people who
submit to local authority in an unsatisfactory way and those who over-
come the shortcomings of their peers. Moreover, it can explain that dif-
ference in terms paralleling the critics’ preferred accounts. For the critics
suggest that inadequate conformity to the prevailing code is separated
from appreciation of real authority, in virtue of the fact that the apprecia-
tors acknowledge external constraints—ethical truth or the commands
of practical reason. Pragmatic naturalism sees these formulations as mis-
guided ways of trying to resist the mere-change view and substitutes
its preferred account: the appreciators make progressive modifications
(better fulfilling ethical functions), and they do so through processes
disposed to yield progressive changes. On the basis of the parallel, prag-
matic naturalism proposes that the demand for this sort of progressive
change constitutes the “real authority” of ethics.
Although this rebuts the criticism in its original form, it may not sat-
isfy critics. For they may declare that the naturalistic account introduces
18. The notions of progress and justification that figure here can be construed accord-
ing to either of the conceptions that figured in the last section, as weak or strong. The
strong concepts provide a close analog of the split between conformity to local authority
and the recognition of ethical truth, on which critics want to insist, but the weak concepts
mark another important distinction, one accommodating our fallibility and the historical
contexts in which we fi nd ourselves.
a contingent element, one failing to capture the absolute authority of eth-

ics. So long as the functions of ethics, the original function of remedying
altruism failure and the derivative functions emerging through the his-
torical ways of fulfilling that original function, are endorsed, it is possi-
ble to recognize the authority of ethics. Those functions could be re-
jected in a way in which it is not possible to reject the realm of ethical fact
or the demands of practical reason. The analogy with belief formation
exposes the critics’ error. The thesis that “facts have authority over be-
lief” must be understood, in the end, as a claim about the functions our
beliefs serve: strictly speaking, a and b are false, and their import can be
properly recognized only by seeing the acquisition of true beliefs as typi-
cally important to the functions of belief formation. In similar fashion, if
the “absolute authority” of practical reason were anything more than a
dogmatic pronouncement, it would rest on showing the role of practical
reason in discharging human functions. The accounts proposed by prag-
matic naturalism and by the critics are on a par. If any of them can de-
liver an explanation of the “real, absolute authority” of ethics, pragmatic
naturalism can do so. If something more is wanted, all are deficient—
and it will be reasonable to judge that the sort of authority demanded
is a myth.
In portraying naturalism as introducing a “merely contingent” au-
thority for ethics, one depending on endorsement of functions that might
be rejected, the criticism does make a “false separation”—for it fails to
appreciate the connection between the problems behind those functions
and the circumstances of human life (§39). The demands leading to eth-
ics are not arbitrary or conventional; they grow out of human needs. To
what higher—or less contingent—standard could one appeal to ground
the authority of ethics?
§42. Troublesome Characters

Surely a major issue remains. Although parallels are provided, and
explanations of the “authority” of ethics offered, traditional (nonnatural-
istic) conceptions of ethics may seem to do something naturalistic sur-
rogates cannot. They can answer the troublesome characters—skeptics,
egoists, nihilists—who crop up here and there in the history of ethical
discussions.19 In relatively polite versions, these troublemakers have al-

ready surfaced in earlier sections, for example, in considering a subjec-
tive criterion for ethical progress (§28). They conceded that particu lar
transitions might be labeled as “progressive” but saw this as merely
providing a sense of false contentment. They level a basic challenge:
once ethics has been seen as the evolving enterprise pragmatic natural-
ism describes, why should anyone conform his or her behavior to its
deliverances?
So far, the issue has been framed as comparative. Allegedly, there are
questions pragmatic naturalism cannot answer, but some other (nonnat-
uralistic) conception of ethics can. Naturalistic proposals are inarticu-
late, or at least unconvincing, in addressing skeptical challenges to which
nonnaturalism can provide fluent responses. My main aim is to show that
anything rivals can do, pragmatic naturalism can do just as well (if not
better).20
Start by disambiguating the skeptical question. The challenger
wants to know why he should conform to “the deliverances” of ethics,
given their emergence from a historical process. What exactly are these
“deliverances”? Skeptics often present their demands as questions to the
authority of current codes, but an evolutionary account will not sup-
pose all the rules included in such codes are to be followed. We can
distinguish:
(a) rules present in current codes

(b) rules that would be sustained in an indefinite sequence of
progressive modifications from current codes
(c) rules we are justified in thinking of as likely to be sustained in an
indefinite sequence of progressive modifications from current
codes
19. I shall not attempt any fi ne distinction between skepticism and nihilism: my focus
will be on a range of questions that might be covered with either label. Egoists will be con-
sidered as skeptics/nihilists who view ethics as a constraint on their personal (solitary—§3)
desires.
20. As we shall see, discussion of the comparative issue leads to a deeper challenge, one
already identified at the end of the last chapter. I am indebted to an anonymous reader who
recommended being explicit about this.
Even this basic differentiation allows various ways of reading the skepti-
cal query.
Nonnaturalistic approaches to ethics clearly suppose that the skeptic
is raising a question about some true ethical precept. Nobody thinks a
skeptic who inquires why he or she should conform to the injunction to lie
whenever one feels like it deserves any more than correction about what
ethics commands. So a very straightforward reading of the challenge
would adopt b as indicating the precepts with respect to which skepti-
cism is to be addressed. That formulation, however, might seem to un-
derestimate human fallibility. We know we do not have any ultrareliable
access to the rules that would be adopted in any indefinite sequence of
progressive modifications, and so, perhaps, we should take ourselves as
reasonably conforming to the rules our best justifications single out
as likely to have the status assigned in b—and take skeptical questions to
be directed against c. Or we might go further, viewing our individual
selves as unlikely to improve on the wisdom of the tradition in which we
stand, and modestly disclaiming our own abilities to make judgments
about future progressive modifications. Consequently, we might take
the rules adopted in our current code as the best available candidates for
b, supposing the skeptical challenge focuses on a. Even formulating the
skeptical worry turns out, ironically, to involve a judgment about the
allowances we ought to make for our acknowledged fallibility.
As before, it is not enough to point to inexact or problematic formula-
tions of an objection. However you circumscribe the queried rules, the
challenge can be posed with respect to examples. Consider any likely
candidate, and imagine the demand directed at it: given that injunctions to
honesty (say) emerge as stable elements in all progressive traditions, why
should that provide grounds for compliance? Skeptics want to know
why a sequence of progressive transitions, conceived as addressing the
functions of the ethical project, should deliver rules to be obeyed, even
when the rules in question seem overwhelmingly likely to be maintained
in all future progressive modifications of current practice.
What conditions should an answer to the skeptical question meet?
One thought is that, to be satisfactory, a response should silence the skep-
tic. Given the response, nothing can be said to continue the skeptical
challenge. Plainly, that sets the standards of success very high, and, in
line with viewing the issue as comparative, we should begin by asking if

the silencing criterion can be satisfied given any understanding of ethics.
Some well-known conceptions clearly fail. Consider the reductive nat-
uralism often attributed to utilitarianism, according to which one defines
the good as that which promotes the greatest total happiness. A familiar
claim about any reductive definition of an ethical concept—the good or
the right, say—is that it leaves a question open: given the defi nition,
you can always ask, “Why should I act in a way that conforms to the
definition?”21 The naturalistic reduction of the good thus fails to address
the skeptical challenge, precisely because it does not silence the skeptic.
Do things go better on alternative pictures? Consider some approaches
found in or inspired by Kant. Ethics expresses the requirements of pure
practical reason: to deny the moral law reason generates is to fall into a
mode of irrationality, in which one contradicts oneself. Or it might be
said: ethics consists in a set of principles ideally rational agents would
agree to under ideal circumstances, so failing to abide by its precepts is
to violate conditions of rationality. Tough-minded skeptics would hardly
be brought to silence by either of these dicta or by any plausible emenda-
tions of them. The skeptic speaks: “You can call the procedures you use
to generate the rules you favor ‘pure practical reason,’ if you like, and
suppose those who don’t go along with them are involved in some sort of
contradiction, but the mere label doesn’t frighten me, and the effects you
envisage don’t appear particularly dreadful. If I reject these rules, I am
hardly doing something similar to asserting a statement and its negation—
and even if I were, it’s not obvious anything I care about would be com-
promised by doing so. The history of science contains episodes in which
people have worked quite well with internally inconsistent ideas (think of
the Bohr model of the atom22), and I have no reason to think my ‘practical
21. The “open-question” argument, in just this form, is usually attributed to Moore. But
a close reading of what Moore actually says reveals a much more complex and rather differ-
ent line of reasoning. For good discussion, see Joyce, The Evolution of Morality (Cam-
bridge, MA: The MIT Press, 2006). Whatever the merits of the interpretation of Moore,
the familiar objection is an important one.
22. As has frequently been observed, the conception of discrete orbits is at odds with
the principles of electromagnetism. Another interesting example is the concept of the rigid
body in relativistic dynamics.
irrationality’ will pose difficulties for me. Nor am I much moved by the
thought of rules hypothetically ideal people (supposedly better than me)
in some fictional situation would agree to. Why should I be bound by
what they would decide? There is no argument for thinking the pur-
poses I care about would be ill served by flouting any such precepts.”
Orthodox Kantians or contractarians may think this alleged reply is
incoherent, tacitly self-undermining, or something like that. Yet however
hard they struggle to reveal the skeptic’s irrationality, they do not render
him speechless—he can cheerfully wave away the gloomy descriptions of
his state by pointing out how well he is able to pursue the things that
matter to him. Silencing is hard to do.
Maybe, though, the skeptic’s answers have exposed something wrong
with him: he has failed to meet ideal conditions of rationality. If that is
so, the criterion for a successful reply to the skeptic is modified. He does
not have to be silenced; one must merely have an account of why his re-
sponses are problematic. Given this understanding, however, pragmatic
naturalism can do just as well as the allegedly superior nonnaturalistic
approaches. Where Kantians and contractarians see failures of ideal
rationality, pragmatic naturalism diagnoses an inability to appreciate
how central the ethical project is to human life. Pragmatic naturalism
will begin answering the challenge by pointing out that the achievement
of normative guidance was central to the origin and development of fully
human society, that ethics served an important original function, and
that progressive shifts in ethical practice consist in fulfilling that func-
tion, and those generated from it, more effectively. Once again, the skep-
tic speaks: “Why should I be bound by the rules emerging from ‘pro-
gressive transitions’ in this ‘project’? No matter how they have helped
discharge the functions you identify, I can manage perfectly well by flout-
ing them. There is no compelling motivation for me to continue in any
ethical tradition.” Pragmatic naturalists will want to insist that desires
the skeptic wants to satisfy by breaking the rules have been made possi-
ble only by the project he rejects, that he fails to understand how the ori-
gin and evolution of ethical practice have framed his life, giving him the
options he wants to pursue. That insistence will not silence the skep-
tic—it will be no more effective than appeals to practical reason or charges
of irrationality. Yet it fares no worse as a diagnosis of the skeptic’s
mistake: he wants to reject something that has made his envisaged way
of life and his preferred choices possible.
This judgment can be elaborated further and defended by consider-
ing some familiar characters from the history of philosophical ethics.
Thrasymachus challenges Socrates, suggesting that ethical principles
(more exactly, the claims of justice) are put forward to advance the inter-
ests of those in power.23 He views ethical practice as a device employed
to keep weaker members of the society in line and refuses to go along.
Pragmatic naturalism credits Thrasymachus with an insight and faults
him for an error. The insight: in many ethical traditions, rules are intro-
duced or emended by a powerful minority, supposedly especially good
at recognizing the will of the unseen enforcer, and the precepts of the
social practice can be arbitrary and oppressive (§§35, 26). The error is to
overlook the fact that pragmatic naturalism diagnoses the oppressive
injunctions as stemming from regressive transitions. Functional ethical
practice is not a tool for asserting the will of the strong and mighty, but
rather grounded in attempts to take into account the desires of all mem-
bers of a society. The original function of remedying altruism failures
acknowledges the wishes and aspirations of all. Thrasymachus can be
enlisted as an ally and invited to continue in the evolving project of eth-
ics by responding to places at which it is dysfunctional. He is wrong in
his characterization of the project as a whole.
A different challenge comes from a later figure, the “sensible knave”
who pops up in the final paragraphs of Hume’s Enquiry Concerning the
Principles of Morals.24 Knave does not offer any general theory of what
ethical practice is and how it has gone wrong—he simply wants his own
plans not to be constrained by it. He has egoistic (solitary) desires con-
travening some of the ethical principles in force, and, where he can get
away with it, he would prefer to satisfy those desires, instead of conforming
23. See Plato Republic Book 1. Socrates replies with a convoluted and unconvincing se-
quence of questions that eventually reduce Thrasymachus to silence. Later, Glaucon and
Adeimantus present the challenge in a milder form, eliciting a far more sophisticated and
interesting account of justice. Whether an adept real-life counterpart of Thrasymachus
would really be rendered speechless by either version of the response is highly dubious.
24. David Hume, Enquiry Concerning the Principles of Morals (Indianapolis, IN: Hack-
ett, 1983), 81.
to the principles. Knave has no interest in persuading others the rules

need reform—in many instances, he is delighted others are prepared to
follow them and happy to be a “free rider.” What can be said to bring
him into line?
Pragmatic naturalism can start with the points Hume makes, in his
quick, almost casual, attempt to deal with the problem. 25 We can inform
knave that he will be disturbed and worried, that he will forfeit that tran-
quility of mind good conscience bestows. We can try to show him he
will not actually achieve the things he wants if he pursues his knavish
tricks. In many instances, these responses will be ineffective. Knave cor-
rectly appreciates that the chances of detection are tiny, that he is the
sort of person who can sleep serenely after his knavery is done, and that
his knavery does not interfere with his other goals. Nor does it help to
tell him he is exhibiting practical irrationality—for he will describe the
ends he has in view and point out how they are attainable even if he
is “practically irrational.” The specific point pragmatic naturalism will
make focuses on his rejection of the very practice that has made his
knavish aspirations possible.
What exactly does knave want to do? If his aim is to satisfy certain
basic desires—the kinds of desires he shares with hominid ancestors
who preceded, and human ancestors who began, the ethical project—
he deserves help and sympathy. If the only way for him to satisfy those
desires is to break the prevalent precepts, something has gone badly
wrong, either with the rules themselves or with the conduct of others. A
knave in desperate straits (Jean Valjean) is no knave at all. Hume’s char-
acter is not in this plight: he simply wants to increase his wealth, his
power, or his position in ways forbidden by ethical rules, even rules em-
bodying those core themes that are the best candidates for ethical truths.
Pragmatic naturalism points out that the aspirations knave harbors are
possible only in virtue of the ethical project that began with the acquisi-
tion of socially embedded normative guidance, and dependent on just the
rules knave proposes to disregard. If our ancestors had not instituted
those rules and found sufficiently effective ways of making their fellows
comply with them, the options he wishes to pursue and the goals he hopes
25. Ibid., 81–82.

to attain would not be available to him. Knave replies that this is all past
history. He is grateful so many of his predecessors went along with the
codes in force and modified them in the progressive ways that have made
contemporary society possible—although he hints that there have prob-
ably been a few of his sort before, people who have quietly made use of
the obedient docility of their fellows. Now that ethical practice has made
a rich and complex society possible, he can take advantage of that fact
and pursue his egoistic goals.
Knave cannot be silenced. Yet pragmatic naturalism can offer a diag-
nosis of what he is doing. He wants to take advantage of the products of
social evolution without acknowledging the functions that have made
those products possible. He wants to operate within a society without
feeling that sympathy to others, that altruistic response to their desires,
whose evolution has formed that society. Being human, we tell him, con-
sists in participating in this project through which altruism failures are
remedied and further ethical functions generated and fulfilled. He shrugs
his shoulders, unmoved by this rhetoric. Yet the diagnosis seems no worse
than that offered by the major rival approaches to ethics. Indeed, one
way of elaborating the notion of practical rationality might suppose that
the knavish incoherence consists in making an exception of oneself while
simultaneously relying on the ethical practice that sustains human coop-
eration, a reading that would erase the differences between the naturalis-
tic and nonnaturalistic alternatives.26
The third and last troublemaker is a Nietzschean persona, the “free
spirit.” Unlike Thrasymachus, free spirit does not want to provide a gen-
eral characterization to convince everyone to view ethics as oppression.
Unlike knave, free spirit is not concerned only with himself and with his
ability to use institutions presupposing the ethical project to advance his
own solitary ends. Free spirit writes for his peers. They are oppressed,
confined, by ethical practice; how matters stand with the rest, with the
herd, is of no concern to him.
26. Pragmatic naturalism does point out a kind of “contradiction” in knave’s attitudes—
and Kantian-inspired approaches might identify this as the crucial failure of practical ratio-
nality. If so, the accounts would be not only on a par, but also almost identical. The residual
difference would lie in the ability of pragmatic naturalism to point to the failure of sympa-
thy knave exhibits.
The charge leveled by free spirit is so far ambiguous. Free spirit might
be rejecting the ethical codes actually developed (or developed in a par-
ticular tradition), seeing these as failing to fulfill important human func-
tions, and, on this basis, demanding a “revaluation of values.”27 Viewed
in this way, free spirit is a reformer, a participant within the evolution of
ethics, not someone who rejects the rules it would deliver if it were to
proceed progressively. He calls attention to particular places—possibly
quite fundamental places—at which he takes the historical development
of ethics to have been regressive. Assuming we can reach agreement with
him about whether the pertinent modifications refine the functions of
ethics, there will be no difficulty.28 For, if the transitions turn out to
be progressive after all, he will acknowledge the rules. If they do not, we
shall not insist on his acknowledgment of the rules actually generated,
and he will honor the prescriptions resulting from progressive replace-
ments of the dysfunctional precepts.
Free spirit may want more, however. He asks why he should care
about the specific recommendations emerging from those transitions
pragmatic naturalism counts as progressive. His question is best met
with another: what alternative does he have in mind? To conceive of the
historical evolution of ethical practice, taken as a whole, as oppressive
is vacuous, unless one can do more than wave vaguely in the direction of
unarticulated possibilities. The more limited version of free spirit, who
is willing to acknowledge the functions of ethics and who clamors for
27. I think most of Nietzsche’s many-sided polemic against “morality” can be inter-
preted in this way. So understood, he turns out to be an interesting and insightful ally of
the pragmatic naturalist project. The derogatory remarks about “English” genealogies,
and the attacks on the value of altruism, [see the early sections of Genealogy of Morals,
Friedrich Nietzsche On the Genealogy of Morality (Cambridge: Cambridge University
Press, 2007)], are fundamentally attacks on the thesis that a historical understanding of the
emergence of contemporary ethical ideas must vindicate those ideas. Instead, Nietzsche
wants to use history in the interests of reform—with the aim, one might say, of advancing
the ethical project. That is entirely in accord with pragmatic naturalism, which is receptive
to the thought that increased historical understanding might expose regressive transitions
and open up new possibilities for us. I am grateful to Jessica Berry, whose insightful com-
ments and questions in response to a lecture I gave at Emory University have led me to view
Nietzsche as less of a threat and more of a fellow traveler.
28. The next section will begin to scrutinize this assumption.
reform so those functions will be more effectively fulfilled, is readily

comprehensible, for he is committed to continuing the progressive evo-
lution of the ethical project. To reject that project entirely is to take a
further step, and we need to be told what free spirit aspires to.
A life cut off from any society, perhaps? That, as Nietzsche sees
clearly, would be dangerous and brutish.29 But what forms of alternative
community or society are available, once the ethical project is renounced?
We do know something of social life outside the tradition of ethical prac-
tice, for a life of this sort is the lot of our evolutionary cousins, the
chimpanzees. Given the psychological dispositions free spirit has ac-
quired, this cannot be a serious possibility for him. Can he offer another?
If not, our response to the challenge is, in the end, a simple one: operating
within the evolving practice of ethics is a central part of our humanity.
Until we are given some description of an alternative—or until the Über-
mensch actually arrives—our choices are confined to the human, the
ethically guided, life and the social state of the chimpanzees, a state
transcended by our first human ancestors.
Three famous troublemakers are enough. They have been taken seri-
ously, as if the task were to find arguments to bring them into line. A
moment’s reflection reveals this to be a distortion of the actual task.
Were we faced with actual figures for whom resistance to ethical pre-
cepts was a live option, it would be foolish to rely on philosophical de-
bate alone. We know too well that ethical injunctions, even those most
stable and central to the main ethical codes, are often ignored or flouted,
and only the most otherworldly of philosophers would think sufficiently
well-crafted arguments would bring potential offenders into compli-
ance. In societies with a high incidence of failure to obey their ethical
precepts, it is usually appropriate to recognize deficiencies in techniques
of socialization and to seek to improve them. Yet some people are beyond
the reach of even the most effective methods for inculcating ethical atti-
tudes. Some real-life knavery results from inadequate education; some
requires sterner measures.30
29. Nietzsche, Genealogy of Morals, 2:9.

30. This point is well made by Philippa Foot, Natural Goodness (Oxford UK: Oxford
Understanding that the task is not to quell the sociopath with some
brilliant philosophical formula helps us see more clearly what the trou-
blesome characters really represent. We should envisage the skeptical
challenge as posed by ordinary people, whose socialization is reason-
ably effective and who feel the tug of ethical commands. They have a
tendency to conform but also want to know why they should be glad to
have this disposition. The troublemakers are devices for giving sub-
stance to this worry, personae with which the skeptical questioner can
vicariously identify but for whom following through on the identification
would be psychologically disastrous. The skeptical question is not a de-
mand to be talked into complying (typically, disobeying the rules is not
a live option), but a request for reassurance. The questioner needs to feel
at home with his or her ethical propensities.
We can now see why certain approaches to ethics, particularly the
nonnaturalistic Kantian and contractarian varieties, seem appropriate
replies to skeptical challenges. They provide reassurance by delineating
an ideal of rational thought and behavior, more or less thoroughly articu-
lated, so people who already feel the ethical tug can identify a mistake
deviants would be making. These philosophical replies cannot (to repeat)
silence deviants or bring sociopaths to heel. But they succeed at a more
modest task.
So too does pragmatic naturalism. To the extent people who wonder
whether they should be glad to have ethical dispositions can be satisfied
with explanations invoking practical rationality, they should be (at least)
equally content with the pragmatic naturalist account. For that account
places ethical practice at the center of our humanity, viewing ever-more
refined attention to altruism failures, ever-increased recognition of the
wants of others, as preconditions of the kinds of lives we live and the
kinds of societies we have. Although one may challenge parts of the ethi-
cal practices we have inherited, there is no escaping the ethical project.
The only social alternative we know is that of our hominid ancestors and
our chimpanzee contemporaries, an alternative from which the intro-
duction of ethics originally liberated us. That reply should be reassurance
enough.
In a justly famous image, Otto Neurath specified our epistemological
predicament, comparing us to sailors who must constantly rebuild the
vessel on which they sail. 31 Pragmatic naturalism takes a similarly anti-

foundationalist approach to ethics, denying any serious human alter-
native to the ethical project. Skepticism is, in the end, nothing more
than an invitation to jump into unknown, and potentially dangerous,
waters.
§43. Settling Disputes

One final version of the objection needs to be confronted. In discussing
the challenge posed by free spirit, I supposed there would be no threat,
provided free spirit was prepared to participate in the ethical project.
Under those conditions, I envisaged him questioning the progressive-
ness of certain transitions and assumed agreement on this issue would
be followed either by his conformity to the rules (if the episodes scruti-
nized were endorsed as progressive) or by vindication of his objection (if
the episodes were found to be regressive). By what right, however, can
one suppose that internal disputes, those arising among parties equally
committed to the ethical project, can be resolved? Champions of rival
approaches to ethics can argue that their favored accounts provide ways
of settling disputes—they offer fundamental principles or methods for
arriving at such principles—and so far pragmatic naturalism provides
no counterpart.
This last concern mixes elements in two previous complaints about
naturalism. In §40, pragmatic naturalism was absolved from any general
commitment to fallacious, allegedly justificatory, inferences. The last
section addressed skeptical challenges by distinguishing the rejection
of the ethical project in toto from attempts to reform ethical practice. To
decide the validity of these local challenges, however, requires methods
of resolving questions about progressiveness, and these must embody
something more specific than the general possibility recognized in the
response to Hume (§40).
31. Otto Neurath, “Protokolsätze,” Erkenntnis 3 (1932–33): 206; translated and re-
printed in Logical Positivism, ed. A. J. Ayer (New York: Free Press, 1959), 199–208, 201.
Quine made the passage famous by using it as the epigraph for his influential book, Word
and Object (Cambridge, MA: The MIT Press, 1960).
Functional conflict is a source of potential trouble. Different partici-

pants in the ethical project may proceed from different views about the
functions to be discharged, and this may lead them to quite radically
divergent judgments about what precepts are to be endorsed and what
actions allowed. Pragmatic naturalism allows for pluralism (§38), but
one might wonder if it has ways of resisting truly rampant pluralism.
Could the troublesome characters of the last section moderate their
stances to commit themselves to the ethical project but recommend pur-
suing it in deviant and threatening ways? The possibility generalizes the
threat perceived at the end of the last chapter, where I envisaged overrid-
ing, or abandoning, the original function of ethics.
If pragmatic naturalism is to serve as a framework for substantive ethi-
cal discussion, it will have to discharge the basic task of showing how
particular disputes can be resolved, how particular claims can be justi-
fied. More exactly, it will need to explain exactly what sorts of justifica-
tion are available, to actors or to analysts, when live issues about the
continuation of the ethical project arise. Hume’s original challenge was
met by identifying a criterion for the cogency of ethical inference. The
latest concern asks for the specification of inferences that meet this crite-
rion. The specification has not yet been given. Hence, as so far elabo-
rated, pragmatic naturalism is inferior to classical accounts of ethics that
appeal to overarching general principles in the resolution of ethical de-
bate. It does not follow that the deficiency is irremediable. To infer that no
method can be given would be a (counternaturalistic) fallacy. The next
two chapters take up this task.
To summarize: without some work in substantive ethics, a normative
stance, the metaethical picture of pragmatic naturalism remains incom-
plete. For the moment, however, without knowing the extent to which it
is possible to elaborate a cogent theory of ethical method, we can offer a
preliminary reply to the final worry. Proposed reforms of ethical prac-
tice are adjudicated by modes of argument tending to promote ethical
progress. If a challenger denies our conclusions, we make explicit the in-
ferences we have made. If the inferences are questioned, we show how
they accord with general rules. If the challenger disputes the rules, we
demonstrate their conformity to the fundamental criterion, reliability
in generating progressive transitions. If the understanding of ethical
progress is disputed, we review the history of our ethical practice, show-

ing how the original function has been served and how new functions
have emerged.32 If the inferences can be shown to accord with the rules,
the rules with the fundamental criterion of promoting progress, and the
account of progress to identify genuine functions, ethical disputes can be
settled. Assuming pragmatic naturalism can elaborate a substantive ethics
furnishing an appropriately constraining method along these lines—one
that will answer worries about the force and relevance of the functions
attributed—the only skeptical option remaining is to question the entire
project, and that, as we saw in the previous section, is to take the leap
favored by free spirit in his most ambitious and uninhibited guise, a leap
into a completely unknown, and doubtfully human, form of life.
32. There is plenty of room for reform here—for challengers might uncover points at
which regressive precepts or institutions have been introduced. Pragmatic naturalism is by
no means committed to the vindicating (“English”) genealogies, of which Nietzsche com-
plained. See On the Genealogy of Morality.
Pa rt T h r e e
A Normative Stance
chapter 8
Progress, Equality, and the Good
§44. Two Visions of Normative Ethics

In the classical conception of normative ethics, espoused by religious
traditions and philosophers alike, normative ethics aims to offer a set of
resources to help people live as they should. Prominent among these
resources are sets of prescriptions for guiding action, but there are often
also vivid stories depicting what is admirable and what is not. Religious
thinkers usually suppose all the fundamental resources have been pro-
vided, in an act of revelation, so the task remaining for the normative
ethicist is to articulate the principles clearly and precisely, showing how
they bear on the circumstances of contemporary life. Philosophers, by
contrast, think the principles of normative ethics must first be generated
and defended before the work of articulation and application can begin.
Rather than invoking divine revelation, they supply reasons for thinking
their favored system is correct.
Despite their differences, almost all approaches to normative ethics
share a static vision. Correct principles and precepts await discovery,
and once apprehended they can be graven in stone. Pragmatic natural-
ism sees things differently. The ethical project evolves indefinitely. Prog-
ress is made not by discovering something independent of us and our
285
societies, but by fulfilling the functions of ethics as they have so far emerged.
The project is something people work out with one another. There are
no experts here.
From this perspective, normative ethics requires continuing efforts to
decide how to live together in a common world. Each generation renews
the project, going on from the point reached by its predecessors. The tasks
facing normative ethics are those of deciding what should be retained,
what modified and how—also to resolve how those decisions are made.
If we are like sailors, repairing our ship at sea, we must determine which
planks to leave in place and which to move—and our efforts have to be
coordinated. The normative ethicist’s role is not to offer the grand plan
but to help the coordination.
Pragmatic naturalism assigns philosophers the task of facilitating dis-
cussion of how we should continue the project of living together. Philos-
ophy makes proposals. (That is itself a proposal.) Given the approach to
progress offered in Chapter 6, one type of proposal should identify the
problems, unsolved and partially solved, to which ethical practice has
responded: call this the diagnostic proposal. Another type of proposal,
the methodological proposal, should offer suggestions about how pro-
posals are to be adjudicated, about the rules of the continuing ethical
conversation. (That does not close off the possibility of reverting to the
classical vision of normative ethics, since one possible conversation is a
monologue: we could decide to listen to the advice of a sage.)
Philosophical proposals can be more or less informed, more or less
articulated and supported. In light of Parts I and II, we can identify the
functions of ethics and try to use factual knowledge to find ways in which
those functions could be more effectively discharged. If there were only
a single function, or if there were no danger that fulfilling the distinct
functions that have emerged would involve choices among them, the im-
portance of deliberation would recede, or even vanish. Ethics could
become a fully empirical discipline, whose task was to discover the ways
of satisfying a single function or a harmonious set of functions. Because
the pursuit of local progress gives rise to functional conflict (§§37–38),
the ethical project cannot be turned over to a group of specialists who
can work on finding ways to fulfill a function (or compatible functions).
A first proposal: we have little alternative to recapitulating the original
Progress, Equality, and the Good 287
predicament of the pioneering ethicists. There are no external con-

straints to which we can appeal, no modes of knowledge to spark ethical
innovation. Because the function of ethics is not to articulate the will
of a transcendent being, the religious authorities who claim to have
special access to the will in question no longer have the last word, or
even any word more convincing than that of their fellows. By the same
token, secular philosophy cannot lay down fundamental principles a
priori, or even through generalization from allegedly basic ethical in-
sights. What resources are left except to renew discussion with one
another, to reject the distortions introduced by advertisements of ethi-
cal expertise, and to continue the conversation our remote ancestors
began?1
This chapter and its successors are ventures in facilitation. The pres-
ent chapter will offer some diagnostic proposals; methodological pro-
posals are postponed to Chapter 9. Any package of proposals should be
internally coherent—the diagnostic suggestions and the methodological
proposals ought to be mutually supportive. As we shall see (§56), a re-
quirement of coherence plays an important role in addressing residual
skeptical concerns.
Those concerns arose (§39) in the form of asking why the “original func-
tion” of ethics should concern us. They were generalized (§43) by noting
that people committed to continuing the ethical project could give priority
to different functions. Pragmatic naturalism’s vision of normative ethics
identifies resolving controversies of this sort as the normative question.
Skeptics are not pale figures of philosophical imagination, but alive and
among us, offering different ideas about functional priority in going on
from where we are. If our efforts are to be coordinated, if the ship is to
stay afloat, these differences need resolution. Pragmatic naturalism needs
to propose a normative stance, and, if it can do so convincingly, the meta-
ethical challenges will be addressed en passant.
The strategy is to proceed in stages: first some diagnostic proposals,
then some methodological proposals, and finally an argument for taking
1. Here I allude to a famous remark of Richard Rorty (see the closing passages in Phi-
losophy and the Mirror of Nature [Princeton, NJ: Princeton University Press, 1979]). I sus-
pect I assign a larger role to philosophical midwifery than Rorty would allow.
the package to have a coherence its (skeptical) rival lacks. I begin by pre-
senting a useful framework for diagnosis.
§45. Dynamic Consequentialism

The ethical project can evolve indefinitely. Its progress is not measured
by the decreasing distance to some fi xed goal—the accumulation of cor-
rect ethical principles, central to the classical vision of normative ethics.
There can be progress from as well as progress to. Once again, the anal-
ogy with technology helps. Inaccessible destinations posed problems to
be solved by developing methods of transportation, and technological
progress in this domain consists in overcoming those problems (and the
derivative problems to which they give rise). No ideal technological sys-
tem beckons us on, and it would be folly to assess our progress in terms
of our proximity to any such thing.
Our predecessors made ethical progress by responding to parts of the
problem background, introducing precepts and other ethical resources.
They did so by framing (partial) solutions, and their recognition of their
innovations as potential solutions depended on seeing the state in which
the novelties were introduced as better than the status quo. They oper-
ated with an implicit notion of the good, fi xed and determined through
their agreements. The good is local, linked to circumstances and prob-
lems; it is constructed through group attempts to solve problems; and it
evolves.
Seeing socially embedded normative guidance as operating with a
tacit notion of the good permits the introduction of a useful normative
framework. The ethical project can be understood as a series of ventures
in dynamic consequentialism. Participants in it respond to their problems
by trying to produce a better world. They have an implicit conception of
the good and take the rightness of actions to depend on their promotion
of the good as they envisage it. Does viewing them in this way foreclose
possibilities? No. Consequentialism is a far more flexible view than its
critics usually assume. Not only are there many different ways in which
consequentialists can characterize the good, but a consequentialist ethi-
cal theory can explicitly acknowledge it has no complete specification of
the good, seeing its judgments as incomplete and provisional. Dynamic
consequentialism makes exactly that admission, supposing that concep-

tions of the good evolve, that some of the transitions among those
conceptions are progressive (in the sense of Chapter 6), and that later
conceptions of the good are (sometimes) superior to their predecessors,
even though none can claim to be the last word.
According to Mill, the central idea of consequentialism—that the
rightness of actions depends on their consequences—is the “doctrine of
rational persons of all schools.”2 To see the problem with ethical systems
that focus on obedience to prior rules and ignore consequences, con-
sider self-sacrifice. Someone who sacrifices himself typically does so for
“the sake of something which he prizes more than his individual happi-
ness,” and, if this is not the case, it is hard to see why the alleged hero
is “more deserving of admiration than the ascetic mounted on his
pillar.”3 There is no ethical justification for following a rule unless one
has grounds for viewing that rule as authoritative, and those grounds
can come not from labeling the source—either as a divine lawgiver or as
its detheologized counterpart “the moral law within”—but only from
recognizing the rule as well adapted to producing good outcomes. Fol-
lowing rules not well adapted to producing good outcomes is a capri-
cious and irresponsible thing to do: that is why consequentialism is “the
doctrine of rational persons of all schools.” Ungrounded deontology
is dangerous.4
There are alternative “schools” because there are very different ways
of measuring the value of consequences. When someone acts, he or she
changes the way the world would have been in the absence of the action,
or would have been if he or she had done something different. At its most
inclusive, consequentialism can compare the different total world histories,
the courses of the world that run from the beginning to the end, and can,
in principle, focus on any or all of the features of those histories, includ-
ing those features inspiring deontologists to lambast the simplest versions of
consequentialism. Typically, consequentialism focuses more narrowly—
2. Mill, Works, J.S. Mill, Collected Works (Toronto: University of Toronto Press, 1970),
10:111, from the essay “Bentham.”
3. Mill, Works, 10:217.
4. The most spectacular examples of ungrounded deontology occur in invocations of
the divine will to prohibit (or command) human actions (§§26, 35).
for utilitarians, the relevant points of comparison are the subjective

experiences of sentient beings, specifically their pleasures and pains. In
moving from the general idea of consequentialism to utilitarianism, vari-
ous important assumptions play a role:
1. The only aspects of the world needing to be considered in

evaluating the consequences of an action are those subsequent to
the action.
2. These pertinent aspects are properties of the lives of existing
beings, all of whom belong to a particular class.
3. The existing beings in question are sentient beings.
4. The relevant facts about their lives can be assessed by consider-
ing the stream of their subjective experiences.
5. The subjective experiences can be conceived as momentary (or
fleeting) states, to which value can be assigned.
6. The relevant states are experiences of pleasure and pain.
7. The value of a pleasure (or pain) is measured by its intensity.
8. The intensities are summed across the interval through which
they persist.
9. The value of an individual’s life course is calculated by summing
the values of pleasures and subtracting the values of pains across
all moments of the person’s life.
10. The value of a world history is calculated by aggregating the
values of the lives of all the individuals from the point of the
action on.
11. In practice, comparisons can typically be made by using small
segments of the lives of a few people as proxies for the values of
world histories.
12. Strictly speaking, an action is right if the value of the world
history it generates is at least as high as the values that would
have resulted from any of the available alternatives (but using the
proxy calculations identified in assumption 11 will almost always
provide an accurate test).
Each of these assumptions deserves scrutiny. You might think both

assumptions 1 and 2 extremely plausible. After all, an agent can do noth-
ing to change the past, so assessment can concentrate on what happens

after the agent’s action. That does not follow, however. For evaluating
the consequences, it should not be assumed that the value of an event or
state of affairs is always independent of its causal history. Consequential-
ists can hold that an effect has greater or lesser value in virtue of its rela-
tion to past actions: even if your spending time with the sick would pro-
mote greater happiness than expressing your gratitude to someone
who has helped you, your overt gratitude and the past aid may confer
sufficient value to outweigh the benefits provided by hospital visiting.
Imagine two world histories involving exactly the same distribution
of pleasures and pains, satisfactions of desires, and anything else about
individual people you might take to be relevant to their individual good,
but diverging in the causal relations. Suppose, for simplicity, all that
matters is pleasure and the absence of pain. In one world history, many
of the pleasures and pains come about because of systematic relations
among people—there are relationships of mutual helping; punishment is
given for harms caused—much as things happen in the world we know.
In the alternative, the causes of the pleasures and pains are quite random:
you do something to please a friend, and, instead of the friend, a complete
stranger acts to give you pleasure; you cause harm to someone, and, out
of the blue, a harm equivalent to the punishment is inflicted on you. If
you think something valuable has been lost in this abnormal world, with
its unsystematic ways of generating pains and pleasures, you think the
causal relations affect the goodness or badness of a world history.5
Assumption 2 is similarly vulnerable. Consequentialism is not com-
mitted to the view that you can consider lives in isolation from one an-
other: it is quite possible that relations among lives, and relations of liv-
ing beings to other constituents of reality, are sources of value (consider
the value that accrues from higher-order altruism; §21). It is even possible
that nonliving things, and the relations among them, could be sources of
value (§47).
5. Perhaps one might claim that the difference in value is explicable because of the atti-
tudes people would have: their pleasures would be less in the causally unsystematic world.
Then grant them the illusion of thinking the causal relations are as they normally are. The
difference in value persists.
There is a very general reason for worrying about almost all the as-
sumptions generating utilitarianism from consequentialism. Utilitarian-
ism derives from consequentialism by a series of reductionist moves. We
aim to compute the values of the worlds that would flow from our envis-
aged actions. We reduce the problem to one of summing the values of the
lives of a class of individuals; we reduce it further by considering only
sentient individuals and further still by ignoring most of these and con-
centrating on those we suppose immediately affected by the actions un-
der scrutiny; we reduce the problem of measuring the values assigned to
the individual lives by decomposing those lives into a sequence of mo-
mentary states; now we assign values to those states by reducing the
aspects we consider to the intensities of pains and pleasures; having
reduced the problem in this way, we can start summing, and arrive at the
measure utilitarians commend. Any or all of the reductions could be
questioned. For there is no reason to think the value of a world will al-
ways consist in the sum of the value of the lives of the individuals we
consider one by one—distribution might be crucial. Further, we have no
reason to suppose that the value of an individual life can be generated by
summing the values of momentary states (or even longer experiences
that occur in people) taken in isolation from one another.
Consider an alternative view of the value of individual lives. It starts
by taking a valuable human life to be one directed by the free choices
of the person whose life it is, as something given coherence by an in-
dividual conception of the good.6 However pleasurable a sequence of dis-
jointed experiences (even a repetitive sequence) might be, it would fall
short unless it had certain global properties. Valuable lives exhibit a
plan.
Similar holistic considerations apply to social states. Rather than
thinking the valuable social state is one in which the total sum of plea-
sure over pain in the population is high, we might conceive a group of
people pursuing a wide diversity of projects for their lives, who are none-
theless bound together in relations of dialogue, joint action, and mutual
6. As Mill claims throughout On Liberty and other Essays [Oxford: Oxford University
Press (The World’s Classics), 1998]. This essay shows his enormous debt to the Greek
thinkers he studied from childhood on.
sympathy.7 Both the heterogeneity of the whole and the interconnec-

tions among the individuals are sources of value. Approximate equality
of material resources might obtain value in terms of its contribution to
the possibilities of all members of society having fair chances to fashion
and to pursue their life plans.
Consider poetry. Discovering a great poet—Wordsworth, say—is
not a pleasurable experience like that of a good meal, or even a soother
of pain, like an analgesic. If a young man, from whom poetry is with-
held, lapses into a state of dejection, from which reading Wordsworth
helps him to recover, it is odd to suppose he has the medicine his con-
dition required. An encounter with Wordsworth is not significant for
its transient effects, but something that can reshape and redirect an
entire life. A calculus of assigning values to passing experiences, based
on intensity and duration, is inadequate to measuring the value of Word-
sworth, because the Wordsworth encounter resonates in subsequent
life.8
In the house of consequentialism, there are many mansions. My exca-
vation of twelve different places at which you could make different deci-
sions about the good, and derivatively about the right, reveals abstract
possibilities. The central point, however, is that the flexibility of conse-
quentialism allows us to understand ethical progress as the fashioning of
conceptions of the good that better discharge the functions of socially
embedded normative guidance and the framing of rules as consequen-
tialist projects directed at promoting the good (according to the preva-
lent conception). Conceptions of the good are duals of the functions as-
sembled in the growth of the ethical project.
7. These holistic considerations are evident in Mill’s own writings, particularly in On

Liberty, Considerations on Representative Government in On Liberty and other Essays [Ox-
ford: Oxford University Press (The World’s Classics)], 1998, and Principles of Political
Economy, vols. 2 and 3 of J.S. Mill, Collected Works (Toronto: University of Toronto Press,
1970).
8. My discussion plainly draws upon Mill’s own experiences and his sense of the impor-
tance of reading Wordsworth and of meeting Harriet Taylor. His writings make it evident
that he understood the holistic significance of these events: see, in particu lar, the essays
“Bentham” and “Coleridge” Mill Works vol. 10, the dedication of On Liberty, the Autobiog-
raphy (New York: Columbia University Press, 1966), and the closing pages of On the Sub-
jection of Women (in On Liberty and other Essays).
A natural response: consequentialism can accommodate so much be-

cause the alleged “flexibility” makes its claims vacuous.9 That is already
indicated by my concession that many of the features deontologists view
as lacking in simple forms of consequentialism can be introduced within
a consequentialist framework: one can, for example, insist that the good-
ness of a state of affairs depends on the character of the psychological
processes (including the intentions) that brought it about (this is a spe-
cial instance of the general point that causal relations matter). The dan-
ger of ungrounded deontology shows, however, that the consequentialist
framework rules out some things, namely prescriptions of what is right
that do not advance ethical functioning (and, in the religious case, often
interfere with it). The insights of traditional deontological thinking can
be viewed as contributions to a progressive understanding of the good,
whereas the deficiencies of other deontological possibilities (those resist-
ing such assimilation) can be exposed.10
§46. Failures and Successes

At the early stages of the ethical project, our ancestors focused on the ba-
sic desires of their fellows. They recognized some wishes as endorsable
and uncontaminated, taking it to be good to divide resources according
to some type of equality under conditions of scarcity and to aim to in-
crease the supply so that all might have enough. Part I saw the emphasis
on supply as fostering social arrangements found in later societies: divi-
sion of labor and the development of roles and institutions (§§19–20).
For present purposes of stocktaking and diagnosis, considering how
features of our current situation might have emerged is not pertinent:
what is crucial is that they have emerged. Contemporary societies have
either abandoned the original emphasis on equality or spectacularly failed
9. Bernard Williams makes some perceptive remarks that anticipate this line of re-
sponse in his contribution to J. J. C. Smart and Bernard Williams, Utilitarianism: For and
Against (Cambridge, UK: Cambridge University Press, 1973), sec. 2.
10. So one might say that the idea of the unseen enforcer is unproblematic for ethical
progress when the deliverances of the commander are grounded in the prior understand-
ing of the good. Things go astray only when the practice of commanding takes on a life of
its own.
to achieve it. Conceive societies as demarcated geographically: start with

a set of people who live in the same place, and include all who are in so-
cial interaction with any member. Even if you begin with the well-to-do,
the society identified not only will be large but will also contain indi-
viduals whose endorsable desires are only incompletely satisfied.11 (This
is evident for poor societies, and for affluent societies with pronounced
inequality; even the most egalitarian social democracies contain pockets
of poverty and want.) If we think of the most elementary form of altruism
failure as one in which individuals do not respond to the basic (endors-
able, uncontaminated) desires of others, when it is possible to satisfy those
desires completely while simultaneously doing the same for everyone in
the society, it is evident that in the most elementary case of the original
function of ethics, contemporary societies do not realize the available ways
of discharging that function. On this score, many experiments in the
history of ethical practice have done better.
Why is this? Most plausibly, the structure of roles and institutions
generated in the evolution of the ethical project has displaced the origi-
nal conception of the good. New ethical functions have emerged, and,
while the roles and institutions we have may be well adapted to discharg-
ing these functions, they forfeit the ability to fulfill the original function.
Our contemporary codes permit maintenance of roles and institutions,
even though those interfere with remedying a very basic type of altruism
failure.
The contemporary conception of the good is surely progressive in
some respects, for it has absorbed ideas of higher forms of altruism, con-
tributions to joint projects, and richer possibilities of human life (§§36–
37). If the original conception of the good were reapplied in the contem-
porary context, it would have to attend to a broader class of endorsable
desires. In terms of offering equal satisfaction of these “higher” desires,
the failures of contemporary societies are even more glaring. Inability to
satisfy basic desires contributes directly to dramatically lowered chances
11. Here I am imagining societies as defi ned recursively. One might try to block the
conclusions by demanding that any two members must engage in social interaction, or by
raising the standards for social interaction. I shall discuss the issue of how to demarcate
societies more extensively below. For the moment, the recursive approach will help to fi x
ideas.
of achieving the “higher” goods. When the most valued social positions
go to those who have performed outstandingly on the tasks assigned by
an educational system, people often fail because of their inability to sat-
isfy basic needs (hungry, ill-sheltered children do not always concen-
trate, children who go to decrepit, dangerous schools do less well on the
exams devised to test them).
Functional conflict underlies social tensions but is not itself perceived.
In practice, the conflict is settled by overriding the original function of
the ethical project, and that suppression is thoroughly institutionalized.
Equipped with an account of the ethical project and its evolution, we can
be more reflective. Dynamic consequentialism sees our normative task
as beginning from respecifying the good. That specification cannot be
achieved without reflecting on roles and institutions—ethics is social.12
The development of a richer repertoire of emotions for responding to
the actions of others, especially the negative sentiments of indignation
and revulsion provoked by injustice and cruelty, underlies current ways
of securing compliance to our codes. People living in a state of blatant
inequality, people whose basic needs are unmet, are often not persuaded
that there has been no real injustice or cruelty. Indignation and revul-
sion easily reinforce the anger and frustration that would have been felt
by their ancestors under similar circumstances. If the amplified emo-
tional reaction is held in check, it is only because the injunction against
initiating violence is powerfully enforced: asymmetries in wealth and
power are used to threaten those inclined to protest with even direr con-
sequences for themselves and their loved ones. (This kind of coercion
figured in the scenarios about dictatorial rule—§34; impossible though
it may have been among the small bands of the Paleolithic, it is readily
contrived within contemporary societies.)
Were we to consider the entire human population as a single society,
the extent and range of inequality would be enormous. Each day, thou-
sands of people die or become disabled through lack of the resources
12. The phrase is Dewey’s, chosen as title for the last chapter of Human Nature and
Conduct (Amherst NY: Prometheus Books, 2002). The point encapsulated in the phrase
recapitulates themes in Foucault and Marx, themes often beyond the horizons of contem-
porary ethical discussions.
needed to satisfy a basic desire: to preserve the body in a healthy state.13

The distribution of opportunities for satisfying less fundamental de-
sires, say, for work of a rewarding kind or for cultural development, is
even more skewed. Concentrating on smaller human units, the less pop-
ulous nations or communities within them, makes the inequalities less
extreme; but affluent democracies and poor countries alike contain siz-
able groups of people who can see how the basic and nonbasic desires
of others are often well, even fully, satisfied, but whose elementary wants
are unmet. Only a few societies, committed to public support and to
egalitarian arrangements, reduce the scale of the inequality. Even in them,
pockets of deprivation remain.
If the difficulties of profound inequality are not as evident to us as the
smaller problems our hominid forebears encountered, it is because those
from whom we might expect hostility are often distant or invisible, and
because the concentrations of resources have generated forms of power
enabling people who enjoy large benefits to threaten the less fortunate
with severe penalties for subversive behavior. From the perspective of
pragmatic naturalism, that cannot be a solution to our magnified analog
of the original problem, for the fragility and tensions of our societies are
addressed only in a way that does not identify their root cause, to wit the
wide-ranging altruism failures. As with the original powerful dictator of
§34, the reduction of social conflict does not work through remedying
altruism failure and therefore fails—spectacularly fails—to discharge the
original function of ethics.
Not all the conflicts and tensions of contemporary societies (whether
the human population is considered as a whole or whether it is split up
into smaller groups) are based on the inequalities toward which I have
gestured. Some versions of the unseen enforcer interpret the will of the
invented being in evangelical terms, so the rules they claim to be expres-
sions of that will are supposed to apply, not only to the society originat-
ing the mythology, but also to the entire human species. Societies gripped
by this conception will certainly be at odds with other groups whom
13. The annual reports of the World Health Orga nization provide (sobering) data. See
James Flory and Philip Kitcher, “Global Health and the Scientific Research Agenda,” Phi-
losophy and Public Affairs 32 (2004): 36– 65, for a small selection from the statistics.
they take to be creating a world inhospitable to their favored ideals of

living and who may even be disposed to spread those ideals by any effec-
tive means (including force). Some social conflicts and tensions arising
within contemporary societies—most evident when our species is viewed
as a single society—result from the dangerous ungrounded deontology
that articulates the initially useful idea of the unseen enforcer. Codes
subordinate concern with the desires and aspirations of individual people
to the prejudices of those who advertise themselves as knowing the en-
forcer’s will.
Part of the emancipation of ethics from the concept of the enforcer
consists in the recognition of previously forbidden desires as endors-
able, the withering of vice (§26). Renewed discussion of the good must
be alert to residues of pressures that have made some desires invisible.
Yet comprehensive diagnosis should recognize what imaginary tran-
scendent policemen have contributed to the ethical project. Most obvi-
ously, they have helped secure greater compliance (§17). Beyond this,
they have sometimes played important roles in relieving or in exposing
the central feature of our contemporary—and recent—ethical predica-
ment, to wit the presence of vast inequalities, resulting from abundant
altruism failures and producing widespread social tension. Religions
have sometimes laid great emphasis on recognizing the needs of the poor,
the oppressed, and the unlucky: they have, in effect, continued to rec-
ognize the desires of those whose aspirations are frustrated, even when
these people are excluded from ethical deliberation. Moreover, religious
community has sometimes fostered the expression of protest against
inequality.
A concomitant of the growth of societies, itself a progressive transi-
tion, made possible through including outsiders, is that the original ethi-
cal venture, in which an entire social group tried to find common
ground, has become practically impossible. Even without deference to
people supposedly having special access to the divine will, it would be
necessary to assign the task of reviewing and possibly revising a code
to a (small) minority. That forfeits an important feature of the extended
conception of the good life (§20), the ability to perceive oneself as play-
ing a role in joint projects, conceived by each participant as directed to-
ward a common good. The shared ethical life begins to disappear.
Diagnosis has so far focused on problems and conflicts, places where

we might consider moving some planks on the ship we sail together. Yet
we should not overlook the successes of the ethical project in achieving
directives for individual behavior. There are core ethical truths (with
associated rules), vague statements attained by all extant ethical traditions
that we reasonably expect will be preserved in an indefinite sequence of
progressive future transitions (§38). Whether or not we can make them
more exact, or find new concepts for their sharper expression, conduct
in accordance with, and out of respect for, these themes is constitutive of
any plausible conception of the good.
Consider honesty. Because there is a very large class of actual or po-
tential altruism failures, in which one individual deceives or manipu-
lates another, regulation of speech acts will figure in progressive ethical
codes at an early stage. Exchange of information is crucial to human life
and has been crucial since we learned to talk. If the informant is to re-
spond to the wishes of the inquirer, it will be important to tell the truth
(or to attempt to convey the truth, to be sincere), and discovery that
truth has been withheld will generate trouble. “Always tell the truth”
looks like a rule to help fulfill the original function of ethics; truth telling
is an activity classified as good by a progressive conception.
For a very large segment of human history, the simple formulation “Al-
ways tell the truth,” was probably apt—just as it is an apt rule for children
today and a serviceable one for many people throughout the course of their
lives. Nevertheless, truthful responses to questions are not always and
everywhere good. A stereotypical scenario involves the demands of the
Gestapo officer at the door, when the householder has hidden Jews in the
attic. The officer asks, “Are there any Jews in the house?” For the story to
work, we must suppose he trusts the householder; given a negative re-
sponse, he will leave and the hidden refugees will be safe. Here, the conse-
quences of honesty are so dire that many people—perhaps everyone not
in the grip of a philosophical theory—thinks lying is good.
Few people encounter situations of this type, but many more face mo-
ments when telling the truth has adverse consequences. One of those you
love most is seriously ill, and the medical report is grave. You know her
determination and resolve will be sapped if she learns the details. Or
perhaps there is really no hope, and you know her fi nal days will be
miserable if she feels there is no chance of recovery (you know her well;
she is sensitive and liable to be crushed by dreadful news). She asks
you point-blank what the doctors have said. You are well prepared and
know you can convince her. Lying would avoid the bad consequences of
telling the truth.14
These examples have to be carefully constructed if they are to call
into question the universal goodness of truth telling. The potential de-
ceiver must be clearheaded, must have thought through the options, and
must have understood that there is no honest alternative to the decep-
tion. The officer at the door cannot be fobbed off with an evasive answer,
nor will your loved one be satisfied with even the most adroit attempt to
change the subject. This means the lie will probably have to be premedi-
tated and its expression carefully prepared—in general, conditions we
take to intensify the wrongness of a misdeed.
14. Thinkers inspired by Kant may claim that allowing any exceptions undermines the
practice of answering questions, or perhaps that it violates the integrity of the speaker, or
perhaps that the good requires that people always act in nonmanipulative ways. When the
odds are set very high, as they are in the case of the officer at the door, none of these replies
succeeds. The lie to the Gestapo officer is not going to undermine any important human
practice, even if all householders in similar situations behave in the same way. If the lies are
discovered, the investigating officials are likely to be more skeptical of what they are told
and hence may investigate more thoroughly—and that will have to be countered in future
practices of asylum granting—but these refugees will have been saved. The example of the
seriously sick illustrates the point very well. Those who are extremely ill know that people
who care about them will be inclined to spare them bad news. Knowledge of that sort does
not undermine any practice of questioning and answering but simply means they are more
suspicious than they would otherwise have been about what their loved ones tell them.
Many deontological traditions, religious as well as philosophical, have taken the bad-
ness of deception to consist in the corruption of the liar—in effect, they offer an account of
the good that gives priority to purity of heart and suppose telling a lie, even to promote
otherwise good ends, compromises this purity. Religious teachers and leaders have been
advised to respond to questions by discovering ambiguous or evasive or misleading (but
not outright false) answers or to take oaths with a private “reservatio.” Faced with the offi-
cer at the door, householders who spend time trying to discover a tricky way out are not so
much worthily protecting themselves from corruption as irresponsibly putting others at
risk. Even if the chances the officer will penetrate the cleverly evasive reply are slim, the
damage done by arousing his suspicions is too great to justify the maneuvers. Further, why
should we suppose that those who concentrate their effort on shielding vulnerable victims
they have taken in are somehow less pure than people whose first thoughts are divided be-
tween the threats to the refugees and their own rectitude?
In both examples, the desires of the questioner are ignored or overrid-

den. The desire to fi nd the Jews is ignored because it is grounded in
a spectacular altruism failure; even a rudimentary account of contami-
nated desires (§34) differentiates the officer’s wish to root out and kill
from the desires of the refugees for escape. The patient wants the truth,
but she also wants the truth to be benign—and because the deceiver under-
stands the strength of that latter wish, as well as the consequences of
knowledge that it cannot be fulfilled for other desires the patient has (the
desire not to live bereft of hope), she answers untruthfully. Here decep-
tion can be seen as altruistic, seeking the most important desires of the
sick one and responding to those. (Plainly, this is an instance of pater-
nalistic altruism—§21.)
The simple prescription of truth telling originates in contexts where
any foreseeable deception involves only altruism failure. Lies advance
the selfish ends of the liar. The desires of the recipient of the false in-
formation are ignored. In simpler societies there will typically be other
ways of heading off the threat posed by a questioner who wants to inflict
harm on a third party—one may summon other group members so com-
plaints can be heard. Similarly, only when special kinds of knowledge
are in play does the problem of protecting others from disturbing truths
arise.15 With the evolution of culture, however, altruistic lying becomes
possible. Altruistic lying can resemble other forms of altruism we en-
courage (we admire those who risk their own safety to protect the perse-
cuted). Insisting on the simple rule overlooks the fundamental role rein-
forcing altruism plays in progressive ethical change.
Yet we do not know precisely how to amend the conception of the
good so it is no longer committed to the universal badness of lies. To ap-
peal to obvious aspects of the conditions under which reinforcing altru-
ism is good (support the refugees, not the wishes of would-be murderers),
and to declare that lying is permissible when it satisfies these conditions,
is not adequate. The Kantian thought that universal permission to lie (in
15. It is interesting to speculate about the length of time for which altruistic lying has
been possible—perhaps there are occasional analogs of the examples I have used that could
have arisen for the small groups of the Paleolithic—but it seems evident that concentrations
of power and advances in certain kinds of knowledge (medical knowledge, for instance)
generate clear cases with increased frequency.
some specified range of circumstances) would undermine the practice of

exchanging information will not support a blanket prohibition of lying,
but weakening the rule has consequences. A consequentialist analog of
the deontological point: one must attend to the consequences of publicly
permitting exceptions.16 Sometimes we have good grounds for thinking
the social consequences slight compared with the benefits of permitting
dishonesty—as, for example, with carefully circumscribed versions of the
two scenarios considered. That is compatible with ignorance about the
costs and benefits in other instances.
This section has offered a preliminary survey on which a respecifica-
tion of the good might be founded. The next task is to take up the (so far
postponed) question of group size. How large should be the scope of our
ethical concerns?
§47. From the Local Community to the Human Population

In the beginning, it was simple. The ethical project was undertaken by
small groups of deliberators, whose agreed-on rules and whose concep-
tion of the good focused on themselves. If we consider renewing the
project, how should we divide the human population (the totality of
members of our species alive at a particular time) into groups within
which versions of the ethical project, “experiments of living,” are tried
out? A simple suggestion appeals to geography: communities are demar-
cated by recognizing territories within which people live and out of which
they rarely, if ever, venture. Or you can emphasize causal interaction: two
16. Mill’s treatment of the permissibility of lying appreciates these points. Convinced
that “all moralists” would allow for breaches of the prohibition against lying, he points to
just the cases I have used: “when the withholding of some fact (as of information from a
malefactor, or of bad news from a person dangerously ill) would preserve someone (espe-
cially a person other than oneself) from great and unmerited evil, and when the withhold-
ing can only be effected by denial” (Works, 10:223). He cautions, however, that not every
good consequence is sufficient to allow for lying, since we also have to take into account the
costs that permitted deception will entail: “inasmuch as any, even unintentional, deviation
from truth, does that much towards weakening the trustworthiness of human assertion,
which is not only the principal support of all present social well-being, but the insufficiency
of which does more than any one thing that can be named to keep back civilization” (Works,
10:223).
individuals belong to the same community if they interact with each

other at greater than some threshold frequency, and the community is
the recursive closure of such interactions.17 Alternatively, appealing to
the altruistic tendencies that make life together possible, communities
are groups of people with dispositions to respond altruistically to one
another, people who lack such dispositions toward outsiders (or who have
them only in reduced form). Or, focusing on the limits of altruism, com-
munities are groups of people for whom altruism failures give rise to social
difficulties.
For almost all of human history it would not matter much which of
these criteria you selected, for all would agree. Most people who have
ever lived have spent most of their lives in the same place, interacting
with the same people, responding altruistically to those around them,
and living by a shared set of rules that attempt to remedy their potential
altruism failures. Nomadic tribes tend to stay together, even in their
wanderings, so that, at any given time, the geograph ical criterion will
align with the others, even though, at different times, the places occupied
are different. Individuals sometimes leave their native group to join
another—up to a certain point in their lives, they belong, according to all
the criteria, in one community; afterward they belong, again by all the
criteria, to a distinct community.
In the contemporary world the criteria no longer coincide. People
who live in one place have some of their most important interactions
with others who live at a great distance from them. If one appeals to a
criterion of causal interaction, connections will be made across geo-
graphical boundaries, and any insistence on transitivity will bind the
human population into one vast community: if I interact frequently with
someone in Australia, and you interact frequently with someone in cen-
tral Asia, and you and I also interact frequently with each other, then the
four of us are linked by a chain of interaction; if it is true that sharing a
17. That is, if A and B meet the interaction criterion, and if B and C do so as well, A and
C belong to the same community, irrespective of whether they also interact above the
threshold; the community consists of all the people who can be related to any of its mem-
bers through a chain whose adjacent members interact at a frequency above the threshold.
The minimal notion of interaction is one in which the behavior of one affects the choices
available and the prospects of success of the other.
community is transitive, we shall all be assigned to a single community,

even though I have no contact with central Asia, and you have none with
Australia, and our partners in Australia and central Asia have none
with each other.
There is a different way to think about the situation.18 One can aban-
don the idea that sharing a community is a transitive relation. Each of us
belongs to many overlapping communities. In terms of the illustrative
example, there is a community to which my Australian friend and I be-
long, another community to which you and your central Asian associate
belong, and a third that embraces you and me, but none that contains all
four of us (or even any three of us). It is possible to introduce community
structure at many different levels, using different criteria singly or in com-
bination. A particularly important way to do so is to appeal to history.
Those who live in a place in which social life has been regulated by a
particular tradition, and who identify themselves as part of that tradition,
might plausibly be counted as an ethical community, for whom the con-
tinuation of the ethical project consists in amending that tradition.
So there are many possibilities for specifying group(s) of people to
work together on respecifying the good. On what basis should we decide
among them? The conceptions introduced should be apt for the prob-
lems faced. Wherever there is a failure to respond to the desires of an-
other person, with respect to whom there is the potential for interaction,
we have a contemporary analog of the problem that underlies the original
function of ethics. Moreover, as at the beginning of the ethical project,
these altruism failures, in a world of far more extensive causal relations
among human lives, provoke social tensions and conflicts. These emerge
today as counterparts of the fragility of the hominid societies that were
transformed by ethics. A proposal: continuation of the ethical project
should include an attempt to frame a conception of the common good re-
sponsive to the desires of the entire human population.
Almost all human beings want a future in which the younger mem-
bers of their societies, and the future off spring of those younger mem-
18. In framing the ideas that follow I have been helped by some acute observations and
questions from Erika Milam. There are also links to the framework proposed by Dewey in
The Public and Its Problems (Athens OH: Swallow Press, 1985).
bers, can grow healthily. All extant ethical traditions count this as an
endorsable desire. Today, the actions of people in some areas of the
world interfere with the realization of such desires. Practices of com-
merce, agriculture, industry, and even medical research in affluent
countries decrease the probability that people in poorer regions will
be able to nurture healthy children.19 All members of our species face
the common problem of avoiding (further) environmental changes
that would dramatically disrupt human lives: global warming will,
almost certainly, make many heavily settled areas uninhabitable and
will leave many others vulnerable to extremities of weather that will
challenge available technologies of shelter. We urgently need a concep-
tion of the good that considers the desires of all people and that will
guide attempts to treat the problems engendered by a thoughtless indus-
trial past.
Causal interaction binds the entire human population together. The
altruism failures in that large community are dangerously magnified ver-
sions of those that prompted the first ventures in ethics. Pragmatic natu-
ralism’s proposal about the good gives priority to the continually more
extensive network of causal relations linking us all. It does not follow
that there is no place for more fine-grained partitions. There are issues
on which a local community could progressively elaborate a conception
of the common good by concentrating only on the desires of its members
and on remedying the altruism failures arising within it. The areas in
which this should occur, however, must be distinct from those covered
by more wide-ranging conceptions (ultimately by that conception that
takes the entire human population as its province): where there are seri-
ous consequences for distant others, there must be an attempt to respond
to them.20 Moreover, the local community’s explorations in this regard
are properly constrained by the vague central themes adduced as the
19. Commercial practices interfere with the supply of basic goods to the poorest areas of
the world; agriculture subordinates the task of feeding the hungry to considerations of
profit; industry squanders resources and neglects long-term energy needs; medical re-
search has, until quite recently, conspicuously neglected the problems of the poor.
20. Plainly, this proposal has links to Mill’s famous “harm principle” (On Liberty); even
more directly, it connects to Dewey’s reformulation in the opening chapter of The Public
and Its Problems.
best candidates for ethical truth. Nevertheless, the pluralism already ac-
knowledged (§38) is recapitulated in the idea of domains in which pro-
gressive elaboration of the conception of the good can proceed through
responses to the desires present in a local community (one that is a
proper subset of the human population).
An important objection: focusing on the global human population
draws the boundary too narrowly. Other things—nonhuman animals,
or parts of the biological or inorganic environment, or human artifacts—
have ethical standing. I shall take the case of nonhuman animals as central
to the objection.
Ethical traditions disagree about the extent to which our conduct to-
ward nonhuman animals should be regulated. Almost everyone agrees,
however, that people should not arbitrarily inflict pain on mammals and
birds (igniting cats is usually viewed as cruel) and that pet owners should
attend to the needs of the animals they have acquired. Some hold that
farmers should not breed mammals and poultry under conditions of close
confinement that distort the animals’ normal patterns of development
and normal metabolic functioning; these critics often hold that consum-
ers should not buy and eat animals bred in these cruel ways. Yet others
suppose the entire practice of eating nonhuman animals—whether birds
and mammals, or all vertebrates, or all invertebrates into the bargain—
is wrong. A similar range of positions applies to using nonhuman animals
in medical experimentation.
The aim here is not to consider which position, if any, would be ad-
opted in a progressive transition from the current state of ethical prac-
tice, but to address the concern that no view of any of these types can
be reconciled with the framework of pragmatic naturalism. It suffices to
focus on the most minimal rules governing conduct toward nonhuman
animals, like the proscription of torturing mammals. What ethical func-
tion would be discharged by introducing any such prohibition into ethi-
cal practice?
It is worth recalling part of the history of ethical experimentation. For
a long period in the ethical project, human groups were very small; out-
siders were not covered by any of the protections afforded by the rules of
the band. The scope of the rules was extended when peaceful exchanges
with neighbors effectively created something akin to a broader society.
“The circle expanded.”21 What occurs here is first that a set of interac-
tions takes place with people who have previously usually been avoided,
and second that failures to respond to some of their desires count as al-
truism failures. My “how possibly explanation” (§19) showed a way of
making that extension. Interactions are set up in an attempt to discharge
an ethical function (satisfying the endorsed desires of members of the
local group through trade), and the expansion of the rules fulfills an-
other function (the original function of remedying altruism failures;
some of the desires of the erstwhile outsiders correspond to endorsed
desires of group members, and not responding to these comes to count
as an altruism failure).
Is anything similar available in the case of nonhuman animals? Ap-
parently so. Just as the late Paleolithic witnessed first occasions of tran-
sient association among neighboring groups and later increases in band
size, so, at the very end of this period, people began to set up more regu-
lar patterns of association with some kinds of nonhuman animals. The
practice of domestication creates something like a society, one including
some nonhuman members. That practice refines the ethical function of
satisfying the endorsed desires of all. Yet, because the animals newly in-
cluded differ in some important properties from the people across the
river, obvious questions arise. Does this expansion really create any-
thing like a society? If so, to what extent can the rules adopted within
the local group be carried over to the new members?
In a minimal sense, a broader society is created simply in virtue of
recurrent patterns of interaction. Human beings feed their domestic ani-
mals, breed them, work them, and consume some of their biological
products. These interactions need to proceed in ways permitting a long
series of repetitions. Yet, because of asymmetries in power, that might
simply be done by brute force. If the “society” is simply constituted by
regular patterns of interaction, its stability requires no more than human
skill in confinement and handling.
21. As §33 observed, this is a prominent mode of ethical progress. Peter Singer’s title
“The Expanding Circle” expresses his conviction that this is the dominant mode of ethical
advance, and he argues that this expansion should bring nonhuman animals within the
scope of our ethical precepts.
Members of the group across the river may ultimately participate in

the ethical project as we practice it. Some generations hence, their de-
scendants and ours may sit at a common campfire and deliberate about
how to go on from where they are. That can never occur with our non-
human animals.22 Yet there are also human beings who do not partici-
pate in ethical deliberation with us and still belong to the scope of our
precepts—and this holds even if “we” are members of one of the early
bands that initiated the ethical project. Those sufficiently young lack a
voice, as do others, perhaps, whose capacities have not grown in normal
human ways; additionally, there are descendants, children and grand-
children yet unborn, some of whose desires we endorse, with the conse-
quence that we work to ensure that necessary conditions for the satisfac-
tion of those desires will obtain. The extension of the agreed-upon rules
to them does not result from the fact that they have been party to the
agreement—for they have not—but rather from a recognition of features
of their lives, including the occurrence in them of desires we endorse for
others. Despite the differences caused by current incapacity, permanent
incapacity, or position in time, there is a common basis on which the no-
tion of an altruism failure can be elaborated.
So too with nonhuman domestic animals. We can recognize their de-
sires for things human beings need (food, rest) and appreciate their wish
to avoid par ticu lar situations (close confi nement, pressure from sharp
objects). Once human emotions have been shaped to resent behavior
causing suffering to other people, there is sufficient community with
nonhuman animals to expect that, among the band of early pastoralists,
deliberations in the cool hour will reprove actions making the livestock
whimper or squeal.
How far will the progressive elaboration of precepts governing ac-
tions toward nonhuman animals go? There are two potential axes of
extension: one leading from domestic animals to their wild counter-
parts, and one encompassing a broader range of taxonomic categories.
So long as wild animals are seen as threatening, they will figure in the
same way threatening human beings from other tribes did, before the
22. Hume saw this point clearly and drew a harsh conclusion. See Enquiry Concerning
the Principles of Morals (Indianapolis, IN: Hackett, 1983), 25–26.
institution of rules for interacting with outsiders brought them within

the circle. Just as participants in the ethical project have learned, over
tens of thousands of years, to deal fi rst with nongroup members in a
limited range of contexts, to form larger societies governed by a common
framework of rules, and ultimately to apply an ethical code to people
with whom interactions are negligible or even nonexistent, so too we can
understand a progressive sequence of changes identifying altruism fail-
ures in inflicting, or even tolerating, pain in animals whose anatomy,
physiology, and behavioral reactions resemble those already embraced
within the ethical framework. Ethical rules could come to cover animals
relatively like those we routinely domesticate—foxes as well as dogs.23
More difficult is the task of deciding just where the taxonomic boundar-
ies are to be drawn. Which kinds of animals have sensations sufficiently
like those of our nonhuman exemplars to warrant counting their pains
as states we should try to avoid or relieve? Reptiles? Fish? Nautiloids?
Molluscs? Worms?
The idea that it is good to relieve pain, wherever it occurs, is a natu-
ral extension of ideas about the good, progressively elaborated during
the evolution of the ethical project. Accompanying that idea is a related
thought about human beings: it is not good for people to be insensitive
to pain, whether it occurs in other people or in nonhuman animals.
Infl icting pain— or even permitting it—produces human beings who
are debased, whose characters and lives are less good than they might
be. As the ethical project evolves, views of the good human life become
richer, and engaging in conduct that causes unnecessary pain to others,
including nonhuman animals, comes to appear detrimental to living
well.
Shared capacities for pain allow the extension of ethical precepts be-
yond the human sphere. How could pragmatic naturalism go further?
Trees and statues do not feel pain. Nevertheless, the same strategy can be
deployed. Concern for individuals who do not participate in the ethical
23. The justification of blood sports would thus have to show that the hunted animals
are threatening in some way (foxes attack livestock) or that the activity of hunting serves
some important human purpose (plays an indispensable role in the maturing of young
men). I am extremely dubious that any such attempt at justification accords with reflection
on the conceptions of the good that have come down to us.
deliberation—juveniles, descendants as yet unborn—has been part of

the ethical project from the beginning. Even when the desires endorsed
are quite basic, not maintaining the conditions that permit those desires
to be satisfied by those who will come later is an altruism failure. For the
small bands of the Paleolithic, introducing precepts requiring conserva-
tion of parts of the habitat would have been a progressive step, at least
once population density was sufficiently high to make a policy of un-
restricted exploitation and moving on impracticable. With the more pow-
erful means of environmental devastation available to us today, it would
be similarly progressive for us to commit ourselves to preserving the one
planet we know we can inhabit.
As more refined conceptions of the good human life emerge, a wider
range of desires is endorsed. Communion with unspoiled nature is seen
as contributing to the quality of a human life; encounters with human
artifacts are regarded as similarly enriching. Preserving the possibilities
of experiencing a range of natural environments and of being moved by
great human achievements is not as crucial to those who will come later
in the history of our species as ensuring that they have an atmosphere fit
to breathe, but it is still important. The basic, long-standing precepts
enjoining us to care for the world inhabited by our descendants underlie
a range of principles of conservation.
It is easy to feel that the approach I have outlined is crassly anthro-
pocentric. Precepts cover behavior toward rain forests and great build-
ings because future people are expected to gain from experiencing these
things. The entities to be conserved have no intrinsic value; their special
status is conferred by us, sometimes because of specific ways in which
our societies, and the ethical project, have evolved. Can that really be an
adequate account of the worth of the environment and of our obligations
to conserve it?
Pragmatic naturalism is committed to an even stronger, perhaps shock-
ing, view of the human dependence of value. Anthropocentrism is at
the core of the ethical project—even to the characterization of ethics as a
human project. That project confers all values. Parts of the environ-
ment are to be conserved because of decisions reached in progressive
transitions—but similar decisions lie behind the injunctions that nonhu-
man animals are to be shielded from pain, that people who belong to
different societies are to be accorded the protections available within the

local group, and even that altruism failures within a small local band are
to be remedied. Valuation is something humans do, and we have been
doing it ever since normative guidance became socially embedded.
There is nothing else that can ground value—no external source, no di-
vine being. If it seems arrogant to view our species as the source of ethi-
cal precepts and values, we should recall that, at least in our corner of
the universe, there is no other entity to do the work for us.
§48. Equality and the Good Life

Pragmatic naturalism views us as facing a scaled-up version of the pre-
dicament of the original ethicists. The primary challenge stems from the
need to address the sources of confl ict: the pronounced inequalities
of the contemporary world and the clash of different systems, whose reli-
giously based imperatives are both mutually hostile and override desires
that might otherwise be endorsed (§46). If these problems are not readily
perceived, that is because, in the manner of the envisaged tyrant of §34,
a small privileged subset of the human population can insulate itself with-
out attending to underlying causes. It is overwhelmingly improbable that
the insulation can be maintained for long, given the technological possi-
bilities for violent retaliation now increasingly available to the poor and
oppressed (or to those who claim to represent them). Even if the comfort-
able few assume their security can be preserved in the long term, the
thought of an imaginary conversation, in which they must discuss respeci-
fying the good, on equal terms with the many who live in want, should
concentrate their attention.
Dynamic consequentialism begins with the conception of the good
world (or the better world) we have and considers how it might be modi-
fied in progressive ways. Contemporary ethical and social theorizing
offers possibilities. Utilitarianism, generated from consequentialism
through a sequence of reductive assumptions, identifies a supposedly
objective measure, the degree of pleasure and absence of pain, and ag-
gregates over sentient beings (§45). Inspired by concerns about how dif-
ferences in ascribed utility could be determined, social theorists either
have focused on a supposed proxy for utility (money, subject to various
potential discounting factors for risk and future possession) or have at-
tempted to analyze the common good without presupposing the compa-
rability of subjective experiences.24 One social situation is said to be su-
perior to another just in case there is no individual whose preference
satisfaction is diminished and at least one whose preference satisfaction
is enhanced. This perspective has been the starting point for some of the
most brilliant and illuminating work in theoretical social science and in
social and political philosophy of the past century.25
Existing consequentialist theoretical traditions offer conceptions of
the good in three distinct forms:
A. The goodness of the world is determined by aggregation of some

objective measure for the states of sentient (or human) beings
(aggregate pleasure minus pain, aggregate desire satisfaction).
B. The goodness of the world is determined by aggregating a
publicly observable proxy (e.g., money or surviving people).
C. One state of the world is better than another when no desire
satisfied in the latter is unsatisfied in the former, and at least one
desire that goes unsatisfied in the latter is satisfied in the former.
Versions of A are notoriously difficult to elaborate. Those elaborations

of B focusing on money as the pertinent proxy, while often deployed as
the basis for social policy, are of dubious ethical significance; counting
heads (or survivors) is only slightly less crude.26 Conception C, on the
24. Lurking behind all this are methodological ideas that have sometimes played a use-
ful role in inquiry (through connecting hypotheses with potential tests and observations)
but that have often hardened into restrictive dogma (most evident, perhaps, in some kinds
of behaviorist psychology). Significant parts of economics and social theory remain in the
grip of sclerotic operationalism, even though its philosophical credentials have long been
subject to important critiques.
25. Kenneth Arrow, Social Choice and Individual Values (Hoboken, NJ: Wiley, 1951);
Amartya Sen, Collective Choice and Social Welfare (San Francisco: Holden-Day, 1970);
John Rawls, A Theory of Justice (Cambridge, MA: Harvard University Press, 1971).
26. It is worth reflecting on the extraordinary combination of a very crude basic mea-
sure of goodness (number of survivors) with the excessively refi ned concentration on
wildly contrary-to-fact hypothetical examples (trolley problems) so evident in some philo-
sophical fashions.
other hand, seems to deprive itself of important resources—for even if

we cannot make all comparisons among the states of different individu-
als, we can surely make some (there is little reason to deny, for example,
that the agony you would suffer from a debilitating illness is worse than
the discomfort I feel from a mild itch—and thus no reason to think a
state, in which I lost by having to endure the itch and you gained by
avoiding the agony, would be “noncomparable” with the status quo.)27
All three versions are open to charges of hedonism, since all start with
a focus on individual states or on variables (like money), viewed as basic
because they are proxies for such states. They are opposed by a far more
pervasive conception, or set of related conceptions, that start with a con-
cern for the quality of individual lives (not simply for survivorship), un-
derstand this in terms of fulfi lling the will of a transcendent being,
and suppose worlds are good insofar as they contain wider multitudes of
people who carry out the divine plan.28 Conceptions of this sort typically
define themselves in contrast to the crudest versions of conception B,
juxtaposing the “materialism” of seeking to maximize money (or material
goods) with the “spiritual” emphasis on fulfillment. So we have:
D. The world is better insofar as it contains an expanding set of

spiritually fulfilled lives.29
27. Sclerotic operationalism makes itself evident in the judgments of noncomparability

that are questioned here. I should note that Rawls attempts to reintroduce a fi ner discrimi-
nation among states by using the device of the original position to distinguish Pareto-
incomparable options; effectively, this is to combine the two approaches I have been sepa-
rating, taking imaginative deliberation to serve as a further constraint on the elaboration of
the good. Since I believe that the operational imperative is overworked in this context, I
suppose that the starting point is unnecessarily restricted and that the attempt to articulate
dynamic consequentialism has more resources than Rawls’s procedure acknowledges.
28. Some secular ethical traditions have a similar structure, taking the world to be good
insofar as its inhabitants conform to the moral law. Kantian conceptions obviously fit here.
29. I have formulated the conception in a way that mimics the emphasis on Pareto-
comparability found in (C). Proponents of (D) probably rarely consider whether salvation
for a greater number at the cost of forfeiting salvation for a lesser would be an advance—and
do so because the world modifications that concern them involve what they take to be ex-
pansion of the community of the faithful without loss. (It is unclear how they would react if
it were shown that sending young people out on missionary work is likely both to increase
the number of converts and to involve some lapses on the part of the missionaries.)
Different religious traditions diverge about what it takes for a life to be

spiritually fulfilled.
The inadequacies of conceptions A– C were prefigured in §45 and are
expressible in terms of an insight of conception D. Instead of breaking
individual lives into small intervals in which pleasures are felt, pains
avoided, or desires satisfied, conception D views lives as wholes, evalu-
able as good or bad. Conception D goes astray in grounding the good-
ness of lives in utterly false doctrines. To develop its insight, we need a
surrogate for the dangerous deontology it offers.
Section 45 pointed in a promising direction. Reductionist emphasis
on pleasure and pain, or on satisfied desire, should give way to recogni-
tion of holistic contributions to the value of lives.30 Replacing a game of
pushpin with a reading of The Prelude or a conversation with a soul mate
does not boost the value assigned to momentary experiences, increasing
aggregate utility across a life, but a life enriched by poetry and intimate
conversation can have more value as a whole than one lacking anything
similar.
It is important to tread cautiously. Holism is welcome in accounts of
the good, but not at the cost of insinuating elitism: “Only the privileged—
those capable of running the polis or of appreciating Wordsworth—can
lead truly worthwhile lives.”31 Elitism can be resisted by proposing that
almost all human beings can advance from a state in which the good life
30. In my interpretation, Mill is innocent of the reductionist errors. Although Utilitari-

anism makes it appear that he is making a minor adjustment in Bentham’s position, intro-
ducing a third, and problematic, evaluation of pleasures and pains (the higher/lower dis-
tinction), he is best read differently. His essay/monograph is an attempt to defend
utilitarianism against popu lar criticisms, and it is readily understandable that Mill would
proceed from the version of the position that was most familiar to his readers. Mill’s conse-
quentialism is expressed in his entire corpus, and Utilitarianism is a response to well-
known objections, rather than a defi nitive exposition of what Mill believes. (See my essay
“Mill’s Consequentialism,” in The Routledge Companion to Nineteenth Century Thought,
ed. Dean Moyar [New York: Routledge, 2010], for more discussion.)
31. The focus of ancient authors on the lives of the socially and econom ical ly privileged
is echoed in contemporary discussions and is evident in that liberal democrat Mill. More
than his major ethical predecessors and contemporaries, Mill is influenced by Greek
thought, so that the question of the good life is central for him, and his formulations often
absorb the elitist perspectives of the ancients (witness the language of “higher” and “lower”
pleasures).
is understood in terms of the satisfaction of basic needs (nourishment,

shelter, protection, and so forth) to a condition in which a richer concep-
tion of the good life makes sense for them. 32 This richer conception
recognizes lives as structured wholes, with direction and point. Emanci-
pating the conception of the good life from unfortunate elitist formula-
tions, we can suppose that the crucial factors in evaluating lives are, first,
whether they reach a conception of the person that can bestow direction
and point, and second, whether the aspirations flowing from that concep-
tion are sufficiently realized.
Educated people may find the shape of their own lives (and of their
selves) in a mix of intellectual work, political activity, appreciation of
nature and of art, and (perhaps above all) communion with others. A
major contribution to thought about the good life opposes the thesis that
this pattern—or any single pattern—is appropriate for everyone.33 Valu-
able lives can be structured quite differently. For some people, physical
activity is more important than it is for many intellectuals—for some, re-
placing a game of pushpin with reading Wordsworth might be a distor-
tion and diminution of a life.
A significant advance over Greek thought suggests that the pattern
one adopts for one’s life should be one’s own.34 In opposition to the par-
ent or the community that specifies the model to which a nascent life
should conform, we can envisage enlightened choice among many op-
tions. Recording and publicizing a diverse collection of “experiments of
living” can acquaint those who come later with an ever-wider repertoire
of possibilities from which they can make their own free choice. Ideally,
once people have the opportunity to advance beyond the stage at which
32. Much of the time, Mill endorses this proposal, thinking of human progress as
making the good life ever more widely available. This is evident in the famous phrase
from On Liberty, “the permanent interests of man as a progressive being,” as well as in
the closing pages of the System of Logic, John Stuart Mill, System of Logic (London:
Longmans Green, 1959). It is crucial to the egalitarian concerns of Principles of Political
Economy.
33. Thus, Mill’s emphasis on choosing one’s own plan of life takes precedence over the
very specific ways in which, as an extraordinarily educated man, he conceived of his own
life as obtaining point.
34. Mill’s writings make this advance in many places. The most well-known occur in
On Liberty, particularly in chaps. 1 and 3.
human beings struggle to satisfy their basic desires, 35 the young will be
invited to draw on the ways of living of many different predecessors to
formulate a conception of what matters for them.
This embryonic account of the good life will be developed a little fur-
ther (§50). The immediate task is to see how an account of this sort could
remedy the inadequacies of the conceptions of the good so far consid-
ered. Taking the notion of the good life as central, understood in the
ways just outlined, preserves the insight offered by conception D, while
rejecting the massively false presuppositions of the various religious
articulations of it. The holistic assessment of lives is primary, and positive
evaluation depends on there being a structure, freely adopted by the in-
dividual and, subsequent to that, on the kinds of satisfactions of desire
figuring in conception C, and tacitly in conceptions A and B as well.
This possibility arises only at a particular stage in human develop-
ment—it requires progress beyond the conditions dominant in the early
phases of the ethical project (§20). There are preconditions, some mate-
rial, some social, for any state in which holistic evaluation can come into
play. When life and health are constantly threatened, it is difficult, if not
impossible, to explore possibilities to fashion one’s own life pattern.
People must be able to interact peacefully with one another (and with
outsiders). Furthermore, if their life projects are to have serious chances
of success, they will need not only the toleration of those around them
but also active cooperation with their fellows.
Consider an imaginary social state: Utopia. In this state, each mem-
ber of the human population has a serious chance of living a good life, a
life in which the person can recognize a number of different possibilities
for living, can make a free choice of a project informed by that recogni-
tion, and realize a significant number of the plans, intentions, and desires
central to that project; moreover, the chances of living such lives are equal
across the population. Where the conditions of human life allow no seri-
ous chance of bringing about Utopia, it cannot figure in the theory of the
good. The good evolves, and this is an ideal available to us, but not to
our ancestors.
35. A condition Mill characterizes as the “puerile” condition of mankind; closing para-
graphs of System of Logic.
Pragmatic naturalism’s proposal: Once a particular stage of techno-

logical development has been reached, a world counts as good to the
extent actualizing it would lead us toward Utopia. More exactly, faced
with a number of possible outcomes, we should rank them in accordance
with the expected sizes of the steps they would take in the direction of
Utopia (subject to the proviso that they do not block further progress).
The reference to “expected” size accommodates the fact that the results
of action may be associated with probabilities, and the proviso takes into
account the possibility that a route leading to an unclimbable cliff will
not take us to the summit.
The preconditions of Utopia include two sorts of equality. First, all
individuals begin their lives in a state of material equality, a state in
which they are assured of being able to meet those basic needs that must
be satisfied before free choice of a life project can be a serious possibility.
Second, enough equality must be maintained so individuals are not co-
erced into pursuing a particular life project out of material need. Plainly,
the present enormously skewed distribution of wealth across our species
is glaringly inconsistent with these fundamental forms of equality, and
transactions enabling the poor to obtain reliable access to food and wa-
ter, protection from the environment and from disease, education for the
young, and so on would be steps in the direction of Utopia (subject, of
course, to the proviso that they did not introduce complications making
further advances impossible).
The proposed conception of the good sees progress in the transitions
that generated our more complex conception of possibilities for human
lives, but it also aims to restore the emphasis on equality abandoned in
the most prominent ethical lineages.36 It thus tries to take advantage of
the enhanced options the inegalitarian developments of our past have
given us, while simultaneously removing their distortions of the ethical
project.
36. Some small societies in the contemporary world have preserved the emphasis on
equality, but they lack any serious possibilities for achieving the richer versions of the good
life available to some members of other societies.
§49. Population Size

My references to “Utopia” and to realizing an “ideal” generate an obvi-
ous question. Is the proposed conception of the good a feasible goal? A
skeptic might insist that the recommended combination is impossible:
an emphasis on equality must inevitably diminish the possibilities for a
good life; the proposal seeks joint attainment of conflicting values, where
we can settle only for pluralism.
Ideals that guide us should not be beyond the horizon of realizability.
Even if one had no basis for thinking an ideal unattainable, there might
be no clear available method of attempting to achieve it. Under those
circumstances, it could not play any role in shaping action. In the case at
hand, however, our knowledge of how people in affluent societies have
sometimes been able to choose profitable projects for their lives provides
a basis for spreading this possibility more widely. If skepticism (that
genuine equality of opportunity for the good life cannot be achieved for
all) is rebutted, it will prove possible to identify steps toward attaining
that goal.
Skepticism rests on a fundamentally economic thought: the resources
human beings collectively can acquire are too limited to make possible
for all the opportunities for a good life (in the rich sense) currently avail-
able to the few. Meeting the basic needs of the entire human population, or
meeting those needs as well as supplying the background circumstances
for good lives (educational and health ser vices among them), would be
possible only if the affluent world impoverished itself. Recent studies of
global poverty cast doubt on this pessimistic assessment.37 Yet we should
pose a different question: Given pragmatic naturalism’s conception of the
good, how many people should there be?
Population size lies at the heart of issues about the good, for two basic
reasons. First, we know that indefinitely continued expansion of the hu-
man population would eventually reduce all human lives to a condition
of wretchedness. Second, we know it is possible for local populations to
resist expansion, and even to diminish their size, without giving up op-
portunities for their members to enjoy good lives. Indeed, statistics show
37. See, for example, Jeff rey Sachs, The End of Poverty (New York: Penguin, 2005).
that significant advances in the quality of individual lives—for example,

advances in educational opportunities, especially for women 38—produce
a decline in birth rate. Even if the skeptic were right about the insuffi-
ciency of the world’s resources to deliver the preconditions of serious
opportunity for the good life with respect to the existing human popula-
tion, the ideal might be attainable for an envisaged future population
with diminished size.
Ecologists use the notion of the carrying capacity of an environment,
with respect to a group of organisms. The carrying capacity is the num-
ber of members that group would have at equilibrium: if the actual pop-
ulation size is lower, the population expands; if the actual population
size exceeds carrying capacity, the population declines because of the
insufficiency of resources. We can introduce an analog of this idea. Say
the proper bound on the human population is that number at or below
which it would be possible, given the material resources of our planet
and available technology, to deliver to a population of that size serious
and approximately equal chances for a good life, through an indefinite
sequence of generations. Pragmatic naturalism proposes that the human
population should be at or below its proper bound.
The proper bound is determined by issues not yet precisely resolved.
The substantive notion of the various forms of the good life might affect
its value. If it is important for people to have the chance to retreat from
the presence of their fellows and to enjoy solitary contemplation of the
natural world, that might entail programs of preservation (or restoration)
limiting the budget of resources. Articulating what counts as a “serious
chance” for a good life will affect the amount of resources individuals
and groups require and thus constrain the proper bound. Suppose these
choices have been made, generating a value for the proper bound. The
skeptic is answered. The proposal offers an account of the good for a
population at or below the proper bound; the feasibility of the ideal is
built in.
Is this cheating? The original worry claimed that meeting the pre-
conditions of equal opportunity for the good life for all human beings
38. See Amartya Sen, Development as Freedom (Oxford, UK: Oxford University Press,
1999), esp. 217ff.
would impoverish the lucky few who actually happen to enjoy richer
prospects for their lives, in a quixotic venture of trying to satisfy every-
one’s most basic needs. That worry has been allayed by limiting the size
of the human population. Can Malthusian restraint be achieved with-
out threatening the quality of the lives of individual people and fami-
lies? Quite reasonably, the skeptic insists that begetting and rearing
children is central to many people’s conceptions of what matters most
to them: consequently, confining the human population to the proper
bound directly and dramatically reduces the chances for good lives.
There are two ways to articulate the objection. One contends that, in a
population already at or below the proper bound, maintaining that popu-
lation size would significantly limit the quality of human lives. A more
modest version argues that any possible route to a population below the
proper bound requires diminishing the quality of life for the generations
traversing that route.
Although family life is central to many people’s conceptions of their
lives, it does not follow that limiting family size must interfere with their
prospects for a good life. As an elementary consequence of the biology
of reproduction, in a world where no couple produced more than two
children, the human population could not increase (and would almost
certainly decline). Thus, the stronger objection can be sustained only if
an ethical restriction of family size to two children excludes the oppor-
tunity for a good life. Plainly, smaller families are vulnerable to contin-
gent events that might diminish the quality of the lives of their members—
because of the impact of the death or disability of children. Nonetheless,
larger family size offers no guarantee against such contingencies, nor
can it necessarily relieve the pain accompanying the death of a child:
children are not intersubstitutable. 39 The most skeptics can rightly
conclude is that an ethical constraint on family size might lower some
people’s chances for a good life, through increased susceptibility to
hostile fortune. In worlds where the human population is maintained
at or below the proper bound, however, an equitable distribution of
39. It should be recognized, however, that the loss of all one’s children, as in cases
where all are killed in war, does strike people as especially tragic. That reinforces the
thought that smaller families are more vulnerable to the whims of fortune.
resources lowers the probability of the major kinds of disruptive

events—for example, through preventative measures against diseases
killing millions of children annually. The stronger version of the skepti-
cal objection fails.
Consider the more modest version of the objection, which focuses on
the trajectory the human population would take to diminish the size
of the population until it is at or below the proper bound. Suppose the
route to the proper bound requires a specific claim about the good: in
worlds at larger sizes, it is good for couples to limit their offspring to a
single child. Assuming it is a greater tragedy to lose all one’s children
than to lose all but one, the imagined ethical constraint increases the
probability that some lives will be devastated. Pragmatic naturalism must
claim that reflective people, envisaging the human future, would be pre-
pared to endorse the ethical constraint, even while recognizing the
potential losses. Their endorsement might be based on appreciating that
not imposing the constraint would continue a situation in which many
people lack the opportunity for good lives, as well as on seeing how a
more egalitarian distribution of resources would provide some protec-
tion against hostile contingencies. Yet, even if one rejected the ethical
constraint and the route associated with it, the ideal would be realizable
through a slower process. An ethical constraint restricting family size to
at most two, coupled with encouragement of voluntary policies of having
only a single child, would steadily decrease the human population. A world
in which the more liberal constraint was honored would be a world in
which the ideal of a sustainable provision of serious and equal chances
for a good life for all members of the population, at or below the proper
bound, eventually became feasible.
Does the current size of the human population exceed the proper
bound? That is an empirical question, and one for which an extended
analysis would be required. Present uncertainties about global warming,
and its potential consequences for the provision of food and shelter, would
have to be resolved to arrive at any convincing judgment. The appropriate
response to the claim that the present human population exceeds the
proper bound is the Scottish verdict: not proven.
One of the most important discussions in ethical theory in past
decades has been the study of consequentialism when population size
varies.40 Begin from a framework analogous to that of orthodox utilitari-

anism: the goodness of a world is measured by aggregating the values of
the qualities of the lives of the human beings who live within it. Accord-
ingly, a world in which a large number of people who live lives as good as
any we can envisage would be inferior to one in which a vaster number
live lives that go slightly less well: if N (population size) and U (the value
attaching to each individual life) are both large, if U* is less than U by
some tiny amount, and if N* is vastly greater than N, then N*U* > NU.
Iterating the process of comparison, a world with a truly gargantuan
number of people whose lives were just above the level at which they
are barely worth living would be superior to the original state. This is
the Repugnant Conclusion, and it seems any viable version of consequen-
tialism ought to reject it.
The problem was generated by using a measure of the goodness of
worlds that, like classical utilitarianism, adds up the values in the individ-
ual elements. Another approach would be to measure by averaging. One
world counts as superior to another just in case the average value of the
lives lived in the first is greater than the average for the second. Although
this revision avoids the Repugnant Conclusion, it entails other unpleasant
judgments. Imagine Eden, a very small world, containing just two people
who live lives of the very highest quality. Compare that with another world
with vastly more people, in which everybody lives a life just infinitesimally
less good than that enjoyed by Eden’s lucky pair. According to the averag-
ing approach, this latter world is inferior to Eden, despite the fact that the
aggregate value of the lives within it is enormously higher. That, too,
seems wrong—for, even though aggregation does not count completely
(witness the Repugnant Conclusion), it cannot be completely dismissed.
Adam and Eve would surely be doing good if they aimed to replace their
Paradise with the more abundantly stocked world, rich in lives of extraor-
dinarily high quality (albeit just slightly inferior to their own).
These two thought experiments pose a challenge: find another mea-
sure of the overall quality of worlds, avoiding these counterintuitive
40. The seminal work here is part 4 of Derek Parfit, Reasons and Persons (Oxford, UK:
Oxford University Press, 1984). Parfit’s brilliant posing of the problem has inspired many
efforts to solve it.
implications and giving both aggregation and averaging their due. A

formal treatment shows there is no solution to the problem.41 The im-
possibility proof locates the fundamental difficulty. So long as it is sup-
posed that the quality of a human life can be represented by a single
number, the Archimedean property of the numbers will hold sway: how-
ever tiny r may be, and however large s may be, there will be some integer
N such that Nr > s. This is the mathematical truth underlying the Re-
pugnant Conclusion, and posing difficulties for aggregation, and there
is no way to limit the trouble this mathematical feature of aggregation
poses.
A formal way to respond would introduce a discrete measure of the
quality of human lives. At its simplest, this would propose that lives go
well or they do not, and the thought of tiny frustrations slightly decreas-
ing the value assigned to a successful life is misguided: if someone’s life
under particular conditions is assessed as one that goes well, and if we
now consider slightly less favorable conditions (an extra headache, a tiny
loss, or whatever), the same discrete measure is assigned; small blem-
ishes do not turn a valuable life into a poor one. That, however, is to
overlook the need for some comparison. Even with respect to two lives
that go well, one can make discriminations—but the discriminations are
secondary. I propose a two-dimensional account of the value of lives.
The dominant dimension accords with a discrete measure: discreteness
is the better part of value. The secondary dimension has a continuous
measure, one recording the minute satisfactions and discomforts of
human existence. Lives counting as successful on the dominant (discrete)
dimension submit to secondary comparison on the subordinate (con-
tinuous) dimension.
The approach of the earlier parts of this section gives substance to
the formal treatment. At early stages of human life, the achievements
pertinent to the dominant dimension have not been made. Early in
the ethical project, the lives of people are measured only by the vari-
ables to which utilitarians point. The two-dimensional account comes
into play once it is possible to provide the material basis for the good
life for all people, and, at that point, the satisfactions measured on the
41. See my article “Parfit’s Puzzle,” Noûs 34 (2000): 550–77.

subordinate dimension are relevant only insofar as they affect the pos-
sibility of doing well on the primary dimension. The formal approach
fits the progressivist and egalitarian aspects of pragmatic naturalism’s
conception of the good. The initial conceptions of the good are di-
rected at obtaining high values on the continuous dimension; later, high-
enough values can be achieved for all to make richer considerations of
living well come into play; thereafter, the good is achieved by realizing
the opportunity for all human beings to have a chance at a good life, and,
beyond that, by trying to increase the number of those who succeed in
living well.
§50. Aspects of the Good Life

To articulate the proposal further requires that more be said about its
egalitarian commitments and about the factors affecting the goodness of
lives. I close this chapter by saying a little about both issues.
Aiming at exact equality with respect to material resources or for
chances of living well is plainly quixotic. Even in Utopia there would be
differences in people’s early lives affecting their probabilities of formu-
lating a satisfactory course for themselves and of successfully pursuing
it. What kinds of differences matter? I propose that differences do not
matter when well-informed parties do not see them as mattering. Two
people can know that one has had more chances than the other and still
see this as no cause for regret or resentment. Education plainly affects
opportunities for choices about one’s life. Even though you and I have
been to different schools, and even though your school clearly provides
a richer range of options for you than mine does for me, I may see the
difference as a tolerable accident, not affecting the fairness of distribu-
tion. Both schools do well; while your options are richer than mine,
mine are also rich. The differences are small in comparison with other
features that affect our lives. With similar amounts of effort, you may
have slightly more chance of living well than I do, but a bit more exertion
on my part would compensate. Furthermore, life has vicissitudes that
could swamp the small increment in your chances for success. Strug-
gling to render our opportunities exactly equal seems pointless: my
slightly inferior chances of living well are very good, and potential
factors thwarting my success would be unaffected by attaining exact

equality.
The injunction to promote equality of opportunity recognizes a range
of tolerable difference. Utopia requires that the differences lie in the
tolerable range—and also requires eliminating them, where that can easily
be done. More intricate issues arise with respect to the second zone of
vagueness in my account, the idea of the good life.
The possibility of formulating one’s own conception of what matters
most is one component of a good life—subject, of course, to the pro-
viso that one’s choice would not interfere with the like choices of oth-
ers.42 It follows that our ability to find our own life project (“our own
good in our own way”) depends, at least up to a point, on an ability
to reflect on the various possibilities, on the characteristics that have
proved to be relevant to realizing them, and on our own propensities
and talents.43 Further, it is crucial that the options not be framed in
terms of an arbitrary deontological perspective, effectively ruling out
the most attractive. It is not simply the law that has often stood in
the way of people’s choice of the life projects they most wanted, but
also a prevailing view that particu lar ways of living are wicked or sin-
ful, where these evaluations have their basis entirely in the alleged de-
liverances of the transcendent policeman. An important part of the
proposed conception of the good is that the circumstances in choice-
of-life conceptions be entirely free from the distortions of religious tradi-
tions. Liberation cannot be achieved unless the context in which people
choose patterns of life for themselves is thoroughly secular, beyond the
myths of the world’s religions, beyond their cramping effects on indi-
vidual choice, and beyond their divisive intrusions among groups of
people.
42. In shaping the conception, and in the formulation given here, my debts to Mill, and
secondarily to Dewey, should be apparent.
43. One might think that our freedom in this regard is always enhanced through the
articulation of further possibilities, so that, the more “experiments of living” someone can
conceive, the more the choice is autonomous. As a matter of psychological fact, I doubt that
this is correct. Too many options can be overwhelming. It remains true that radically new
potential ways of living can show us opportunities that had previously been beyond our
horizons, so that Mill’s plea for further experiments remains cogent.
One par tic u lar way in which some religions have distorted our
conception of the human good is by allowing for possibilities of valu-
able lives detached from other people. Solitary communion with tran-
scendent beings (or with the universe), exemplified by hermits who
live in remote places or those who pledge themselves to silence, is
viewed as one way of living a good life. The source of the value here is
surely the attunement of the individual psyche to the transcendent,
and when the myths about the transcendental realm are abandoned,
the idea that lives can be made significant in this way should go too. It
is equally important to repudiate secular ways in which this distort-
ing idea continues to manifest itself. Some of the most militant oppo-
nents of the world’s religions commend participation in the project of
understanding the natural world as an especially valuable way to live.44
Although there is an important insight here, it needs to be carefully
understood. The great discoverers achieve two things: they enjoy private
states of recognizing hitherto uncomprehended aspects of nature, and
they facilitate the understanding of others. The states of understand-
ing, while superior to the momentary satisfactions of the hermits (be-
cause based on genuine achievement rather than on illusion), are not
the primary determinant of the value of the discoverer’s life. Instead,
value accrues through the contribution to understanding on the part
of others.
To take this perspective is to emphasize the centrality to the good life
of our relationships with other people and of our contributions to their
lives. Refined theoretical contemplation has its place among the catalog
of factors that promote the good life precisely because of its potential to
promote the value of good lives. Consequently, the life of the priest or
scientist, the doctor or nurse, the teacher or social organizer, the tireless
participant in the maintenance of community and family, become valu-
able in similar ways, through the various human relationships the per-
son’s actions sustain. The emphasis on individual freedom, on the abil-
44. This is prominent in the writings of Richard Dawkins, Unweaving the Rainbow
(Boston: Houghton Miffl in, 1998). But the idea is very old—Aristotle’s fi nal chapter of the
Nicomachean Ethics, (Terence Irwin trans. Indianapolis: Hackett, 1999) sounds a similar
theme.
ity of each of us to choose our own conception of what matters, needs to

be accompanied with recognition that any choice that does not incorpo-
rate interactions with others and see their good as involved with one’s
own is inadequate.45
This judgment recapitulates the historical story of Part I. Ethics be-
gan with our joint reflection on how to live together. From the experi-
ments of living of the distant past emerged the forms of higher-order
altruism (§21)—and the thought of joint action as proceeding from atten-
tion to the aspirations and interests of one another as especially valuable.
The first ethicists were probably concerned with the altruism failures
that arose with respect to basic needs, but their efforts led to an evolved
conception of the good life, one in which our interactions and relation-
ships with others are fundamental. That aspect of the evolutionary pro-
cess should be seen as thoroughly progressive.
To recognize progress here is not to suppose that the evolving forms
of human life have made it easier to achieve the valuable relationships,
marked by higher-order altruism and participation in a rich variety of
joint projects, that are principal constituents of the goodness of lives.
Indeed, as human societies grow larger, as they emphasize the division
of labor, even as they proliferate possibilities for living well, they may
make it increasingly difficult to actualize those possibilities and to live
lives with real value. Religious myths may distort the conception of the
good life—through proposals to the effect that what really matters is our
relationship to a nonexisting being—but they are not the only culprits.
As we become coordinated parts of larger social machines, it is easy for
us to think that what we achieve is determined by the specific contribu-
tions we make, the deals we bring off, the things we discover or invent or
compose, the tasks of whatever kind we complete, without any reflection
on the impact on other people. Our self-conceptions are further debased
when we measure our worth in terms of the proxies for success in any of
these particular directions, the cash rewards we receive for doing them
and the trophies we thereby amass. We miss the fact that all of this effort
45. Thus Dewey advances on the Millian formula from On Liberty, when he opens The
Public and Its Problems by posing the issue in terms of the freedom from coercion of joint
projects.
obtains its significance from effects on, or more exactly contributions to,
the lives of others. We miss also the important point that, independent of
any large-scale public success, lives may be interlocked in mutual depen-
dence and mutual contribution and thus be genuinely and completely
worthwhile.
From the moment at which the issue of how to live well surfaced
among the ancients, it has been vulnerable to three types of major error.
The hedonist mistake is to decompose our lives into sequences of mo-
mentary experiences and measure value by the balance of pleasures and
pains. The individualist mistake, prominent in some religious traditions
but also retained in some versions of secularism, proposes that some
particular nonsocial condition of the individual—the receipt of divine
grace, the making of great discoveries, the amassing of wealth—is the
major source of value. The elitist mistake, already evident in the restric-
tion of the question to the male aristocrats of the polis, is to suppose that
something very large and uncommon is a precondition of a life’s going
well. By contrast, in the approach I have taken, good lives are in princi-
ple available to almost all members of our species.46 Philosophers tend
to talk grandly of “life projects,” as if the good life required both a type
of intellectual reflection and an exalted focus that can be managed by
only a select few. Instead, I offer a schematic account of the good life that
celebrates the ordinary. Although in almost all places at almost all times,
people have been coerced or led into lives that should not be counted as
worthwhile, what they have lacked are certain basic forms of freedom,
everyday awareness of possibilities, not exceptional resources or un-
usual talents. Moreover, in many times and places, ordinary people whose
lives are permeated by actions with and for others have sometimes, if not
often, lived well.
I have pointed only to aspects of the good life: freedom of choice, lack
of deontological encumbrance (most evidently the distortions of reli-
gions), joint activity and reciprocal relationships with others. To indi-
46. The exceptions are those whose cognitive and emotional possibilities cut them off
from fully developed relationships with others. This is, I believe, why we fi nd prenatal ge-
netic testing for some sorts of traits a merciful way of proceeding. I have discussed these
questions in more detail in the later chapters of The Lives to Come (New York: Simon and
Schuster, 1996).
cate these features is not to provide an articulated theory, but merely to

avoid some dangers that have beset discussions of what makes lives
worthwhile. That is, I hope, enough to develop the proposal respecify-
ing the conception of the good, with which this chapter has been princi-
pally concerned.
chapter 9
Method in Ethics
§51. Varieties of Ethical Change

Although each of us acquires, early in youth, an ethical code from
older members of our society, our ethical convictions and attitudes do
not remain constant throughout our lives. There have been societies
(most probably in the distant past) in which members were forbidden
to add or subtract from what they had been taught, and others, more
common, in which only additions, articulations of the group lore, were
allowed. That is not the way we live now, nor is it the way in which our
species has lived throughout most of recorded history. In recent mil-
lennia, societies have equipped their young members with starting points
for individual ethical development and exploration, permitting them
to add, revise, refi ne, and subtract. Few people who grew up in the
English-speaking world in the 1950s have retained exactly the constel-
lation of attitudes toward sexual behavior originally passed on to us
by parents and other ethical teachers. Only the most confident and un-
imaginative believe they acquired, early on, the complete, correct ethical
system.
In ethics, as in biology, ontogeny needs to be distinguished from
phylogeny. Ontogenetic changes are those occurring in an individual
330
Method in Ethics 331
life, from the beginning of ethical consciousness to its termination.

Ontogenetic changes may produce phylogenetic change, either when
the explicit discussions of the next generation about what should be
passed on involve discussants who have made parallel shifts during
their ethical lives, or, without public discussion, when similar modifi-
cations occur on a broad scale, so the amended code passed on by par-
ents is reinforced by other members of the community. Plenty of onto-
genetic changes have no phylogenetic impact, most notably when people’s
articulations of an initially shared ethical code diverge from one another.
You and I may begin with a vague maxim, one needing more precision
in situations we both encounter, but your precise version may be at
odds with mine; indeed, society-wide emendations may be so various
that the vagueness persists as part of ethical counsel. How exactly the
distribution of ontogenetic changes bears on phylogenetic change varies
from society to society, depending on the role of prevalent religions or
the presence of officially designated ethical teachers. When particu lar
people, or institutions, dominate the training of the young, the onto-
getic changes taking place in large segments of the population may be
irrelevant.
Much of the time, people are not moved to change. Ethical discus-
sions often fail to mention the obvious fact that large parts of our lives are
based on routines and habits. You have acquired an ethical code that set
up patterns of behavior, associating them with roles, which, if you re-
flected, you would understand yourself as playing. Your lack of thought
need not be problematic; indeed, to step back and pose questions about
habits or roles might be at odds with the ethical code you endorse.
Sometimes, however, you do deliberate. The current occasion provokes
thought, perhaps because a maxim you accept seems to dictate an action
contrary to the one produced by following routine: you are on your way
to carrying out your role (as friend, or spouse, or worker), and you en-
counter someone who plainly needs help.1 The role-sanctioned activity
1. As psychological experiments have shown, commitment to the role may be so strong

that people do not deliberate, even though their background ethical code calls for delibera-
tion in the circumstances. See John Sabini and Maury Silver, Moralities of Everyday Life
(New York: Oxford University Press, 1982).
is interrupted, and you must decide if the maxim has sufficient force to
suspend it.2 Deliberation need not lead to any modification—for the
code, as it stands, may give a clear directive; it may, for example, pro-
nounce on the importance of giving aid to people in the predicament you
see before you, assigning that priority over the role-associated duties.
Yet though episodes of this sort need not lead to ontogenetic change,
they can do so.
Imagine your habitual performance is interrupted, and you are brought
to think about how to go on. You canvass the code you presently endorse,
but there is no unambiguous result. Perhaps there are conflicting vague
maxims with no clear priority. Reflecting on other facets of the code, and
emotionally reacting to the situation, you introduce a priority, perhaps
making one or both of the maxims more precise than they previously
were. The judgment about what to do remains stable in the aftermath of
your action. As you explain what you have done to others, particularly
to those your action affected, their reactions do not prompt a change of
mind. From this point on, the claim about relative priority and the some-
what sharpened maxim(s) belong to your code.
Other types of ontogenetic change do not begin with thoughtful de-
liberation. Your ethical code contains no prohibition against speaking
in a par ticu lar way to members of a par ticu lar class. You are accus-
tomed to talk that way, until one day your remarks are met with protest.
Somebody affected by your words shows you, convincingly, how pain-
ful it is for him, and, in appealing to the authority of the code, you feel
your response is feeble. Suddenly and unexpectedly, you are moved to
sympathy with a person whom you have previously seen under the
shadow of a label, and your regret about what you said translates into a
resolution not to repeat such phrases. You have modified your ethical
code by inserting a prohibition where there was none before. The inter-
actions that fuel changes like this are dim echoes of the original ethical
discussions.
For most people all the ontogenetic changes made in ethical codes, and
for all people most of the ontogenetic changes made in ethical codes,
2. See Dewey, Human Nature and Conduct (Amherst, NY: Prometheus Books, 2002),
52ff., 103–5.
are normal changes. 3 Given the ethical resources she has, an agent faces
a puzzle, and the ontogenetic changes she makes are the products of her
efforts to solve the puzzle. One cannot, of course, speak of a “solution”
to a puzzle, without some standard of correctness: what makes life hard
for the agent is the sense that there is some way of going on that is right
and others that are not. People do not seek “mere change.” Although
ethical lives are largely matters of following the precepts and patterns of
our ethical codes, they are also—and importantly—ventures in trying to
improve those codes, by recognizing and solving puzzles that arise for
us. As individuals, we solve puzzles by making progressive shifts in the
codes to which we subscribe. Normal ethical change consists in attempts
to find a way to amend or extend the code, one refining the functions at
which it is directed.4 Puzzles arise within an ethical practice in which
a particular set of functions is to be discharged; a solution to a puzzle
modifies the code so that some of these functions are better served and
none is worse served. Puzzle solving occurs only when there is no func-
tional conflict (§36). Revolutionary change, by contrast, is marked by the
presence of functional conflict.
How can normal phylogenetic change occur? Perhaps as the product
of uncoordinated ontogenetic changes (although this would be highly
unlikely at early stages of the ethical project). Imagine most people encoun-
ter a type of situation where the ethical code they endorse fails to provide
clear counsel; through these encounters, they amend their codes in simi-
lar ways; in the next generation, the lore passed down from parents to
children absorbs the modifications, and, because the amendments are so
similar, the parental advice is reinforced in wider socialization. Yet if a
type of situation arises frequently, challenging individuals to articulate
their ethical code through puzzle solving, a public response can occur.
Especially when people respond differently, ethicists—including social
critics, religious teachers, and individuals with intimate knowledge of
3. My terminology of “normal” and “revolutionary” change plainly derives from Kuhn,

The Structure of Scientific Revolutions (Chicago: University of Chicago Press, 1962 and
1970).
4. Ethical agents do not conceive things in this way. They want the “right” solution.
According to pragmatic naturalism, the notion of rightness must stem from the concept of
progress (Chapter 6).
the pertinent type of situation, as well as philosophers—attempt to iden-

tify the proper solution to the puzzle. If the ensuing conversation
achieves consensus on a way of amending the core ethical code the par-
ticipants share, the result will be a phylogenetic change.
Chapter 6 identifies a difference between reaching agreement on an
amendment and solving the puzzle that sparked discussion. A solution
refines the functions of the ethical practice in which the conversational-
ists are engaged. There is a set of functions to be addressed, and some
are discharged better by the amendment and none is discharged worse.
Consensus is by no means guaranteed to meet this criterion. Public con-
versation goes best if the conclusions reached are based on processes
likely to yield functional refinement: to discharge some existing func-
tions better and none worse.5
At this point, we can appraise a method underlying some of the best
philosophical discussions of ethical puzzles. Guided by an analogy be-
tween ethical and scientific argumentation, philosophers seek reflective
equilibrium between general ethical principles and so-called intuitive
judgments.6 Sophisticated versions of the analogy do not suppose intui-
tive judgments correspond to the statements provoked by observation,
thereby providing access to sources of external constraint on ethical in-
quiry. They regard “intuitions” as judgments people socialized within
a par ticu lar code would be inclined to make about individual cases.
Effectively, this is to embed the method of reflective equilibrium within
a dynamic approach to ethical practice, one supposing users of the
method already have an ethical code. The task is not to justify the code
from scratch but to identify successful ways of modifying it, to increase
its internal coherence. Puzzles are generated when parts of an ethical
code jar with one another, and the search for equilibrium, when done
thoroughly, places the local conf lict within a broader perspective of
principles and responses to par tic u lar situations, some real and some
5. Appeal to religious texts is highly unlikely to satisfy this constraint on public ethical
discussion, since there is no positive correlation between the deliverances of those texts
and the refi nement of major ethical functions. In fact there is probably a negative
correlation.
6. A classical discussion of this method is given by John Rawls, A Theory of Justice
(Cambridge, MA: Harvard University Press, 1999), 18–19, 42–45.
imagined, responses expressing the characteristic habits, emotions, and

dispositions the code supplies.7 Philosophical analysis seeks the best
way of achieving overall harmony, not treating the “intuitive judgments”
as apprehensions of external constraints, but approaching the problem as
one of internal tension (in the extreme case: inconsistency) within ethical
practice.
Is this method likely to solve ethical puzzles? Agreement on a modifi-
cation is distinct from refining the functions of ethical practice, and thus
a consensus proposal is not necessarily a solution. By the same token, an
optimal smoothing out of tensions within an ethical code, the most har-
monious amendment of it, might not improve its ability to discharge
the underlying functions (indeed, the modification might make matters
worse). In both cases, divergence stems from global trouble: if consensus
fails to solve the puzzle, there must be some recurrent blindness among
the participants in the conversation; if the broad search for coherence
fails to deliver an amendment enhancing the code’s functioning, some
pervasive feature of the code must be at odds with the underlying func-
tions.8 If discussants are thinking clearly, we expect their combined ef-
forts to correct for individual biases, errors, and prejudices—we think of
many heads as being better than one. If the ethical code is relatively well
attuned to its functions, we should anticipate that a global exploration
would yield improved functioning—and hence view the method of reflec-
tive equilibrium as reliable, unless our background situation is seriously
problematic.
So: a qualified endorsement of a prominent approach to contempo-
rary ethical discussion. 9 The method of reflective equilibrium might
be further explained and refined through more precise specification of
7. Many of the “thought experiments” beloved of analytic ethics provoke serious doubt
whether responses express any socially inculcated skill. The situations are too remote from
the mundane situations in which ethical judgment is exercised. I am grateful to Gerd
Gigerenzer for discussion of this point.
8. This was probably so for some codes considered in Chapter 6: the Greek honor code
and the Puritan code defending slavery seem systematically blind to the bungling of the
original function of ethics.
9. One carried out with great skill in the writings of some phi losophers, for example,
John Rawls, Thomas Nagel, and Peter Singer.
the virtues of coherence to be attained, and the task of improving our

understanding and practice of the method might be advanced by using
the criterion for puzzle solution to show why particular types of coher-
ence are pertinent. Yet that criterion suggests a different method for
normal phylogenetic change, one to which the search for reflective equi-
librium can be viewed as a preliminary approximation. Since the funda-
mental condition of puzzle solution is refinement of functioning, we can
envisage a more direct strategy:
1. Identify the functions the ethical code is to serve.

2. Show how an amended code directs action in the situation that
gives rise to the puzzle.
3. Show how the amended code improves discharge of some func-
tions without compromising any others.
The indicated method preserves all the virtues of seeking reflective

equilibrium, while also forestalling the possibility that some pervasive
factor introduces a discrepancy between the functions and the elements
of the code. Of course, when there is no discrepancy, global attention to
principles and intuitive judgments (formed through exercise of disposi-
tions the code inculcates) will represent the ways in which the underlying
functions are served—which is why the method of reflective equilibrium
is typically a good approximation.
The strategy just outlined is sometimes available: in contexts of nor-
mal change. For revolutionary change it breaks down. The reason is evi-
dent: its success presupposes the existence of a way of amending the
code that will refine some functions while leaving others no worse than
they were before. When functions conflict, the presupposition fails: there
is no modification of the ethical code that will (a) give direction in the situ-
ation confronted, (b) refine some function(s), and (c) leave other functions
uncompromised. Revolutionary changes are those transitions in which
ethical codes are modified in ways involving both gains and losses with
respect to functions the codes endorse.
Ethical codes may tacitly treat their underlying functions as elements
of an unordered list: all these functions should be satisfied as completely
as possible, and no function is given higher or lower priority. Alterna-
tively, they may already embody judgments about relative importance:

these functions are to be satisfied to this degree provided others are
satisfied to different (lesser) degrees. In the former case, revolutionary
change will introduce an assignment of priorities where there was none
before; in the latter, it will be generated by the impossibility of honoring
the present ordering of priorities, and it will introduce a revised order-
ing. Neither type of transition can be defended by appealing to the
method for normal change, since there are alternative ways of introduc-
ing priorities or of revising the priority ordering, all of which, from the
perspective of seeking refinement of some functions without compromise
of any, are equally justified (or unjustified) relative to the original form of
the code.
Section 40 offered a standard for assessing inferences: good ethical
inferences are those likely to generate ethical progress. If all ethical
change were normal change, invoking that standard would suffice. The
multiple functions of ethical practice, and consequent possibilities of
functional confl ict and ethical pluralism, opened the door to further
worries about pragmatic naturalism (§43). When functional confl ict
arises and revolutionary changes lead to new forms of ethical practice,
can the resultant codes be seen as progressive with respect to the earlier
ones, and can their adoption be defended through nonfallacious modes
of inference? Consider, again, a challenger who asks why he should
be bound by the ethical project. If the challenger questions a normal
change, the strategy outlined above will yield nonfallacious modes of in-
ference to answer him. If he is concerned with the ethical enterprise as a
whole, it will be legitimate to demand that he present some alternative
(other than the hominid state from which the ethical project liberated
us). These are the ways in which the challenge was previously turned
back (§§40, 42).
Trouble arises from an intermediate case. When functional conflict
erupts, the challenger questions any proposed resolution. In doing
so, he is neither casting doubt on modifications of the code defensible
by inferences meeting the straightforward standard (suited to normal
changes), nor rejecting the ethical project in toto. His point is that within
that project, there are other ways of going on, different ways to assign
priorities among the competing functions, and those can serve him as
alternatives, from whose perspective he can query the choices actually

made. He raises the specter of rampant pluralism (§43).
We need a method for justifying revolutionary change. The following
sections develop one.
§52. Method and the Good

Start with an apparently dangerous proposal. Section 45 suggested that
the unfolding of ethical practice could be understood in terms of the
evolution of the good. The idea that proper ethical inferences are those
likely to generate ethical progress founders on instances of revolutionary
change, because of functional conflict. Substantive normative concep-
tions of the good offer proposals to resolve functional conflict. Very well.
Use a proposed conception of the good to assess the progressiveness of
transitions, and apply the old idea of proper inference as promoting
progress, given the determination of progress so achieved. A proposal
about the good can be the starting point for articulating a standard for
inferential justification, and methods shown to accord with that standard
can then be used to support the proposal.
The danger is obvious. Using the account of the good to frame the
methods to be employed in ethical discussion, and then defending the
account by appeal to the methods so generated and approved, looks bla-
tantly and viciously circular. Isn’t this simply a matter of justifying a par-
ticular conception of the ethical project by taking it for granted at the
beginning? Or, to take up skeptical concerns explicitly, isn’t this a game
challengers can play equally well? Despite the importance of these ques-
tions, an attempt to address them will be postponed. The idea will first
be developed on the basis of the conception of the good proposed in the
last chapter—a concrete instance of the apparently circular strategy will
show it to have justificatory force. The conception of the good offered
in Chapter 8 will generate ideas about ethical method, and then we shall
consider the possibilities for defending the conception by using the
method that emerges. Until we return, at that point, to the skeptical
challenge, it is a (small?) consolation that, as with the treatment of nor-
mal change, there is an apparent analogy with the search for reflective
equilibrium.
Consider some fundamental features of the conception of the good

offered in Chapter 8. Today we face a scaled-up version of the circum-
stances that originally provoked the ethical project. Instead of conceiv-
ing altruism failures as arising within a small band, and calling for remedy,
we should think of the current membership of our species as a society in
which altruism failures occur on a spectacular scale—and equally demand
attention. The original function of ethics, that of remedying altruism
failures, is thus given priority. To propose Utopia as an ideal (§48) takes
equality of opportunity for a worthwhile life as a central constituent of
the good. The human population should be limited so as to render this
form of equal opportunity possible. Joint projects and higher-order al-
truism are important features of worthwhile lives (§50). The world at
which we aim would thus be one in which all individuals acknowledge,
and try to promote, the existence for each other person of a surrounding
set of people with whom the person engages in cooperative activity and
with whom that person enjoys long-term relationships.
These features translate the original ethical predicament into the con-
temporary situation. The first ethicists focused on the altruism failures
within a small group, treating the members of that group as equal with
respect to the simplest preconditions of the good life (satisfaction of
basic desires), and seeking the cooperation of all with all. When the perti-
nent population (the set of people whose actions impinge on one another)
expands dramatically—to include all of us—and when the evolution of
ethics has bequeathed to us a richer conception of the good life, the goal
is modified: equality with respect to basic needs gives way to equality of
opportunity for a worthwhile life; the cooperation of all with all gives
way to the fostering of cooperation within subsets of the human popula-
tion (because, when the “group” is numbered in the billions, the coop-
eration of all with all on a wide range of ventures is manifestly impossi-
ble). Essentially, then, the conception of the good proposed results from
an attempt to renew the original project of ethics, while retaining some
of the functions since introduced.
Consider the method employed by the early ethicists. Normative guid-
ance was socially embedded: group members discussed potential ways
of orga nizing their lives together. They deliberated under conditions
where all were present and all were given equal voice, attempting to arrive
at rules all could accept. It will be useful to have a term for these condi-
tions: call them the circumstances of mutual engagement. Mutual en-
gagement was well-suited to finding a solution to the social problems
generating the ethical project. Acquainting group members with the
needs and difficulties of their fellows, giving them equal status in the
conversation, and aiming at solutions everyone can endorse constitutes a
good strategy for coping with the altruism failures manifested in social
tension.
Chapter 8 proposed to renew the original function of ethics and scale
up the initial conception of the good. Likewise, our ethical method, for
use in cases of revolutionary change, can scale up the original method of
the fi rst ethicists. We should seek a notion of mutual engagement as
well suited to the renewed ethical project as the original version of mutual
engagement—the deliberations among band members—was to the origi-
nal venture. We need an analog of those constructive conversations out
of which the earliest rules for conduct emerged.
On the face of it, this suggestion must appear ludicrous, for any
actual conversation among all affected individuals—that is, among all
human beings—is impossible. Public ethical deliberation, however, can
proceed by attempting to simulate a conversation of the pertinent kind.
Faced with functional conflict, so revolutionary change is in order,
public contributors to ethical discussion are judged by their ability to
ground their proposals in mutual engagement: that is, to introduce the
considerations and lines of reasoning that would be brought forward to
achieve consensus were the entire human population to participate,
under conditions of mutual engagement, in a conversation about the
regulation of conduct. The italicized phrase is crucial here. An ethical
discussion seeking to replicate the conversation that would occur if the
entire human population were simply brought together in some vast
arena would be a useless exercise in cacophony. Because of their exist-
ing dispositions to psychological altruism, limited though these were,
because of the pressures on the group and the perceived need for joint
action, the original ethicists were forced into mutual engagement with
people who lived beside them every day. For us, however, mutual en-
gagement is not automatic (except with respect to small subsets of the
human population). If the proposal about ethical method is to have
substance, a clear sense must be given to a notion of mutual engage-

ment, suitable for us.
The conditions of mutual engagement will be specified by incorporat-
ing insights from the historical understanding of the ethical project, by
combining them with a proposed conception of the good, and, on this
basis, devising rules to govern the imagined conversation. Worries about
circularity emerge precisely here, and, even before we have introduced
rules, and so given substance to the concept of mutual engagement, the
danger can be made more concrete. One obvious conversational rule,
implicit in the idea of scaling up, would require the conversationalists to
participate on terms of equality: all are to have a voice, and all to be taken
as authoritative with respect to their own wants and needs.10 The concep-
tion of the good drawn from the last chapter, already used to shape the
proposed method, also makes equality central. Hence, it might be al-
leged, equality is built into the method of justification of a proposal to aim
at equal treatment for all, and this is simply to beg the question.
Articulating the charge of circularity in this way enables us to appre-
ciate both its force and its limitations. The emphases on equality in
the conception of the good and in the putative method of justification are
different: the conception of the good aims at equality of opportunity for
a worthwhile life; the conversational rule requires equality of participa-
tion in the ideal simulated conversation. The difference in the types
of equality is important, for an imaginary conversation undertaken on
the condition of equal participation is not guaranteed to issue in the en-
dorsement of equal opportunities for worthwhile lives. Substantive work
is required to show how the rules governing that conversation would
generate consensus on this conception of the good. If a demonstration of
this kind were at hand, it would reveal a particular sort of coherence in
the two proposals offered: the proposal that ethical discussions justify
their conclusions if they rehearse the considerations that would bring
ideal conversationalists to consensus, where part of the ideal consists in
the equal participation of all, would be seen as providing support for the
proposal that the good requires equal opportunity for a worthwhile life,
10. As we shall see in the next section, this is too crude as it stands. But the crude ver-
sion will help to bring out the potential problem of circularity.
through exposing the ways in which discussants under the conditions of

mutual engagement would endorse the pertinent conception of the good.
Because that sort of coherence is not guaranteed, exhibiting it reveals
something important about the package of proposals. It achieves some-
thing analogous to reflective equilibrium.
Once this point is recognized, we can glimpse a possible response to
the concerns that surfaced at the end of the last section. Suppose my pair
of proposals (one concerning the good, one concerning ethical method)
is coherent in the suggested fashion. That might be enough to distin-
guish it from alleged alternatives, recommendations about the good and
about ethical method offered by putative challengers. There would be a
basis for turning back the challenge and completing the line of argument
begun in §§42–43.
Whether this possible response succeeds depends on articulating the
account of method. Turn, then, to the concept of mutual engagement.
§53. Mutual Engagement

The search for a method for ethical discussion stems from the wish to
make our modifications of ethical practice more reliable than they have
typically been. Any defense of the possibility of ethical progress should
concede that there may not be very much of it and that what there is may
be achieved blindly (§§28, 30). Pragmatic naturalists hope an explicit
account of ethical practice might generate a higher frequency of progres-
sive transitions, through the articulation of criteria for the discussion
and resolution of ethical questions.11 In the case of normal change, it is
relatively clear how to proceed, and my attempt to refine the method of
reflective equilibrium (§51) offers a putative basis. The thought of a con-
versation under conditions of mutual engagement has been introduced
in an attempt to do something similar for revolutionary change.
The search for method can be guided here by comparison with the
sciences. According to a popular story, scientific research became more
11. This is a major theme in Dewey’s writings; see, for example, chaps. 9 and 10 of The
Quest for Certainty John Dewey: The Later Works, vol. 4 (Carbondale, IL: Southern Illi-
nois University Press, 1984).
successful as it was more self-consciously directed by ideas about

method, this joint process occurring in the seventeenth century.12
It is worth reflecting on this supposed analogy, because attention to
the scientific case can temper expectations about method in ethics.
Despite the frequency with which pioneers of early modern science
linked their hypotheses and discoveries to claims about method, one
does not have to read very far in their writings to understand that,
fi rst, their methodological conceptions, while related, have important
differences, and, second, that their methodological counsel is often
imprecise. If we now have a more detailed account of methods of sci-
entific justification, that is because the initial vague thoughts about
method have inspired scientific research whose successes could then be
used to refi ne and revise the preliminary ideas about method. To take
the comparison between ethics and science seriously should accus-
tom us to the possibility that an initially imprecise account of method
might spark ethical deliberations, whose results lead to further preci-
sion about method. There are no fi nal defi nitions, but an evolutionary
process.
According to my account of method in revolutionary ethical change, a
discussion of an ethical problem (generated by functional conflict) is as-
sessed against a standard of replicating the course of an ideal delibera-
tion under conditions of mutual engagement in which all members of
the human population participate. The first condition on replication re-
quires the conclusion to be the consensus the ideal conversation would
reach:
(RC) An ethically adequate discussion concludes that p only if an

ideal deliberation under conditions of mutual engagement of
the issue whether p would reach a consensus on p.13
The point of ethical discussion is not, however, simply to state a con-

clusion, but to show others why they should adopt it. Hence there is a
12. Dewey emphasizes the comparison. See, for example, Quest for Certainty.
13. For the moment I focus on what may appear to be the implausible thought of perfect
consensus. Later, this requirement will be relaxed.
second condition on replication, directed at bringing to the fore the con-

siderations moving the ideal conversationalists:
(RJ) An ethically adequate discussion discloses those features of an

ideal deliberation under conditions of mutual engagement that
would prompt each participant to reach consensus.
I postpone issues about how one might apply any such standard, through
testing (confirming, refuting) claims that replication of an ideal delibera-
tion has been achieved.
Now to the concept of mutual engagement that lies at the heart of
these standards. The conditions of mutual engagement are partly cogni-
tive and partly affective (and in this they resemble the dimensions of
psychological altruism, §5). The first cognitive condition eliminates any
erroneous factual beliefs from the ideal conversation:
(KE) In their deliberations, the participants do not rely on any false

beliefs about the natural world.
This condition seems innocuous and well motivated, for it appears that
consensus achieved on the basis of error would be problematic—a group
of intensely altruistic people would arrive at quite peculiar conclusions if
they assumed that extremely severe pain has all sorts of wonderful con-
sequences for sufferers. (KE) entails that the aspirations and wishes the
participants form, and the premises from which they reason, must be
thoroughly secular. They cannot announce that certain actions are re-
quired or forbidden, or certain elements characteristic of the highest
good, on the grounds that there is a transcendent being who commands
us in the relevant ways or who offers us an infinitely valuable immortal-
ity. Positive beliefs in transcendent beings, and in the arrangements such
beings make, are errors that cannot be permitted to distort the ideal
conversation.14
14. Unlike some contractarian writers (Rawls, Scanlon), I exclude appeals to literalist
readings of religious texts not because they introduce reasons not all participants share,
but because they are simply false. (KE) embodies the idea that false beliefs will distort the
Two further cognitive conditions run parallel to the cognitive dimen-

sions of altruism (§5). First, in an ideal conversation, the participants
know the consequences for all of the types of actions they consider re-
quiring, allowing, or forbidding. Hence:
(KC) Each conversationalist knows the consequences for each other

conversationalist of the actions and institutional arrangements
under discussion.
Moreover, the empathetic understanding of these deliberators is not

faulty; they recognize the wishes of those with whom they converse. As
we shall see shortly, those wishes are themselves modified through the
interaction of the participants, and ideal conversationalists are able to
keep track of these modifications, much in the way that people do under
successful conditions of higher-order altruism (§21).
(KW) Each participant has complete knowledge of the

wishes of others, and of the ways in which these wishes are
modified through the course of their interactions with one
another.
Plainly, these cognitive conditions are extremely strong, and unlikely to

be realized in human exchanges; because ethical discussions meeting
the replication conditions must be sensitive to the desires of all mem-
bers of the human population (and of the ways in which those desires
would be altered through attention to others), the authors of such dis-
cussions must apparently be attuned to an impossibly broad range of
psychological facts. Once again, I emphasize that I am postponing the
issue of how the standards of ethical discussion might be applied to
particu lar cases, and how the credentials of such discussions might be
tested.
conversation, an idea easily supported by considering a vast number of examples of delib-

eration based on error ( just one of which—the mistake about the consequences of pain—is
invoked in the text). Falsehood is to be avoided, and there is no reason to think religious
falsehoods should be treated differently from others.
The heart of my account of mutual engagement consists in the affec-

tive conditions. Start from the thought that genuine engagement with
others begins from an expansion of one’s sympathies, in which the per-
ceived desires of those with whom one deliberates are given equal weight
with one’s own. Because of conflicts, that cannot be carried out consis-
tently across the board. As repeatedly noted, if two people have in-
compatible desires, it is impossible for a third party to behave as a
golden-rule altruist (§5) with respect to both of them. Consequently, ex-
panded sympathy cannot simply be understood in terms of responses to
the desires of others that give equal weight to the wishes of each. Nor
will it do to seek identification with the wishes of some harmonious ma-
jority, for the majority may be blinded by failures of sympathy. How,
then, is expansion of sympathy to be conceived?
I tackle this question by introducing the notion of mirroring others.15
The simplest sort of mirroring is that just considered (and found prob-
lematic as a general account of mutual engagement). For A to engage in
primitive mirroring of B is for A’s desire to give equal weight to the soli-
tary desire of B, and to A’s own solitary desire.16 Now, in an ideal conver-
sation aimed at addressing functional conflict in an ethical code, the
solitary desires from which mirroring begins are not those individuals
actually adopt, but rather those in harmony with the functions to which
the ethical code responds and sustainable if the cognitive conditions on
ideal conversationalists were satisfied. Hence:
15. Here I draw on ideas present in the eighteenth- century sentimentalist tradition,
notably in Smith’s theory of moral sentiments. I have explored the mirroring metaphor,
as he develops it, in “The Hall of Mirrors,” Proceedings and Addresses of the American
Philosophical Association 79, no. 2 (2005): 67– 84. As with the original discussion
of psychological altruism in §§3– 5, my formulations of conditions here incline to a
position that attributes conscious recognition of the states of others. It is possible, how-
ever, that human beings sometimes engage with the feelings of those around them in
more automatic ways (perhaps through the activity of mirror neurons). The account of
mutual engagement is easily amended to permit any mechanisms of this sort to play a
role.
16. The notion of a solitary desire was introduced in §3. Notice that I am not distin-
guishing A’s perception of B’s solitary desire from B’s solitary desire itself, since, under the
cognitive conditions of the ideal conversation, A has an accurate understanding of B’s
desires.
(DS) The solitary desires of an ideal participant include only such

desires as are compatible with precepts of the participant’s
ethical code (precepts contributing to the functions the code is
to discharge) and also retained when the participant accords
with (KE) and (KC).17
Effectively, this filters the original solitary desires, eliminating those

ruled out by the conversationalist’s current ethical code and those pres-
ent solely because of some form of ignorance.
Primitive mirroring cannot provide a general account of the affective
part of mutual engagement, because A can encounter situations in which
two others, B and B*, have incompatible desires, so A cannot accommo-
date both of them. Given the filtering required by (DS) however, some of
these problems can disappear, and primitive mirroring of others may be-
come possible. When that occurs, an ideal participant adopts the pertinent
desire (one that gives equal weight to the solitary desires resulting from the
filtering). This requires the filtered solitary desires of B and B* to be the
same, so filtering has already done some of the work of primitive mirroring.
The challenge is to understand how to engage with others when dif-
ferences remain. This is undertaken through extended mirroring. In
extended mirroring, A attends not only to B’s solitary desires, but also to
B*’s assessment of B’s desires, B*’s assessment of A’s assessment of B’s
desires, B**’s assessment of B’s desires, B**’s assessment of B*’s assess-
ment of B’s desires, and so forth. Through consideration of a variety of
perspectives, a conversationalist seeks the best balance among the ethi-
cally permissible and factually well-grounded desires present in the
population.18 So we have:
(DM) An ideal conversationalist forms a desire by extended mirror-

ing of the desires of others, achieving that desire he or she
17. Since we are concerned here only with solitary desires, the relevant constraint from
(KC) is that the participant would continue to desire what he or she actually does, if the
participant knew all the consequences for him- or herself.
18. This is effectively to construct an analog of the “ideal spectator,” but one informed
by extensive factual knowledge. It articulates further the synthesis of the methodological
ideas of Smith and Dewey I propose in “The Hall of Mirrors.”
judges to be the best balance among the varying assessments

(indefinitely iterated) made by fellow participants.
If there is complete agreement about how the balancing is to be done,

there is no need for further conversation. If there is not, the ideal con-
versation consists in attempts to support or reject various ways of
balancing.
(IC) Ideal conversation consists in attempts to show that proposals

that participants desire to implement as ways of responding to
functional conflict either accord or fail to accord either with
ethical functions all participants recognize or with their shared
understanding of the need to respond to the wishes of all.
For the moment, focus on cases in which ideal conversation achieves

consensus.19
Ethical discussions addressing functional conflict can be genuinely
helpful if they can show how the actual considerations adduced in op-
position to a potential way of solving the conflict fall foul of one or more
of the conditions on ideal conversation. That might be achieved by ex-
posing the fact that particular kinds of solitary desires are at odds with
some of the functions endorsed in the current state of ethical practice
(one identifies these functions, showing how existing precepts discharg-
ing those functions would prohibit the action or state of affairs desired);
or it might be done by showing that the desire is undermined by facts
about the world (people who express a desire for a particular outcome
do not recognize that it has consequences they strongly detest; people
have the desire because of some background false belief); or it might be
done by showing that the desire persists because of a failure to take into
account the wishes of some group of people—most obviously by show-
ing how these people are systematically ignored by those who have the
desire, but also by questioning the ways in which the balance among the
variety of human desires has been struck). An obvious skeptical com-
19. This apparently optimistic thought about possible consensus will be modified
in §55.
plaint about revolutionary ethical change would insist that there is noth-
ing for ethical discussion to do when functional conflict arises, that
spoken or written words can be only the expression of an attitude others
are free to reject. By contrast, there is plenty to be said, much that can be
done to expose factual errors and false presuppositions, disharmonies
with background features of the prevalent ethical code, and, most im-
portant, shortcomings in accommodating the wishes of classes of other
people—failures of mutual engagement.
In the next sections, I attempt to show how this approach to ethical
method might sometimes manage to achieve consensus (and also con-
sider appropriate conclusions when consensus proves unreachable). First,
however, I acknowledge the considerable idealizations introduced and
examine how, given our limited perspectives, actual ethical discussions
might be assessed.
§54. Ethical Debate

Typically, we are moved to ethical discussion by disagreement. In recur-
rent circumstances involving different agents, the actors make incompat-
ible choices and defend what they do with judgments that contradict one
another. Some of these episodes are occasions of normal ethical change.
The prevalent ethical code needs extension to address the circumstances
in question, but there is a way to extend it that refines some underlying
function(s) without cost to others. Here the task of public ethical discus-
sion is to apply the method of §51 to demonstrate to all who share the
code that this is so.
Awareness of functional conflict typically emerges from efforts to adapt
that method to a problem not yet recognized as calling for revolutionary
change. As discussants find that those attempts do not succeed in resolv-
ing the issue, they come to suspect functional conflict.20 Their first task
is to expose, as clearly as they can, the nature of the conflict. Once that is
done, the focus will be on proposals for setting priorities among functions,
20. Here I suggest a parallel to Kuhn’s claims that normal science may break down as
repeated efforts to solve puzzles fail, so what appeared as a puzzle becomes seen as an
anomaly.
or for revising attributions of priorities currently in force. The stage is then

set for ethical debate, where the aim is to simulate an ideal population-
wide conversation.
Those who engage in this debate will rarely, if ever, even approximate
the cognitive conditions (KC) and (KW). Although they will argue that
the proposal for modification of ethical practice they favor responds to the
ideal desires of all (or achieves the best balance among those desires), their
presentations will be vulnerable to charges that they have misconceived
the consequences for particular classes of people, or that they have failed
to identify the solitary desires of some groups of people, or that they have
misunderstood the ways in which solitary desires would be transformed
under attempts at primitive mirroring, or that they have a misguided view
of how to balance in extended mirroring. Even more fundamentally, they
may be criticized for basing their claims on factually false premises or for
overlooking consequences for parts of ethical practice outside the purview
of discussion.
Examples of the last two cognitive failures are easy to find. A defense
of some modification of ethical practice invoking the commands of an
allegedly transcendent being or introducing suspect psychological enti-
ties or processes would rightly be rejected on grounds of violating (KE).
Similarly, if someone aims to resolve a conflict between different ethical
functions but fails to recognize the ways in which the proposed resolu-
tion impinges on functions beyond those explicitly considered, it is im-
portant to make these implications manifest. Doing so is a continuation
of the search for reflective equilibrium, and it is a familiar feature of
ethical debates.
The method of §53 emphasizes other ways in which proposals can be
challenged, tested, confirmed and undermined. Often a thesis about
the consequences for people will be neither supported nor refuted by
the available evidence—or the evidence we have supports only a partial
version of the thesis, one restricting attention to the consequences for
a small subgroup of the human population. Many debates about the
proper distribution of resources involve large claims about the effect of
particular socioeconomic arrangements. We are told, for example, that
if certain kinds of competition are encouraged, everybody will achieve
various things they are presumed to want, or that attempts to divide the
products of labor along roughly egalitarian lines will interfere with mo-
tivation to work, and consequently with production. These claims
about the effects of general types of social arrangement—types includ-
ing a wide variety of possible implementations—are sometimes sup-
ported by assuming psychological generalizations, sometimes but-
tressed by mentioning a handful of specific historical cases.21 The
psychology invoked is simplistic, and the sample of instances ludi-
crously small. A genuine test of rival ethical proposals, with their vari-
ant conceptions of the consequences for people who live in quite differ-
ent situations, requires either further work in relevant social sciences
or a direct experiment.
Some hypotheses about the consequences of social arrangements
could be confirmed or disconfirmed through indirect investigations.
Without attempting to create conditions under which economic incen-
tives, claimed to be crucial for motivation to work, were absent, one
might aim for a general theory of human motivation, testing it in labora-
tory experiments or in existing social situations, and then deriving
from it some conclusion about likely behavior under hypothetical egali-
tarian conditions. When that strategy is considered more thoroughly,
however, it seems unlikely to resolve the question: any combination
of laboratory and field studies would be inadequate to the full range
of possibilities for realizing an egalitarian ideal. If that is so, the only
chance of replacing our— ethically crucial—ignorance about conse-
quences is to bring about the conditions egalitarians envisage and see
what happens. Rational ethical debate may require further experiments
of living.22
21. Adam Smith and John Stuart Mill disagree about the first issue (Smith holding that
increased rates of production will translate into increased economic benefits for all, Mill
maintaining that distribution is not an immediate effect of increased productivity), but they
are united on the second. Smith, Wealth of Nations (New York: Modern Library, 2000), bk.
1, chap. 8; Mill, Works, 2 [Principles of Political Economy (Toronto: University of Toronto
Press, 1970), bk. 2, chaps. 1–2]. With respect to this latter question, both appeal to a simple
generalization about human motivation. Contemporary critics of egalitarianism are fond of
mentioning the prominent instances in which they take egalitarian ideas to have been ap-
plied, and they conclude from a few failures that all ways of implementing egalitarian
ideals are doomed.
22. Mill envisages this possibility, and it is emphasized even more strongly by Dewey.
The experimental design seems relatively obvious. Those who cham-

pion some version of the egalitarian ideal can be encouraged to develop
the social conditions they envisage and to discover empirically what it is
like to live in this way. In their efforts, they will presumably be mindful
of the difficulties previous ventures of this kind have experienced, and
perhaps heartened by the well-known difficulty of making even the sim-
plest natural scientific experiments work.23 Experiments involving hu-
man subjects are properly subjected to ethical scrutiny—that is probably
a stable part of our current ethical practice—and hence there will be
constraints on how such experiments are carried out. One important
constraint is already hinted at in my suggestion that the experiment be
carried out by those who champion the ideals: commitment to it should
be voluntary. The question of appropriate constraints will be considered
more systematically in the next section, in exploring situations where the
search for consensus breaks down.
The method of §53 also emphasizes the need to engage with the de-
sires of people who live in very different ways. Ethical proposals are
subject to the conditions (DS) and (DM), but attributions of desire easily
go awry because we misrepresent the circumstances under which other
people live and the solitary desires they form.24 We can make headway,
even though ideal identification with all human perspectives remains
unattainable, through acquiring detailed information about the lives of
others. History and ethnography are pertinent sources, often unduly
neglected in the course of ethical debate.
To succeed in replicating the conversations of people who mirror one
another, ethical discussants need the ability to mirror others who are often
quite remote from them. Knowledge of those others’ circumstances and
solitary desires is a precondition for doing that, but the bare apprehension
23. With respect to egalitarian experiments, there is a tendency to overlook some of the
small societies that achieved partial success (Robert Owen’s Lanark colony), or that were
set at a disadvantage because of relations to other communities (the Diggers and Levellers
in Britain, the Israeli kibbutzim).
24. Hume is rightly criticized for the parochialism of his remarks that we can reconstruct
the wishes of temporally and spatially distant people by considering our acquaintances—
but it is far from obvious that anyone (including Hume’s critics) really avoids partiality on
this issue.
of particular wants will almost certainly be insufficient. Successful mir-

roring depends on a capacity for thinking and feeling oneself into an-
other person’s life, a form of vivid knowledge through which the imagi-
nation is exercised.25 Good ethnographies and skilled histories can
guide us toward that—but they are often usefully supplemented, or even
replaced, by imaginative fictions that focus particular aspects of human
existence. Great works of fiction and drama aid our ethical debates in
two quite distinct ways, first by developing capacities for imaginative
identification and for vivid knowledge, and second by distilling aspects
of ourselves that enable us to proceed from a bare factual knowledge of
another’s circumstances to a more vivid awareness of how that person’s
life is and feels. To cite a simple example: it was surely no news to Victo-
rian readers that impoverished slum dwellers wanted food and shelter,
but in his depictions of the squalid conditions of urban life, Dickens
made possible a far more vivid apprehension of those wants and of their
urgency.26
Vivid knowledge is also pertinent to the final (and most intricate) way
in which ethical proposals can be tested and confirmed. Extended mir-
roring involves an attempt to balance the varying assessments people
make of one another’s desires and aspirations, and (IC) allows for the
possibility of achieving this balance in distinct ways. In the envisaged
ideal conversation, alternatives are compared, considerations are ad-
vanced in support of particular ways of proceeding, and those consider-
ations are debated. That debate focuses on the various ways in which a
coherent synthesis of divergent points of view might be achieved, attending
to abstract features of the envisaged ways of balancing. In the end, how-
ever, the relevance of the abstract features lies in the impact on individu-
als with their different perspectives. If the wishes of some are overridden,
one can seek vivid identification with the people in question, posing the
question of whether, from that particular point of view, the approved way
of balancing (and the abstract features it embodies) can be understood
25. Here I am indebted to discussions with Moira Gatens and to her illuminating work
on imaginative knowledge in philosophy and literature.
26. Perhaps the most striking instance is the description of the life of Jo, the crossing
sweeper in Bleak House. But there are many others.
and accepted. Suppose the proposal supported by a particular way of

balancing is at odds with your wishes, even when those wishes have
been refined as the method requires. As you consider the desires of oth-
ers, you cannot find any obvious flaw—there has been no apparent ne-
glect of others’ points of view, you can see their eventual decision as one
involving something recognizable as balancing. If so, you should be able
to acknowledge the outcome as one you can live with. Convincing ethi-
cal discussion, when there are competing modes of balancing, must pro-
ceed through vivid identification with those whose wishes would be
overridden, revealing that the outcome has been thoroughly considered
from their perspective and found to be acceptable.
So abstract a description is vulnerable to the charge that differences
in ways of carrying out the process of extended mirroring (different ways
of realizing [DM]) are ultimately irreconcilable. Disputes that cannot be
resolved at earlier stages—through correction of factual mistakes, through
inclusion of all points of view, through vivid identification in mirroring—
prove irresoluble. That conclusion is premature—but so equally would
be the judgment that more can always be said, that considerations about
different ways of balancing can be adduced, that further exercises in iden-
tification (particularly with those whose desires are overridden or less
fully represented) can settle the debate. The method would be complicated
to apply to any serious instance of revolutionary debate, and until one
has tried to work through the complexities, it is impossible to decide
whether there is hope for resolution or not.
Once again, the analogy with scientific cases may help. Skeptics worry
that the large changes labeled “scientific revolutions” prove rationally
irresoluble. On my own account of how reasoning works in these scien-
tific episodes, the debates proceed by efforts to show how one position is
able to overcome problems, while its rival encounters recurrent difficul-
ties in tackling the problems confronting it.27 Because both parties can
point to successes, and because their problem solving is inevitably in-
complete, it is always possible, apparently, to contend that “we” have
addressed the crucial issues, whereas “they” have not. Such judgments
27. For details, see chap. 7 of my Advancement of Science (New York: Oxford University
Press, 1993).
about crucial issues, problems that have highest priority, involve the
same sort of “balancing” as that found in the schematic account of
extended mirroring. In the scientific cases, we are often confident that
some proposals for “balancing,” some claims about the truly crucial prob-
lems, are defensible and others are not. Our contemporary views about
the evolution of life do not resolve every issue in the domain, and yet
considerations of the “balance” of successes and failures demonstrate
how vastly more successful these views are, in comparison with the posi-
tions invoked as challenges to them; it is no less reasonable to judge that
Darwinism solves a huge number of significant problems that its rivals
do not, than it is to prefer a clean, well-maintained car to a rusty clunker
whose only virtue is its fetching hood ornament. “Balance” resists pre-
cise definition in the scientific case, and that should caution us against
demanding more in the ethical sphere.28
I have tried to counter the impression that any search for method to
deal with revolutionary change is hopeless. We cannot tell how far we
can get until we make a serious attempt at a complicated and taxing ven-
ture (one requiring considerable pooling of ideas and recognition of
alternative perspectives). Yet it is worth considering, in advance, not only
the cases in which consensus can be reached, but also those in which it
cannot.
§55. Dissent and the Limits of Tolerance

In characterizing ethical truth, §38 made room for the possibility of plu-
ralism. Different ethical traditions, moved by the felt urgency of discharg-
ing different ethical functions, may be drawn to values whose complete
reconciliation proves impossible. The members of the traditions, aware
of the character of the rival practice, feel themselves drawn to both values.
They would like to find ways of satisfying better that function (or those
functions) to which, as things stand, they assign lesser priority. Their
28. The earliest efforts to specify methods of scientific confirmation were schematic and
imprecise. More specific formulations emerged from attempts to put those preliminary
conceptions to work. Thoughts about method coevolve with discoveries about the natural
world. I envisage a similar process in the ethical sphere: attempts to elaborate the method
will yield more precise versions of it.
situation may thus be one of functional conflict, and, for both traditions,
the confl ict may be resoluble in alternative ways. Because there is no
ultimate resolution here, because we can envisage a progressive sequence
of changes, proceeding indefi nitely without convergence, it would be
wrong to require a method for revolutionary change always to achieve con-
sensus. Even an ideal conversation should sometimes leave participants
with alternatives.
Furthermore, even along the route to full consensus, even with re-
spect to ethical debates allowing eventual resolution, there are likely to
be intermediate stages at which agreement cannot be forced. Conse-
quences hard to predict might be recognized through direct ethical ex-
perimentation. To carry out such experiments is, in effect, to fragment
a previously united community (although the fragmentation is initially
conceived as an intermediate step toward resolution of a debated ques-
tion). As such experiments are envisaged, how they are to be conducted
will itself be a matter for ethical reflection, and the considerations in-
volved here are the same as those figuring in appraisal of the relations
among rival ethical traditions. For in both instances, questions arise
about the extent to which consensus must be sought or difference toler-
ated. These questions may be focused by asking when a single commu-
nity debating a question of revolutionary change can properly allow
division into two subgroups that adopt alternative answers.
Imagine a discussion, under the conditions of mutual engagement,
that has given rise to two rival parties whose differences cannot cur-
rently be resolved. Perhaps this comes about because there are facts
about the consequences of the alternative proposals espoused by these
parties, facts not determined by the available body of knowledge, and
because the only way members of the society can envisage settling these
facts is the route of direct experimentation. Or perhaps it is the product
of different ways of extended mirroring, the result of alternative concep-
tions of how to balance the desires of group members (desires conform-
ing to the cognitive and affective conditions of §53). Each party regards
the other as espousing an incorrect (or an inferior) modification of their
shared current ethical practice. They recognize, however, that no com-
pelling ways of showing the mistake (the inferiority) are at hand, nothing
can be done to show the factual error (fi rst case), nothing can be said
to bring home the inadequacies of the approach to balancing (second

case).
If the debate has reached stalemate in this fashion, then each of the
parties must acknowledge its rival to be in a state of conditional mutual
engagement. Given the perspective adopted by the rival party, it has
completely attained mutual engagement with respect to its members.
Consider the dispute over empirical facts. Members of each party be-
lieve matters are not as their rivals believe (or, perhaps, hope) and, in
consequence, the attempt at mutual engagement by the rival party has
gone awry (it violates [KE]). But they must also acknowledge that, if the
rivals were correct in the disputed matter, they would have succeeded in
mutual engagement within their party. Similarly, with respect to the dif-
ference in ways of extended mirroring, each group thinks the other has
done the balancing in an inferior way, but it must also acknowledge that
if their extended mirroring were adequate, they would have achieved
mutual engagement within the party.
When debates reach stalemate, rivals see one another as attaining
conditional mutual engagement. Why must this be? The answer is simple.
If it were not, the debate would not have reached stalemate. Something
more could be said: to wit, that independently of the disputed feature of
the situation (the disagreement about facts, the alternative approaches to
extended mirroring), the rival party fails to achieve an internal condition
of mutual engagement. So if the method of §53 is applied to the debate,
the point at which differences become irresoluble has to be one at which
the contending parties have attained, subject to their conception of the
point in dispute, an internal equilibrium that is as good a candidate as
any for mutual engagement. In light of this, I propose, each should be
prepared to tolerate the other.
This may appear to be a strong, even dangerous, suggestion. There
are familiar worries about permitting societies to follow their own cus-
toms, and to decide how their members shall be treated. We know that
actual societies, inspired by allegedly authoritative religious texts, be-
lieve they are entitled to take aggressive actions against other groups and
to treat some of their members in ways outsiders view as harsh, restrictive,
and cruel. We know also that the recipients of this treatment sometimes
endorse what is done to them, genuinely desiring to live in the way assigned
them. Is the method then committed to permitting local groups to flout

what liberals typically hail as universal values?
No. The conditions of mutual engagement are sufficiently strong to
limit tolerance of alternatives so as to exclude these problematic cases. As
§§38 and 46 argued, there are imprecise prescriptions, such as those in
favor of honesty and opposed to initiating violence, likely to be stable in
progressive ethical traditions. These core truths cannot be abandoned
if the proposed method is adopted—for dishonesty and unprovoked ag-
gression are, in the overwhelming majority of instances, expressions of
failure to engage with the desires of others.29 This means the core truths
will have to be shared among the tolerable rivals, and groups rejecting
them will be at odds with the conditions of mutual engagement.
Even more evidently, (KE) rules out appeal to religious texts as the
foundation for any sort of treatment of group members, and any set of
attitudes to outsiders. If it is proposed that women should be confined in
various ways or that men should proselytize aggressively, and the pro-
posal is grounded in the supposed commands of a deity, a fundamental
cognitive condition on mutual engagement is violated. If the ideal con-
versation is to defend proposals of these kinds, it will have to do so in
different ways. It is hard to conceive how any policy of proselytizing
could be detached from a foundation in a false belief about transcendent
entities, and aggressive attempts to spread the word would be debarred
by the core truth opposing unprovoked violence. The case of female
confinement, however, and analogous proposals with respect to other sub-
groups, can seem more difficult. For, it may be suggested, the individuals
who receive the treatment, unpleasant as it may appear to outsiders, really
want to live as they do. Suppose their wish is independent of any faulty
religious belief. Is the societal practice then to be tolerated?
Typically not. Although the current expression of the desire is
sincere—the woman declares that this is the way she wants her life to be,
this is her choice—it is usually infected by a false belief, the belief that
29. The treatment of honesty in §46 argued that there are exceptional instances and that
the core truth is thus vague. One can envisage rival ethical traditions proposing different
more precise versions of it, but abandonment of it would be ruled out by the method sug-
gested here.
her socialization and education have elicited her wants. If she were to
see clearly how her options have been narrowed from the beginning, if
others who care about her were to understand how there were alternative
possibilities for her, the desires held by her and by them would be quite
different. The actual desires owe their presence to a pervasive error, the
mistaken belief that the existing processes of socialization are non-
coercive. Consequently, there is a failure of mutual engagement: in effect,
those who have led her to her present state have engaged not with her but
with their preconceived image of her. Only in the case of people who,
under alternative schemes of socialization, would remain entirely un-
concerned by thinking their choices had been narrowly foreclosed from
the beginning can one suppose that the desire is uninfected by error.
There may be such people, but we have every reason to think that they
are rare.
Discussion of such cases exposes an important facet of the limits of
tolerance, one best illustrated by considering the permissibility of ethi-
cal experiments. The last section envisaged groups, moved by an egali-
tarian ideal, trying out their preferred versions of it. If those who sign
on to the experiment are coerced, either through the overriding of their
present desires or through prior socialization that has narrowed their
choices, the experiment will not accord with the conditions of mutual
engagement. At times, however, people voluntarily resolve—by our every-
day standards of voluntary choice—to undertake this kind of venture.
There is a danger they will coerce others, in particular their children. The
forms of coercion making the examples of confining social practices sub-
ject to ethical critique may be absent in the first generation, but they can
easily emerge thereafter. Do all ethical experiments under mutual engage-
ment have to be short-term?
To avoid that conclusion the education and socialization of the young
must carefully preserve options. The experiment should allow children
to choose from as wide a range of alternatives as those available to their
peers, whose parents and teachers are not part of the experiment. The
experimenters must recognize and tolerate their children’s opting out—
and, perhaps, others’ opting in. As a result, some possibilities for living
(forms of family relationships) may be distorted, a distortion to be recog-
nized if the probative value of the experiment is to be properly assessed.
In virtue of their status as minority ventures, ethical experiments are al-

ways vulnerable to breakdown. Judgments of failure should take into
account the stresses to which they are subject.
To sum up: when conversations under the conditions of mutual en-
gagement break down, competing parties should see one another as in
an internal state of conditional mutual engagement. Groups achieving
this state can be allowed to pursue their favored ethical practices. Although
the suggestion initially seems to allow for overindulgence of alternative
ways of living, including those oppressing members of some subgroups,
the conditions of mutual engagement are sufficiently strong to debar the
cases that provoke condemnation. The method provides a basis for ethi-
cal criticism of groups engaging in externally aggressive or internally
coercive practices. 30
§56. The Challenger Revisited

Section 51 closed with a challenge. Functional conflict seemed to make
room for ways of rejecting some major functions of ethics in favor of oth-
ers, and thus for radical differences in continuing the ethical project. To
give the challenge clear force, we need only envisage someone who pro-
poses to go in a dramatically different direction from the one pursued in
this chapter and its predecessor, someone who has little patience with
the original function of ethics. Even if he concedes that the ethical proj-
ect began with the need to remedy altruism failures, 31 he sees no reason
to be governed by that beginning. The subsequent course of ethical
evolution proliferated possibilities for human life, and, aware of these
possibilities, the challenger emphasizes the freedom of the strong indi-
vidual, who charts his own way quite independently of the protests (or
30. Whether it will also justify outside pressure to amend the practices, or even the use
of force, is a further question, one that would have to be settled by looking carefully at the
details in individual instances. To justify ethical criticism is not yet to warrant particu lar
ways of translating the criticism into action.
31. That is, he concedes the story as I have narrated it. The challenger, shortly to be
identified as a version of Nietzsche, might prefer an alternative account, one viewing the
emphasis on cooperation as a late corruption [see the first essay of Friedrich Nietzsche On
the Genealogy of Morality (Cambridge: Cambridge University Press, 2007)].
the applause) of others. This challenger can be viewed as a version of

Nietzsche, and he sees the principal function of ethics to lie in opening
the way for the emergence of higher types of beings, who will carry some
of the characteristics of the greatest individuals of the human past to
new heights.
What can be said to someone who poses this challenge—and poses it
seriously? The method developed in previous sections offers an obvious
line of response: the rival possibility spectacularly fails to meet the con-
ditions of mutual engagement.32 Rejection of the original function of
ethical practice goes hand in hand with a repudiation of the method, so
the circularity noted in §52 appears to doom the response. Although one
could bring the challenger into line, if he agreed to argue according to
the conditions proposed, he will claim the conditions take for granted
just the emphasis on the original function of ethics he is concerned to
reject. In this point he will be entirely correct.
My earlier attempts to confront troublesome characters in §42 gener-
ated an understanding of the criteria for success. Turning back chal-
lenges requires less than absolute silencing. What must be done is to
settle genuine doubts and worries at least as well as alternative ethical
pictures are able to do. In the present case the comparative approach
yields little solace. Some alternative approaches to ethics, Kantianism
and other forms of rationalism, seem to have weapons with which to
fight the radical Nietzschean challenge. Proponents of such approaches
will claim to have a priori methods of establishing fundamental ethical
principles, and thus a priori knowledge of the falsity of the challenger’s
claims. Pragmatic naturalism denies both the supposed a priori meth-
ods (or at least the a priori status of the methods) and the “fundamental
ethical principles” known on their basis. Hence it may be seen as prepar-
ing the way for the challenge—precisely because of the emphasis on the
historical evolution of ethical practice, on the emergence of different
ethical functions, and on the lack of any a priori insight into ethical
truth, it appears to have jettisoned the weapons needed to fight back.
32. The challenger may even flaunt this rejection, comparing the masses whose desires
are ignored or crushed to lambs that serve as prey for stronger, “nobler” beasts. See the fa-
mous image from Genealogy of Morality, essay 1, sec. 13.
Advertising the resources of rationalism has more show than sub-

stance. True, rationalists can produce some sentences after the challenger
has had his say, but there is no great force in invoking mysterious pro-
cesses announced as yielding knowledge of an especially robust type.
Swords can be used to threaten people, but the shadow of a sword should
frighten no one. Although rationalists may pride themselves on their
ability to keep talking in a situation where naturalists stand speechless,
the difference is not particularly significant if the phrases they utter are
empty. Nevertheless, I do not want to leave matters here: as I shall now
try to show, pragmatic naturalism has considerably more to say.
Two points from earlier discussions need elaboration. First, even
though the use of a conception of the good to frame an ethical method
threatens to be circular, as we discovered at the end of §52, the method
is not guaranteed to generate the preferred conception of the good. If it
could be shown to do so, the package of proposals would turn out to
be coherent in a way they might not have been. Second, the original
reply to the troublesome characters of §42 demanded the challenger
provide some alternative. The current challenge has begun to do that,
in proposing to dismiss the initial function of ethics in favor of the sub-
sequent function of individual development, but that is only the begin-
ning. This idea needs to be articulated as a full account of the good,
without presupposing ethical functions supposedly dismissed or down-
graded. It must be supplemented with ways (if not a method) of further
ethical development, and the conception of the good must cohere with
the preferred modes of ethical evolution. If pragmatic naturalism can
be shown to be coherent, and if the rival proves to be difficult to articu-
late, tacitly dependent on ethical ideas it officially rejects, and not easily
combined with any strategy for ethical change, enough will have been
said.
Start with the coherence of pragmatic naturalism. Under conditions
of mutual engagement, deliberators attempt to respond to one another’s
desires. Which desires strike them as particularly crucial? They recognize
that we all have common basic needs, and that, as things stand, these are
satisfied for some people but by no means for all. They know, too, that
the more fortunate people have a richer set of aspirations, that from their
own perspectives, some desires are much more crucial than others, and
that the relative centrality of these desires is the result of a conception of

what matters in their lives. In light of our shared humanity, they suppose
people who struggle to meet their basic needs would also develop richer
aspirations of a similar type if they were provided with the chance to do
so. From this, they conclude that the appropriate focus of mutual en-
gagement would be the satisfaction of people’s central desires, and thus
the promotion of worthwhile lives.
This line of reasoning assumes that we try to reach mutual engagement
with one another by attending to what each person takes to be most im-
portant, and we know enough about human possibilities to move beyond
the atomism of hedonism, with its decomposition of lives into sequences
of momentary states of pleasure and pain, and to appreciate the holism
of the desires people form, once they are freed from painful struggles.
People who currently are absorbed by the need to survive can appreciate
the value of further aspirations, in conditions where their basic require-
ments are met. Moreover, people who come to see their options as nar-
rowed through social coercion would prefer to form their own assess-
ment of what matters in their lives. If these hypotheses are correct, the
deliberators will want primarily (i) the option of forming for themselves
a conception of what matters, and of doing so in an uncoerced way; (ii)
the satisfaction of the material preconditions of this option; and (iii) the
satisfaction of the desires endorsed as central. Mutual engagement takes
all three types of desires to matter for all. Does it also require conceiving
the good in terms of providing i and ii for all, and offering roughly equal
chances with respect to iii?
Apparently. On what basis could the deliberators introduce inequali-
ties here? One possibility would be to maintain that serious chances of
satisfying desires of type iii for some people require the failure to satisfy
desires of types i and ii for others: one simply cannot give all people seri-
ous chances for a worthwhile life, and it is better that some should have
such chances, even if it means others live in want. But the conception of
the good advanced in Chapter 8 blocked this proposal by taking the
question of population size as fundamental (§49). The human population
is to be maintained at or below the proper bound, and this means resources
are available to satisfy i and ii for all, and to give everybody serious chances
with respect to iii. If the deliberators are to arrive at an alternative, they
will have to suppose it is better not to impose this condition on popula-

tion size. That will commit them to holding as good a world in which
some people (possibly most people) are doomed not to have the most
important desires satisfied. How could this be anything other than a
failure of mutual engagement? 33
Another possibility would be to suppose that a conception of the
good can tolerate significant differences in the probabilities of people’s
satisfying type iii desires. The need for universal satisfaction of i and ii
is conceded, but the opportunities for some can be significantly richer
than those available to others. Now it is agreed that resources are suf-
ficient to provide serious chances of a worthwhile life for all. Hence
there is a distribution of those resources allowing everybody roughly
equal, serious chances of satisfying desires of type iii, where the standard
of rough equality is set by the recognition of the perturbations uncon-
trollable factors can introduce (§50). Lowering the chances for some,
while raising them for others, would apparently be a failure of mutual
engagement.
Perhaps the proponent of inequalities could appeal to an alternative
strategy for extended mirroring. From the standpoint of the egalitarian
distribution, the inegalitarian proposal divides the population into two
classes: the Augmented, whose chances are increased, and the Dimin-
ished, whose chances are lessened. Introducing differences plainly gives
less weight to the wishes of the Diminished. Yet it might be viewed as
responding to the desires of the Augmented and as supposing that the
balancing of conflicting desires is conducted by giving weight to the re-
action of the Augmented to the desires of the Diminished. That will
happen only if the Augmented judge that the wishes of the Diminished
are not to figure in forming their own desires. That could happen if they
simply ignored the wishes of the Diminished—a failure of mutual en-
gagement. Or it could happen if they used a strategy for extended mir-
roring that relied on the judgment of others. But these others would also
33. One way in which my conclusion here might be resisted would be to argue that set-
ting bounds on the population is incompatible with major components of worthwhile lives.
This line of reasoning was already considered in §49, in confronting a form of skepticism,
and I rely on the response developed there.
have to ascribe low weight to the wishes of the Diminished, and exactly
the same problem would arise. Thus, the only way for extended mirror-
ing to generate the inegalitarian desires is for a failure of primitive mir-
roring to occur somewhere among the class of the deliberators—and
that means that the discussion cannot satisfy the conditions of mutual
engagement.
Application of the method of §53 is likely to yield pragmatic natural-
ism’s conception of the good, as the provision for all of serious, and
roughly equal, chances for worthwhile lives. Further, the conception
of human relationships as central elements in worthwhile human lives
can be viewed as in harmony with the method’s insistence on the con-
ditions of mutual engagement. Yet the conclusion should not be over-
stated. Any rehearsal of an ideal conversation must appeal to psycho-
logical assumptions about human reactions, and these can be falsified
by actual reactions to the line of reasoning people are supposed to
adopt. In advance of those reactions, a tentative conclusion: pragmatic
naturalism’s proposal about the good coheres with its view of ethical
method.
Turn now to the challenger’s position. The challenger’s conception of
the good sees the development of the few as crucial and the lives of the
many as unimportant. There are special people—call them “free spirits”—
supposed to be capable of a form of human existence superior to any
attainable by the rest. A world is good to the extent that it allows for the full
development of the free spirits; what happens to the others is irrelevant.
This conception is evidently incompatible with the conception offered
in Chapter 8, and the incompatibility descends from a difference about
the functions the ethical project is to discharge. The egalitarian concep-
tion is embedded in an attempt to take the original function as primary,
to see ethics as continuing to be directed at remedying altruism failures
(and to view ethical method as an attempt at mutual engagement). Its ri-
val views the original function of ethics as a trigger that eventually led to
enhanced possibilities of human living. Whereas the egalitarian concep-
tion attempts to integrate the emergent possibilities with the original
function (through the focus on equal opportunities for worthwhile lives),
the challenger takes the emergence of the richer possibilities to be a turn-
ing point, after which the original function can be forgotten, in favor of
ever further enhancement of these possibilities, even if the resultant

opportunities are available only to the few.
So far, the rival approach has no method for ethical discussion—
perhaps because there is to be no further ethical discussion. It is not
obvious, however, how we arrive at a point beyond the need for further
consideration of what to do. Note first that the enhanced human possi-
bilities realized by free spirits had better be considerably richer than
those available to the human population generally, for, if they were not,
we would all be approximations to the free spirits, and the challenger’s
position would collapse into the egalitarian view it is supposed to re-
place. Now the human possibilities open to ordinary people under
favorable circumstances—the kinds of things people take to be impor-
tant elements of worthwhile lives—are typically dependent on extensive
social coordination. Material resources must be provided and social
institutions must be in place for us to enjoy those possibilities. Even
without any definite understanding of the splendid ways in which free
spirits are able to live, there is no basis for thinking their predicament is
different. They will depend on a society and rely on coordinated efforts.
Questions will arise about how the requisite basis is to be achieved, how
the relations between free spirits and the herd are to be governed, how
the potentially confl icting projects of free spirits are to be reconciled.
Apparently, then, ethics cannot come to an end, and there will have to
be some way of continuing ethical discussion (even if it is supposed that
the discussion is one in which only free spirits take part).
We do not know what the possibilities are to which the free spirits
(and only they) can aspire—hence, the conception of the good is woe-
fully indefinite. We also have no ethical method (despite the need for
ethics to continue), and hence no possibility of showing that the method
and the conception of the good are mutually coherent. We cannot even
decide whether the enhanced possibilities for the free spirits presuppose
a society, whose coordinated efforts already rest on ways of discharging
the original function of ethics (the function allegedly transcended). These
are serious defects, but, because of elementary facts about human biology
and psychology, further trouble arises.
Any attempt to promote the splendid flourishing of the free spirits
would first have to identify them. Perhaps the challenger hopes they
will identify themselves, soaring above the mediocrities around them.

Yet a banal truth about us is that our potential is not marked on our
foreheads at birth. Human beings develop, and, as they do, their tal-
ents and capabilities sometimes become clear. Not always, however,
since it is a sad and familiar fact that people are often limited and even
crippled by inadequate social environments. No free spirit will be able
to soar unless he has previously had a sufficiently supportive environ-
ment. One way to allow for that (hypothetical) self-identification would
be to provide ample opportunities for all—but that would absorb the
central features of the conception of the good currently subjected to
challenge. Promoting the development of free spirits would then presup-
pose the provision of equal opportunities for worthwhile lives for all.
What other possibilities are there? Could we use the identified free
spirits in one generation to pick out the likely free spirits of the next?
Even with an understanding of the phenomena of heredity far less ar-
ticulated than that provided by contemporary genetics, it is easy to see
that this is hopeless. 34 Without a background provision of equal op-
portunities, the supposedly incompatible approach cannot even get off
the ground.
Section 42 responded to radical challenges by arguing that they de-
serted the ethical project without offering any clear alternative to it (ex-
cept for the fragile hominid social state from which that project liberated
us). Section 51 concluded with a way of reviving those challenges, in the
exploitation of functional conflict: it appeared to be possible to develop
rival ways of continuing the ethical project, ways highlighting or subor-
dinating different functions. One version of the challenge has been con-
sidered, and we have discovered (tentatively) that proposals grounded in
the original function of ethics cohere with one another, whereas elabora-
tion of the alleged alternative encounters severe difficulties. Not only is it
unclear what method of ethical discussion it can offer, or how method
and conception of the good can be coherently developed, but there is
34. As Plato already did. The thought of concentrating on biological heredity is central
to his way of providing different courses of education for differently talented people [Re-
public (trans Grube) (Indianapolis: Hackett, 1992)]. Yet Plato understands that this will at
best be an approximation: golden children can emerge from leaden parents, and golden
parents can produce baser offspring.
significant danger that the alternative will presuppose the function it is

supposed to transcend.
Consideration of a single radical challenge can easily inspire the ob-
jection that the alternative has been carefully chosen: pragmatic natural-
ism succeeds only because it is juxtaposed with an implausible rival
position. On the contrary: this version of the challenge is important
precisely because it demotes the original function of ethics and thus under-
cuts the idea of ethical decision through mutual engagement. Less
sweeping rival approaches would allow the importance of the original
function. Once remedying altruism failures and mutual engagement are
taken seriously, the range of alternatives for continuing the ethical proj-
ect is greatly narrowed. We are left with those functional conflicts that
give rise to tolerable pluralism.
Both the conception of the good presented in Chapter 8 and the
ethical methods outlined here seek ways of replicating, within the con-
temporary context, characteristics of the earliest stages of the ethical
project. They can be viewed as endeavoring to undo the distortions
introduced in later phases. Moreover, the conditions imposed on mu-
tual engagement deliberately reflect the account of psychological altru-
ism (§§3–5, 21). Two motivational corollaries follow. First, the renewal
of the ethical project should proceed by demanding of deliberators
that they exhibit, in the most complete possible form, the characteris-
tics of psychological altruists, for that is to take the original project, the
remedying of altruism failures, as primary. Second, to the extent one is
concerned that the concept of an altruism failure is sufficiently open or
ambiguous to allow different elaborations by participants in ethical
discussion, it is appropriate to demand of those discussants wide-
ranging forms of psychological altruism. The judgments about pro-
gressiveness should be vindicated by submitting them to discussion
under conditions of mutual engagement. Those conditions can be in-
voked to fi ll out the account of progress of Part II: to the extent that
there are questions about whether a proposed transition remedies
altruism failures, questions turning on what is to count as an altruism
failure, those questions are resolved by identifying the altruism failures
as those recognized by discussants who conform to the conditions of
mutual engagement.
Section 44 denied the existence of ethical expertise. The consider-

ations offered in the past two chapters are hardly knockdown arguments.
They are invitations to consider our ethical choices in a particular
way—as a large-scale emulation of those who began the ethical project—
and they are to be tested by reactions, based on a variety of human per-
spectives. That is entirely in the spirit of the method proposed, a method
emphasizing conversation, simulated or perhaps real.
chapter 10
Renewing the Project
§57. Philosophical Midwifery

How does the normative stance developed in the last two chapters, the
egalitarian conception of the good, and the method for ethical decision
that aims to simulate wide-ranging deliberation under conditions of
mutual engagement, apply to our contemporary predicament? To reca-
pitulate: it is not for any single author to answer. Philosophers can make
proposals, attempting to facilitate the conversation that would deliver
answers. Call the work of facilitation philosophical midwifery. The most
obvious forms of philosophical midwifery consist in proposing topics
for consideration (places on our common vessel where planks might de-
serve attention) and suggestions about those topics (specific ways of
rearranging the timber in those places). Almost all of this chapter will be
devoted to efforts in this vein. First, however, a more basic type of mid-
wifery is needed if the ethical project is to be reborn.
The ancients wondered if virtue could be taught. Pragmatic natural-
ism takes the catalog of virtues to be something generated from ideal
conversations. A prior issue, then, is how to produce good simulations
of those discussions. To renew the ethical project in the ways suggested,
deliberators capable of approximating the conditions of mutual engage-
ment are needed. Where will they come from?
370
Renewing the Project 371
When the ethical community becomes the entire species, all the in-
volved parties can no longer sit down together and discuss potential
modifications to a shared code. Socialization of the young can no longer
stop with transmitting the code as it has been passed on, leaving its fur-
ther evolution to conversations in which they will engage with one an-
other. At least some people in each generation, perhaps the entire cohort,
need to understand what the ideal of mutual engagement is and how
they might exemplify it. The proposals of the last two chapters call for
further proposals about ethical education.
The cognitive conditions on mutual engagement require recognition
of the predicaments of others. In many affluent societies, the general idea
of learning about people who live under very different circumstances
receives official approval, but, for serious ethical discussion, it deserves
more systematic emphasis. Young people need to acquire vivid knowl-
edge of a range of lives lived by their fellows. They should be conscious
also that their power of vivid representation reveals only a sample of the
actual forms of contemporary existence. When they reflect on proposals
for action, they should be aware of how those proposals would be felt by
people whose situations are very different from their own. If they eventu-
ally conclude that nothing can be done to improve the predicaments of
the less fortunate, they should do so with an intimate understanding of
how the neglect would be experienced. The aspirations of others should
be present in their own psychological lives.
Education should also foster an ability to satisfy the affective condi-
tions of mutual engagement. Besides a general encouragement of empa-
thy, potential deliberators need an ability to shape altruistic responses
along the lines of the inherited ethical code, to recognize endorsable and
contaminated desires, insofar as these have been discriminated. They
should have practice in mirroring, in contemplating an outcome from
various angles. They should become aware of the value of higher-order
altruism. All these skills can be developed by simulating the early phases
of the ethical project. Children can learn, from early in their lives, how to
work out a life with others, to find their way to rules all can accept—and
they can experience the joys of fluent cooperative interactions according
to those rules, as they attune their own conduct to that of their peers.
Although some issues, most notably the custody of our planet, arise at
a specieswide level, others can be negotiated by smaller groups, and these
provide practice in reinforcing capacities for ethical conversations at any

scale. Sensitivity with respect to the largest questions, where discussion
involves a small number of (with luck) representative voices, and in
which eventual action depends on voting, can be cultivated if those who
frame the options and publicly debate them, as well as those who re-
spond to the conversation, have practice in working on smaller topics,
where the coordination of social life can be carried on as it was by the
ethical pioneers. Ideally, as people grow, they should gain experience
of ethical deliberation at many different scales.
To the extent that this already occurs in contemporary societies, it
often does so within religious communities. Joint action to address the
problems felt locally emerges from conversations in the synagogue, the
mosque, or the church. Previous chapters have characterized the embed-
ding of ethics within religion as a distortion of the ethical project, and to
insist on that description seems doubly at odds with the educational pro-
posals of this section. For not only does secularism undermine the most
prominent institutions in which local ethical deliberation is found, but
the emphasis on mutual engagement appears at odds with dismissal of
aspirations and ideals figuring centrally in most human lives.
It is possible to admire the fact that ethical discussion occurs in lo-
cal religious communities, and even to approve its responsiveness to
human problems, while maintaining that the conversation should be
freed from the potentially damaging reliance on taking particu lar
texts, interpreted by leaders credited with special insights, as overrid-
ingly authoritative. Secularism should aim to replicate the dedicated
attempts to work through problems together, without succumbing to
the distortions induced by supposing par tic u lar commands to repre-
sent the divine will. Surrogate institutions are needed to do the work
once done in the “cool hour” around the campfi re.1 Nor is the empha-
1. The secular humanism I defend cannot be a purely negative position, one that stops
with the declaration that religion is bunk. For elaboration of the point, see the closing
pages of my Living with Darwin (New York: Oxford University Press, 2007), as well as
some further essays [“Beyond Disbelief” in Russell Blackford and Udo Schüklenk (eds)
Fifty Voices of Disbelief (Oxford: Blackwell, 2009), “Challenges for Secularism” in George
Levine (ed) The Joy of Secularism (Princeton, NJ: Princeton University Press, 2011), and
“Militant Modern Atheism” Journal of Applied Philosophy, 28, 2011].
sis on secular discussion a failure of mutual engagement. Insisting that

appeal to religious ideas not be part of ethical conversation, embodied
in the requirement (KE) of §53, is based on three important ideas.
First, false belief can misdirect the discussion; second, proselytizing
religions systematically fail to respond to the aspirations and desires of
large classes of people (those currently outside the faith, as well as in-
siders whose propensities diverge from those taken to suit the divine
taste); third, the reasons advanced in deliberation should be shared
among the ideal discussants, and for this a focus on the fi ltered desires
of others is appropriate. Yet it is not enough to leave matters there. Re-
ligion is central to the lives of many people, and, for them, to remove it
from the ethical forum can be felt as disenfranchisement. A mutually
engaged secularism should take seriously the psychological and social
needs religion, and religious community, satisfies, recognizing and re-
sponding to the desires and aspirations out of which religious commit-
ment grows. Especially for the world’s poor, for whom basic material
needs are not met and for whom it is difficult to think in terms of a
freely chosen structure for their lives, religion can provide both conso-
lation and framework. The conception of the good, proposed in Chap-
ter 8, is intended to recognize and respond to the predicaments of the
poor, including the desires that have made the world’s religions so at-
tractive. This egalitarian conception is a deeper embodiment of mu-
tual engagement.
The proposals sketched here recapitulate themes often found in edu-
cational theorizing during the past century. Seen in the context of the
ethical project, their systematic importance may be appreciated. Ethical
discussion, in undistorted forms, is urgently required of us, and the dis-
cussants have to be prepared. In a world dominated by crude conse-
quentialist visions of the good (A and B of §48), it is easy to marginalize
educational programs fostering mutual engagement: so much firm tech-
nical knowledge needs to be impressed on the future worker! Pragmatic
naturalism suggests a different vision, one in which learning has its place
as a component of the worthwhile life.
§58. Scarce Resources

The original problems of sharing basic resources, providing food, secu-
rity, and shelter for all, do not arise for many people, so long as they fo-
cus attention on those with whom they have most contact. Around the
world, however, these necessities are often lacking, and there are pockets
of poverty even in the most affluent societies. Pragmatic naturalism’s
conception of the good yields an imperative to correct this situation, as
immediately as possible. Making food and clean water available to all is
almost certainly not beyond contemporary human ability. Problems of
shelter are harder to solve, especially in a period when many areas of the
world are increasingly vulnerable to natural catastrophes. Most difficult
of all is the provision of security for many of the world’s poor, especially
women and children. The constant threat of violence arises from many
causes: lack of reliable supplies of basic necessities, historical tensions
among groups, and religious differences. Only when it becomes clear
that shelter is available for all, that escalating vendettas are avoidable
(and better avoided), and that satisfying lives are possible on a secular
basis, are the risks likely to decline to tolerable levels. For that to occur,
significant educational resources will have to be provided, enabling rival
groups to appreciate the tortured history of their antagonisms and sap-
ping the power of militant religions (§60). Even if the will to tackle the
problem of security were present, it would have to be sustained over a
long term.
To recognize these issues of global inequality and the need to remedy
them is hardly novel. For pragmatic naturalism, however, they are an
absolutely necessary first step toward realizing the conception of the
good. Until food, water, shelter, and security are provided, talk of “op-
portunities for a worthwhile life,” supposedly available to all, will be
idle. Should we then direct all our energy toward satisfying the basic
needs of the world’s population? If so, would the affluent world have
to make severe sacrifices, possibly even undermining our struggle to ad-
vance the good?
There is no reason to think an energetic and resolute effort to provide
basic resources for the poor would bankrupt the affluent world. The is-
sue is, however, too many-sided to be resolved here. Instead, I want to
identify an important constraint on attention to global inequality. The

fundamental task is to bring about a state of affairs in which the necessi-
ties of life can be stably maintained for all. To reach that goal, some insti-
tutions and practices in the affluent world must be maintained. Specifi-
cally, it will be crucial to support schools, colleges, universities, and
research centers, because of their role both in providing sustainable tech-
niques for agriculture, hydrology, and engineering, and in serving as the
basis for realizing further aspects of the conception of the good (specifi-
cally, enabling young people to grow into sensitive ethical deliberators
and to learn to make free choices about worthwhile life projects for
themselves). The conception of the good may well demand that affluent
people sacrifice certain luxuries, for doing so could ease the burdens of
the poor: the reorientation of biomedical research toward tackling infec-
tious diseases, in the environments in which they kill and disable mil-
lions, rather than providing diet pills and cosmetic treatments, is one
specific example.2 By contrast, educational systems are likely to serve as
sources of new technologies for solving the problems of providing basic
necessities for all, and they provide models for the extension of educa-
tion to give all young people opportunities to develop their life projects.
Is this to desert pragmatic naturalism’s conception of the good? Com-
mitted egalitarians may suppose too much has been conceded to the
notorious trickle-down defenses of capitalism: let competition thrive and
entrepreneurs enjoy enormous revenues, for out of this will come bene-
fits for those who are currently poor. My claim seems to run in parallel:
let funds be invested in high-quality schools and universities in the afflu-
ent world, for out of these may come new ideas and technologies to help
transform the conditions of the world’s indigent people. The defenses
are parallel but by no means identical, for there are two different classes
of empirical facts, one about money flows under unfettered capitalism,
and one about the genesis of helpful ideas and techniques under high-
quality education. We have plenty of evidence that the first class of facts
does not favor the trickle-down defense: the condition of the poorest citi-
zens is better in those countries in which inequalities are curbed. By
2. See James Flory and Philip Kitcher, “Global Health and the Scientific Research
Agenda,” Philosophy and Public Affairs 32 (2004): 36– 65.
contrast, even though the transfer of technologies to the impoverished

areas of the world is imperfect—largely because of the intrusions of
unregulated capitalism—it does offer actual successes, and the potential
for many more, if the orientation of research toward the needs of all were
taken more seriously (for example, by penalizing research and develop-
ment that caters to trivial luxuries at the cost of delivering necessities).3
Ultimately, pragmatic naturalism proposes that further resources, in-
cluding quality education and medical care, be distributed to all. Focus-
ing on these cases should remind us that serious problems about scarce
resources remain, felt even in the affluent world. Only the children of
lucky parents receive the education they deserve. Here, then, we find
ourselves in a situation analogous to that of the first ethicists. Three dif-
ferent desiderata need to be weighed in a decision about sharing scarce
educational resources: first, there is the question of what to do, here and
now, with these children, for whom variable educational opportunities
are available; second, there is the question of improving the local supply,
that is, of trying to ensure that the educational prospects for future
groups of children, in this segment of society, are better than those cur-
rently existing; third comes the issue of realizing the long-term concep-
tion of the good, a world in which all people are provided with the basic
necessities and all children have the educational opportunities they
deserve.
The first question requires a plan, a scheme for dividing up the edu-
cational goods, which all those affected would acquiesce in as the best
available, provided they were to deliberate with one another under the
conditions of mutual engagement. Yet to concentrate on it alone would
be inadequate, for the consensus view obtained under that deliberation
might slight the wants and aspirations of another affected group, those
who will face the same problem in coming years. They might protest
that their own choices were constrained because of the inaction of pre-
decessors who merely focused on a scheme for sharing and did not think
3. To cite just one example, techniques for genetically manipulating crops might dra-
matically increase the sustainable food supply in the parts of the world most subject to
famine, if agricultural companies were less concerned to secure an indefi nite return of
large profits by modifying plants so poor farmers could not harvest their own seeds.
about increasing supply—and suggest that increasing the supply of edu-

cational opportunities for later generations should factor into the deter-
mination of a policy for sharing educational opportunities now. By the
same token, deliberators who embodied the viewpoints of the world’s poor,
included in a broader discussion under conditions of mutual engagement,
might criticize the results of narrower conversations on grounds that they
achieved local improvements without doing anything to advance the
universalist egalitarian conception of the good. The challenge is to dis-
cover integrated solutions acceptable to the broad group of ideal delib-
erators as the best available—a package of proposals deliberators under
conditions of mutual engagement would endorse.
Consider briefly various factors that might play a role in devising a
scheme for apportioning children to schools, or young people to univer-
sities. One possibility would be to institute a lottery. Another would al-
low access to those who can pay—in the extreme form, permitting an
auction in which parents bid for positions. Yet another would propose
that all children be evaluated by administering tests and the distribution
be chosen according to what maximizes their potential for development.
Still another would look to the effects of distributions on the future of
various subcommunities, on considering, for example, whether people
of particular backgrounds can emerge to play respected social roles (or
perhaps emerge as future teachers who will raise the overall level of edu-
cation). No deliberation that fails to consider these potential factors (and
several others) can be considered ideal, but each of them brings in its
train a large set of empirical questions. We still lack firm knowledge
about how providing educational opportunities for underprivileged
groups enhances motivation for members of those groups—and what
other changes in the social environment might be needed to magnify the
effects existing under the status quo.4 In fact, much current discussion
4. An obvious biological analogy is frequently unappreciated. When opponents of

schemes offering educational opportunities to the underprivileged complain that the moti-
vational effects on young people in these groups are small, they fail to recognize that pro-
viding such opportunities may be analogous to watering a plant—without water the plant
will surely die, but even with water, it will not thrive unless other nutrients are available in
the soil. Debates about educational policies are entangled with other issues about the
handicaps and obstacles faced by the poorer members of competitive capitalist societies.
of the proper modes of sharing scarce resources, like access to the best
schools and universities, proceeds by guesswork.
That said, the ways of considering the issue adopted by educational
institutions and by policy advisers are not entirely bad. Instead of trying
to draw conclusions from abstract philosophical principles, they typi-
cally proceed by involving a number of people who represent different
points of view. Procedures of this sort can be regarded as an approxima-
tion to pragmatic naturalism’s method—even though they fall far short of
the ideal conditions of mutual engagement. There are obvious ways of
improving them: the population of those represented in the conversation
could be far more varied and far less skewed toward the affluent than is
currently the case; the deliberators could make far greater efforts toward
understanding rival perspectives; and they could benefit from experi-
ments to resolve questions currently answered by guessing.
Issues about scarce medical resources seem simpler. A society com-
mitted to offering health care for all its citizens must consider the ways in
which a budget of medical resources is to be divided up. Focusing on
giving equal opportunities for a worthwhile life inspires the proposal
that preventative procedures promoting health in childhood and matu-
rity are more important than late rescue efforts enabling old people to
hang on for a few weeks or months at greatly reduced capacity. Accept-
ing that proposal would reshape the character of medical expenditure in
the United States, in which, at present, half of the money spent on a
person’s health is expended in the last years of life.5 One might even go
further, proposing that, for some elderly patients, the principal task of
medical care is to ease the process of dying. A good death is, for many
people, a component of a good life.
Public appreciation of that fact would mark a salutary shift in our ethi-
cal life. In many hospitals, many doctors quietly appreciate that preserv-
ing life is not a fundamental goal. A more basic aim is to help lives realize
the projects to which those who live them are committed. Many active,
passionate, and thoughtful people reflectively view an envisaged future
state of lingering in a state of dementia, or subject to a condition in which
5. See, for example, Berhanu Alemayehu and Kenneth Warner, “The Lifetime Distri-
bution of Healthcare Costs,” Health Ser vices Research 39 (2004): 627–42.
a failing body is patched together with a tangle of tubes, as a travesty of

who they are. For people who conceive their lives in this way, the solace
of medicine is not that physicians can maintain an absurd life but that
they can offer the gift of a death suited to the life, sometimes by letting
the patient die, sometimes by accelerating the process. Moreover, those
who view medicine as protection against a bad death will often also re-
gard themselves as remedying an altruism failure: instead of receiving
expensive forms of life support, they wish to give to others who might
benefit more from the supply of medical resources. Proposals of this sort
must be evaluated by people with different points of view, fully informed
about the possibilities and committed to mutual engagement.
Consider a medical case where scarcity of resources is likely to en-
dure. However successful a society may be at recruiting people for or-
gan donation, it is hard to have organs with the right properties avail-
able at the right times in the right places to be transferred to those
whose own organs are failing. Hence, priorities have to be set. Doctors
have to decide which patients will be chosen to receive the hearts, liv-
ers, and kidneys that can be transplanted. (Their predicament is, in
fact, the original “doctor’s dilemma.”6) Many of the criteria to which
they appeal look to the future: they assess the probability that the trans-
plant will be successful, evaluate the future life prospects of the patients,
consider how the patients’ lives impinge upon others, and so forth. Yet
one historical consideration comes into play. It is often considered rel-
evant to distinguish cases in which the failure of the organ has been
caused by something beyond the patient’s control (a genetic defect, or a
trauma early in life) from cases in which the patient’s own behavior has
generated the problem (the liver is failing because of a career of excessive
drinking).
Pragmatic naturalism allows us to understand and refine the forward-
looking criteria. If our actions should promote equal opportunities for
the good life, it is appropriate to favor those whose life projects are more
gravely threatened because bad luck has struck early, rather than those
6. See George Bernard Shaw, The Doctor’s Dilemma (New York: Penguin, 1956), in
which a doctor has a newly discovered cure in a quantity sufficient to save one patient—and
there are two potential recipients.
who have already been able to realize many of their aspirations before
the current misfortune. If the lives of many others are bound up with
the patient’s continued existence (as with parents of young children, or
people whose work contributes to the welfare of many others), it is ap-
propriate to give the patient more weight. The ideal doctor—or the ideal
committee of medical planners—would thus attend to the various ways
in which patients’ failures to survive would affect the future quality of
lives, both their own and those affected by their presence in the world.
Ideally, there would be a process of mirroring, in which the possible
futures for an inclusive group of people were surveyed, out of which a
decision could come.
So far, however, no attention has been paid to the past. If the alcoholic
scores more highly on the future-oriented measures than his or her tem-
perate counterpart, he or she will qualify for the single available liver.7
Can the recommended form of consequentialism be sensitive to the
thought that the carelessness of a patient’s earlier life is pertinent, that
people who neglect their health, or even actively cause their own medical
problems, are less worthy as recipients of corrective treatment? It can. A
commitment to equal sharing of scarce resources must consider whether,
in the initial distribution, everyone received the same. In the original
context of food sharing, for example, if the portion given to one member
of the group were to spoil quickly, our first ethical ancestors would likely
try to make up the lack, perhaps by having the others contribute from
their own allotted shares, viewing this as producing the equal distribu-
tion they had aimed at, but failed to achieve, in the first place. Similarly,
a person whose liver has failed because of a genetic defect was not
granted an equal portion of the resource healthy liver, and a commit-
ment to equality (equal opportunity for a worthwhile life) has to take ac-
count of this fact.
An ideal distribution of the scarce medical resources would attend to
the variety of forward- and backward-looking factors, balancing them in
7. In practice, of course, the alcoholic’s past excesses may diminish his or her future
prospects, and thus reduce his or her score on the forward-looking measures. But to ex-
clude the alcoholic in this way would respond to the future consequences of his or her be-
havior, not to the behavior itself. Many people would maintain that, even if this yields an
acceptable result, it does so in the wrong way.
ways representatives of different perspectives, committed to terms of

mutual engagement, would endorse. In contemporary contexts of medi-
cal decision, however, an attempt to work out how the idealized discus-
sion would go is only rarely appropriate. Judgments frequently have to
be made quickly, and those affected by them need to be assured that
their own cases are not prey to capricious or partial decisions. Hence,
knowing there will rarely be time to investigate all the details of indi-
vidual cases, to understand the complicated ways the continued exis-
tence of one person affects other lives, well-informed individuals com-
mitted to mutual engagement are likely to prefer rules and guidelines
elaborated in advance. These guidelines are to be applied in the urgent
cases, subject to the proviso that more refined judgments—judgments
closer to the ideal survey of particular details and consequences—can be
introduced when time permits.
§59. Habits and Their Limits

Turn next to mundane contexts of ethical judgment, as they arise for
people whose lives are comfortable. Although most of what these people
do is a matter of routine, the exercise of habits that go unquestioned,
each day brings occasions for judgment. There are tasks to be carried
out, and agents need to organize their activity so as to achieve those things
others rely on them to produce. Despite the fact that many ways of pro-
ceeding will typically suffice, some minimal planning is usually required.
Besides the discharge of their duties, however, there are also occasions
for doing more, for helping those around them or for giving unantici-
pated pleasures. People whom others regard as particularly thoughtful
are more alert to these opportunities and take them into account in plan-
ning their daily actions—they make time to telephone a distant friend
who has had some setbacks, to prepare a surprise for the children, to ask
the elderly neighbor whether he or she needs a ride to the doctor. Those
not so sensitive to the possibilities of helping often chide themselves
for allowing their lives to be so constantly guided by habit—and make
resolutions to do better.
This is the humdrum stuff of ethical life for many well-to-do people.
The social environment consists of an interwoven tapestry of routines,
which can be followed to promote what the people involved identify as

the well-being of all. Institutions (such as marriage and private prop-
erty), roles (wife, teacher, citizen), and conceptions of good individual
lives (economic comfort, family harmony) are taken for granted. Given
these institutions, roles, and assumptions about what promotes the good
life, the routines work smoothly together. If each of the people involved
carries out the duties associated with his or her role, the individual com-
ponents of welfare will be advanced and good lives for all in the local
circle will ensue. If people are sensitive to the lives of others, taking ad-
vantage of opportunities to provide extra aid and comfort, there will be
additional benefits. Some roles—family member or friend, say—may
even require a certain level of sensitivity, expressed in a willingness to
deviate from routine, even though more than the minimum would often
be welcome.
The anatomy of everyday ethical judgment in contexts like these is
easy to discern. The agents draw on the ethical resources passed down
to them, general injunctions to do certain things and avoid others, pre-
scriptions associated with roles and institutions, often illustrated by sto-
ries of exemplary figures, vague ideas about what makes for human wel-
fare. In everyday planning, these ethical resources mark out some goals
and rule out particular means—food must be bought so the family can
eat, but the money cannot simply be taken from someone else’s pocket.
Even within the framework supplied by the accepted ethical perspec-
tive, however, there is an enormous range of possibilities. The agents
know that particular things have to be done, and they can recognize,
more or less clearly, some possibilities for going beyond their assigned
duties. How should they budget their time, dividing it among various
domains of action? Because the options are so numerous, a thorough
exploration of what plan of action would be best would preclude any ac-
tivity whatsoever. Some thought is required, but not too much. As we
judge our own days, and those of others, we criticize the thoughtless
who run entirely by routine and their polar opposites who are paralyzed
by the difficulties of choosing, praising those who give enough consider-
ation to pick a plan that allows discharging responsibilities to be com-
bined with some extra benefits to others but who stop deliberating in
time to carry it out. Praise can be combined with the recognition that
some alternative might have been a little better, and that many would
have been equally good, provided the superior possibilities would have
been hard to appreciate in the time available for decision. Reflections on
everyday life are pervaded by concessions to our limitations.
The mundane character of ethical deliberation and ethical judgment
is easy to overlook and takes for granted so much that might be exam-
ined and questioned. Talk of ethics gravitates too quickly to striking
cases, often to cases both striking and artificial. The reality of everyday
ethical life, however, consists in attempts to use and extend the collec-
tion of resources previously acquired, and, in doing this, there is no
single answer an ideal system of ethics might in principle supply, nor do
we judge ourselves and others from the perspective of any such system.
Against the background of what we already have, an ethical code that
presupposes institutions, roles, facets of the good life, and ways of allow-
ing for human limitations, each of us considers what to do, and, retrospec-
tively, appraises his or her own conduct and the deeds of others. Judgment
turns on a number of obvious questions: Is the conduct in accordance
with the general prescriptions of the code? Does it discharge the duties
associated with the agent’s role(s)? Does it appreciate opportunities for
contributing to welfare that go beyond those roles? Does it exhibit too
much thought or too little? Are any lapses understandable as consequences
of human limitations affecting (almost) all of us? Answers typically allow
a significant range of plans of action to count as permissible, even praise-
worthy, applications and extensions of the code.
The everyday practice of ethical judgment might itself be judged from
a more extensive perspective. Is there occasion for modifying the code
itself, the prescriptions it offers, the institutions and roles it presup-
poses, its conception (tacit or explicit) of the good life, its recommenda-
tions about the proper extent and the proper occasion of deliberation, its
ways of accommodating human frailty? Stepping back and judging all
aspects of our code at all times is plainly impossible. Decisions about
what to do next have to be made, and they cannot await protracted scru-
tiny of everything we have inherited. Equally, failing to raise questions
about the various features of our codes, about the ways we answer the
questions on which everyday judgment turns, is the unthinking devotion
to routine our mundane practice condemns. While most of our ethical
life consists in using the ethical resources we have to deal with novel sit-
uations, our current ethical code—with its appreciation of the dangers of
thinking too little as well as those of thinking too much—requires us to
probe its constituent resources. How does the conception of the good (in
terms of equal opportunity for the good life) bear on institutions and
roles we normally take for granted? How does the method of delibera-
tion (in terms of ideal mutual engagement) affect the principles we try to
apply? Are we too slack in allowing people to steer by their routines, and
not to engage, more frequently, in attempts to mirror the attitudes of
others? Is the practice of everyday judgment itself too comfortable?8
Almost certainly, contemporary members of affluent societies (philoso-
phers included) are often caught up in the unexamined life. Operating by
habit, we narrow the horizons of ethical concern, so the plight of many
people becomes invisible to us—hence the tolerance for worldwide in-
equalities, even those as glaring as those just considered (§58). Routine
inclines us to consider a restricted range of options, among which we seek
one as good as any rival. Often, everyday ethical judgment involves prob-
lems to which there are many adequate solutions. The challenging exam-
ples, those able to jar us into considering reform, are occasions on which,
given the resources of the ethical code, there appears to be no available
solution. Ethical dilemmas are not the principal part, or even a very large
part, of ethical life—at least, not for those who live comfortably.9 Their
value lies in causing us to pay attention.
Dilemmas can be more or less isolated, or conversely more or less sys-
tematic. Systematic dilemmas are those that can erupt for different people
8. Although we have seen (§27) why ethics cannot be founded on the will of the deity,
there is an important question about the positive impact of embedding ethics within reli-
gion. William James argued that a connection between ethical precepts and ideas about
powers transcending the human makes ethics properly “strenuous” (“The Moral Phi loso-
pher and the Moral Life,” in The Will to Believe [New York: Dover, 1956], 211). For all the
problems of that connection, one might maintain that, as a matter of fact, the discipline of
religious participation provides structures that encourage people to examine their conduct
more closely and thus prevent certain types of slackness. If this is so, secular societies re-
quire substitute ways of fostering self-examination. (The example of Socrates, whose refer-
ences to deities are peripheral to his probing interrogations, shows us that uninstitutional-
ized surrogates are possible.)
9. They are surely more prominent in the lives of those who live in want.
under a range of contexts, and if this is the case, a proper diagnosis may
require us to understand accepted institutions or roles as responsible.
Suddenly, the background to our habits needs examination. The well-
brought-up conversationalists imagined in §57 might be less dependent
on these occasional provocations, more constantly attuned to the possi-
bility of habit-induced myopia.
§60. Conflicting Roles

Many people in affluent societies, predominantly but not only women,
fi nd themselves torn between the demands of the workplace and those
of the home. As §24 reported, late-nineteenth-century opposition to the
entry of women into public life centered on claims that working women
would inevitably fail to discharge important family duties—and that
argument is often heard today. More to the point, almost every mother
who works outside the home has felt, at some point in the early life of her
children, that she must choose between doing things required by her job
(often things that must be done to retain her job) and things good for her
children (often things her children need). To a lesser extent, the same
dilemmas are felt by fathers too, and one might even assess the degree of
equality within a relationship by gauging the frequency of these pres-
sures on both partners. Often the solutions are suboptimal, achieved by
cutting corners in the workplace (in ways that curtail opportunities for
advancement or even put employment at risk) or by substituting ex-
tended periods of inferior care for parental attention. Richer families
sometimes buy the ser vices of caregivers (usually poor women) who can
offer closer approximations of the focus the parents would like to pro-
vide themselves, but, in doing so, they frequently transfer the problem:
the caregivers must themselves provide substitute care for their own
children.
Oversimplifying, one can distinguish two cases. There are women
whose central project for themselves involves raising children in ways
equipping them to pursue their own chosen course, one selected from a
wide range of possibilities. Because of their economic circumstances,
these women are forced to take on a secondary role, that of worker in the
public sphere. A job outside the home is simply a necessity if they are to
provide for their children and make the envisaged opportunities avail-
able. Yet to keep their job they are required to frustrate their central
purpose. Sometimes, even when the children are sick or distressed,
they must be left with overburdened caregivers or sent from school to
after-school activities because parents cannot leave the workplace to
provide support and comfort. The internal logic of examples like these
is that a chosen role (child rearing) requires a secondary role (working)
that sometimes can be pursued only at cost to performance in the pri-
mary role.
The second class of cases involves women who hope to combine
career satisfactions with responsible motherhood. In many of these in-
stances, more options are available: the aspiring lawyer, doctor, busi-
nesswoman, or researcher has enough disposable income to purchase
better substitutes for her own presence. If, however, the career goals are
to be realized, the same squeeze is felt in a different place. Maybe a ten-
hour workday can be managed, if sufficiently hefty expenditures on
child care secure nurturing and stimulating environments for the grow-
ing children. Career advancement, even career maintenance, may de-
mand even more. Ten hours at the desk, in the hospital, on the road, or
at the bench is not sufficient. The childless—and some of the fathers—
are able to devote more time and energy, and they are likely to receive
the most prized opportunities. Eventually, perhaps, the career of the
promising young woman stagnates—she has put in “only” her ten-hour
days, leaving for a few hours with the children as others remain—and as
she recognizes stagnation, even the ten-hour day becomes unbearable.
Here, the logic of the situation is that a life centered on filling two roles,
both viewed as important and neither prior to the other, is constantly
affected by their competing demands.
Given luck in the form of sufficiently favorable circumstances, women
(and men) who fall under these types may stagger through, achieving
something they retrospectively identify as satisfactory. The parent for
whom work is simply a means to support the children sees them happily
launched on lives they have chosen; the parent who hopes to combine
career and family reaches a similar goal while also managing a satisfying
professional life (perhaps the career could have gone farther without the
“distractions” of the family, but it has gone far enough). With great good
fortune, women and men can feel even that they have discharged both
roles to the best of their abilities: their child raising has gone as well as it
would have done, even given more “investment” in family life; their
career has gained whatever rewards10 it would have garnered, even if full-
time efforts had been devoted to it. Yet the most successful ways of bal-
ancing may require squeezing out other roles, limiting the range or depth
of friendships, for example.
If we compare the women and men who strive to combine family and
professional life, and who constantly feel these dilemmas, with the class
of men (and a smaller number of women) who either pursue careers
single-mindedly or else rely on another (a spouse or a partner) to take
care of the family responsibilities, successes, especially at the highest
levels, will be found predominantly in the latter group. These people,
after all, do not require great good fortune to achieve what they might
have attained with maximum effort—for maximum effort is just what
they devote. Even among the most egalitarian communities in the afflu-
ent world, the sex ratios in the two groups are different: more women
than men find themselves confronting serial dilemmas, generated by
clash of roles; more men than women are freed from those serial dilem-
mas because they can depend on someone else to discharge the poten-
tially clashing role. Hence, an imbalance in performance will persist.
The most spectacular successes in public life will typically be achieved
by men rather than by women, even though opportunities are, supposedly,
now equal.11
The recurrently arising dilemmas stemming from clashes among
domestic and public roles give rise to two distinct ethical problems. The
more obvious of these is that which confronts the individual agents as
they must make their everyday decisions. So long as the demands of the
confl icting roles are not too severe, these people can resolve the sequen-
tial predicaments by judicious accounting. Those who stagger through
to success typically do so by keeping a running tabulation of their
10. These need not be matters of salary, or public honors, but the sense of having made
a contribution to an important enterprise.
11. For a brilliant summary of the manifold ways, many of them small, in which equality
is undermined, and of the large cumulative effect of apparently minute differences, see
Virginia Valian, Why So Slow? (Cambridge, MA: The MIT Press, 1998).
shortfalls in each department of their lives: the urgent need to work

late this evening requires the children to be left longer at the day-care
center or after-school program, but that can be made up to them on an
upcoming state holiday; staying with the sick child this morning can
be compensated by working late at night, once the other parent is
home. Ethical decision making under these relatively benign condi-
tions is straightforwardly consequentialist: one has a pair of goals to be
attained, and by borrowing time here and returning it there it is possi-
ble to achieve tolerable levels with respect to each. But things easily
go wrong, as severe demands from both roles make any such flexible
accounting impossible. An accident of some kind requires the parental
presence for an extended period; during that same period, significant
time must be put in if the work of years is not to be entirely wasted.
Dilemmas of this intensity are felt by a not inconsiderable number
of parents, and they are as difficult as any of the famous classical
examples.
The second ethical problem arises against the background of a com-
mitment to pragmatic naturalism’s conception of the good. To adopt
the ideal of equal opportunity for a worthwhile life is to insist on genu-
ine freedom of choice in life projects for all. Even in the most affluent
societies, opportunities have traditionally been limited to one sex, and
the mere declaration of permitted access to all modes of public life is
only a formal advance, so long as the most severe clashes of work and
family roles remain unresolved. Free formation of aspirations depends
on expectation that the goal is attainable, and, in a situation in which
attempts to combine career and family are always and evidently under
threat from dilemmas of varying severity, clearheaded young women
will feel that they must make an early choice. As they contemplate the
imbalance in achievement at the very highest levels—the predominance
of men in the most rewarding positions in many areas of public life—
they believe that certain career choices simply cannot be successfully
combined with the parental life they want. Formally available they may
be, but women simply do not have equal opportunities for successful
pursuit of them. Even the luckiest women, whose efforts open up a wide
spectrum of possibilities for their daughters, cannot free those daugh-
ters from the shadow cast by less fortunate parents, whose patent diffi-
culties offer cautionary lessons to young women who are deciding on

the direction of their lives.
When dilemmas arise so systematically for a particular class of people,
the principal ethical challenge is to consider ways in which the underly-
ing conditions might be changed so future generations are not similarly
affl icted. It is worth being completely explicit about the structure of
the situation. For some people in affluent societies who meet the formal
conditions for choosing goals others are able realistically to pursue
(women, for example, who are formally able to choose, like their male
counterparts, to combine family and career), the consequences of pursuing
those goals will be to face a recurrent series of dilemmas. At best, that
series of dilemmas will be navigable by keeping a firm eye on both aspects
of the combination and maintaining a balance in both accounts. Quite
frequently, however, that strategy will not be available, and the dilem-
mas arising demand a selection of one goal at the expense of the other.
The effect of these unresolvable dilemmas erodes further the condition
of genuine equality of opportunity for later members of the group. Thus
any acceptance of the egalitarian conception of the good should lead
us to seek changes in the underlying social conditions that generate the
series of dilemmas. We should scrutinize the pertinent roles and the social
institutions in which they are embedded.
In the evolution of the ethical project, roles and institutions have been
introduced to solve prior problems.12 When we discover that the roles
and institutions we have generate difficulties in realizing our conception
of the good, it is necessary to reexamine those roles and institutions.
Can we secure the benefits of having roles of caregiver and worker, while
pursuing sexual equality with respect to opportunities for choice? Under
some kinds of conditions of work, that does not seem to be very difficult,
for, when the needs to be satisfied are quite limited, the strains on equal
participation in both roles are diminished. Agrarian societies in benign
environments, for example, can, if they choose, combine high-quality
parental care with sexual equality. The challenge arises for contemporary
12. The “how possibly” account of §20 proposed particu lar problems as giving rise to
specific roles. Even if that account is historically inaccurate, the more general thesis that
roles and institutions respond to problems would not be falsified.
societies in which citizens aspire to far more, societies usually domi-

nated by intense competition for scarce goods and forms of distinction.
Serious thought about the conflict of roles ought to consider the charac-
ter of these aspirations and the value of the supposed goods for which
people compete. We also need to ponder the value of social distinctions
we routinely make about kinds of work and the ways those distinctions
are reflected in individual choices about what is worthwhile.13 Empirical
investigation—and possibly experimentation—should settle unresolved
issues about what, if anything, would be lost by relaxing the constraints
on the role worker, or by making the care of children a more collective
commitment. Differential distribution of material rewards and of social
approval is often viewed as a necessity for high productivity: if no atten-
tion is paid to differences in effort, we are told, workers will be lazy and
the fruits of their labor will be shoddy and scanty. Energetic workers
prepared to work longer hours should be rewarded for their exertions. It
follows, then, that those who give no time to caregiving will, other things
being equal, receive the largest rewards. Arguments of this sort presup-
pose a number of theses for which there is only impressionistic evidence:
we do not know enough about human psychology to determine whether
differentiation of rewards is necessary for involvement in work; we do
not know the extent to which differentiations have to be made to gener-
ate high performance (do we really need highly skewed systems of re-
warding people?); we do not know whether the differentiations must
be intrinsically connected with the work done, so that those who do not
receive them effectively lose the opportunity for engaging in particu lar
types of work.
The last point is especially pertinent to the clash between the roles of
caregiver and worker. From the perspective of the aspiring doctor, law-
yer, or scientific researcher, the especially high salary, the gongs, and the
plumes may not matter. What she wants is the opportunity to carry on a
particular kind of work at a satisfying level. If an inegalitarian distribu-
tion of material rewards were combined with an egalitarian distribution
13. This is a point that has been emphasized in feminist analyses for decades. It is elo-
quently made, for example, by Virginia Woolf. See Three Guineas (San Diego: Harcourt
Brace Jovanovich, 1966).
of opportunities for work, she would be happy to trade the external

goods for the satisfactions of contributing to family life. Trouble arises
because the differentiation extends into the work itself. She does all the
routine work at the understaffed hospital with the personally (but not
intellectually) demanding patients; she works on mundane legal cases;
or she is able to do scientific research only as an assistant on somebody
else’s project. None of that is required by a commitment to differential
rewards. For it would, in principle, be possible to pay more to her male
counterparts (who are freed from burdensome family commitments),
and to honor them with prizes and titles, while distributing the actual
work more evenly. (Of course, whether those prizes and titles are them-
selves worthy of retention is an issue deserving scrutiny.) Without pur-
suing questions about the balance of work and leisure, the roles could be
revised so that putting in the longest possible workday was no longer a
requirement for having the opportunity to do the most satisfying kinds
of work.
This possible reshaping of the role worker is a very conservative one.
From the stance of pragmatic naturalism, it would be unwise to leap at it
without both acquiring more empirical information and also consider-
ing more extensive revisions. It is, after all, entirely possible that a world
in which many people are educated to aspire to lives in which they work
at the expense of leisure time with their partners, parents, and children
is not quite the best world from the perspective of the quality of the lives
achieved.
§61. Ethically Insulated Spheres

As §§38 and 46 emphasized, our ethical codes contain vague maxims,
likely to remain as we make ethical progress: “Tell the truth” and “Do
not initiate violence” have served as exemplars. Yet our practice divides
conduct into two spheres, in one of which the maxims that hold sway in
the other are relaxed.14
14. Here, I consider only the toleration of violence in warfare. Concealment, even dis-
honesty, is allowed in some commercial and political contexts. For preliminary analysis,
see my “Varieties of Altruism,” Economics and Philosophy 26 (2010): 121–48.
Modern warfare employs techniques often leading to the deaths of

numerous civilians, and contemporary ethical codes tolerate a certain
amount of “collateral damage.” There are limits. Almost nobody thinks
relentless policies of raping and murdering people who belong to an en-
emy nation, but who have played no role in combat, can be condoned.
Yet if an overall campaign is viewed as justified, perhaps because it
resists those who would subjugate or destroy others, we allow means to
pursue it that are recognized, in advance, as likely to cause the deaths of
noncombatants—and even of people not associated with the enemy.
There is an intricate philosophical defense of such uses of violence, one
that turns on claiming that the deaths in question are “foreseen but un-
intended.” Even though their occurrence is regrettable, actions causing
them are excused because of a larger purpose.15
When the allegedly greater end is dubious, any such defense of vio-
lence is undermined: the dropping of bombs on innocent civilians
unfortunate enough to belong to a nation falsely accused of threatening
another is an act of murder, as culpable as any other (and the criminals
who knowingly falsify facts to gain support for their aggressive acts
deserve punishment). Even when the end to be achieved has properties
justifying initiating aggression, when (for example) the goal is to stop a
dictator bent on a systematic policy of genocide, there are serious ques-
tions about acceptable means. Is it permissible to take a greater number
of civilian lives in order to avoid a smaller number of deaths on the part
of combatants? Can one legitimately save “our” troops by dramatic explo-
sions of bombs that kill (immediately or eventually) large numbers of
“their” noncombatants (and even many lives of people unaligned with
either side)? We think there are relatively clear cases, determined by the
balance of the accounting. When the threat posed by the enemy is very
large, so the lives of millions of people are already at stake, a policy
causing the deaths of thousands of noncombatants might appear ac-
ceptable (assuming no alternative would require a lesser loss of life). A
decision to fire-bomb a city containing hundreds of thousands of non-
15. See the essays collected in P. A. Woodward, ed., The Doctrine of Double Effect
(Notre Dame, IN: University of Notre Dame Press, 2001), for lively debate about this
issue.
combatants, and a similar number of refugees who have suffered at the

hands of both sides, in order to shorten a war by, at most, a couple of weeks,
does not.16
You might think there is some intermediate point where the loss of
civilian life exactly balances the gain from promoting the just cause of
resisting violence and aggression. The egalitarian conception of the
good makes that thought problematic. Appreciating the value of human
lives does not fit the accountant’s perspective. Is there some more pre-
cise version of our ethical code that would introduce specific numerical
ratios marking the limits of “collateral damage,” a function assigning to
the number of soldier-lives spared, or the number of war-weeks averted,
some definite maximal number of civilian deaths to be allowed? Would
that specification be progressive, in revealing where the limits of permis-
sible violence lie? A defense might tout the advantages of knowing just
how far military tactics can go. Nevertheless, the defense ignores the fact
that the search for a precise limit is part of the callousness about human
life to which it seeks to respond. A world in which policies were straight-
forwardly justified by showing how they fell within the permitted range
would be one failing to recognize the real problem, a world with too little
regret. Hard choices need to be made, but in making them we should not
forget that they are hard.17
For the dominant ethical task is not so much to find some way of
drawing lines, as to find ways of reconfiguring human life so the lines
become unnecessary. From the perspective of pragmatic naturalism, the
clashes of nations (and other groups) in the contemporary world are a
scaled-up version of the intrasocietal conflicts of our ancestors. At the
earliest stages of the ethical project, groups devising the most successful
experiments would have had to confront the threat of violence that is
constant in chimpanzee social life (and was prevalent in hominid social
16. The bombing of Dresden was probably inspired by the need of the Western Allies to
satisfy Stalin’s demand for action. The official claim was that this act would break the Ger-
man morale (in odd neglect of the fact that the Blitz had only stiffened British resistance to
Hitler). Those who crafted the policy knew that Dresden had become a sanctuary for many
people from Eastern Europe who were fleeing the advancing Russian army.
17. My formulation here is intended to allude to Isaac Levi’s pioneering work on this
topic. See his Hard Choices (Cambridge, UK: Cambridge University Press, 1985).
life).18 Part of the ethical task was to frame rules for dealing with aggres-
sive interactions once they had broken out. Even more crucial was the
formulation of rules for preventing conflicts from erupting in the first place.
Egalitarian arrangements were essential at those early stages because the
group could not manage without the ser vices of any of its members and
because failing to attend to the needs of any member would lead to defec-
tion or outright conflict.
Pragmatic naturalism treats the contemporary world as a scaled-up
version of this initial stage in our ethical life, one in which our inherited
social technology is inadequate to the expanded field on which conflict
has emerged. Huge inequalities between groups that have become aware
of one another’s relative situations inspire the have-nots, when they can,
to wrest what they lack by force. When the economic differences become
highly skewed, the most disadvantaged often outnumber those who live
in comfort, and, when they are able to band together, they have sufficient
chances of success to make aggressive action pay. As technology produces
weapons that can be used to inflict massive destruction, great numbers
of people are no longer needed. Armed with atomic bombs, even in crude
and “dirty” versions, small groups of disadvantaged people can threaten
to take hundreds of thousands of lives. Our predicament exacerbates
that faced by our human ancestors—as if, in their small bands, each
member, even the weakest, had access to tools that would realistically
threaten the life of any four or five others the person selected. The chal-
lenge to amend our ethical practice so as to forestall situations provoking
such threats is more urgent than the fragile social life that inspired the
origin of ethics.
The egalitarian approach to the good addresses that challenge as our
ancestors did when they tackled their problems: they introduced rules
for sharing, and so should we (§58). But this is only part of the difficulty
we face. Not all human conflict arises out of the protests of those hitherto
18. As recognized in §10, the evolution of hominid behavior after the split with the com-
mon ancestor of human beings and chimpanzees may have broadened or intensified the
original tendencies to psychological altruism. See, for example, the proposals made by
Sarah Hrdy about cooperative parenting (Mothers and Others [Cambridge, MA: Harvard
University Press, 2009]). Despite these steps, there remained plenty of limitations of homi-
nid altruism and resultant tensions within hominid societies.
deprived. Human history and the human present are full of wars sparked
by religious differences. Equally, some of the most brutal hostilities re-
sult from the ambition of megalomaniacs who succeed in inspiring oth-
ers. To address all the factors generating conflict in the global society to
which we now belong, egalitarianism must be combined with secular-
ism, at least to the extent of undercutting the military enterprises of
zealots who would impose their conception of the divine will by force.
If we could achieve a world in which there were no economic causes to
invade others, and in which any religious exhortation to conflict was
viewed as illegitimate, if not absurd, two of the major sources of warfare
would be dammed up.19
Even though we can hope to reverse the biblical (and Malthusian)
dictum that the poor will always be with us, the same is surely not true
for megalomaniacs and psychopaths. Their power to generate wars would
be radically diminished if they were deprived of their two major means
of securing support: through rallying those who suffer from various forms
of inequality and through invoking a divine mission. Pragmatic naturalism
does not suppose that the pacifist project of working to eliminate war is
unrealistic, and that we have to settle for articulating the precise condi-
tions under which various kinds of lives should be traded. It can draw
inspiration from the thought that our first ethical ancestors learned how
to solve the problem in the small—and also from the more recent fact that,
on a continent dominated, through centuries, by fierce wars among neigh-
bors, Eu ropean states have entered into relationships that are broadly
(if insufficiently) egalitarian and secular. To many who look at their ac-
complishments, the hostilities of the past seem no longer possible. It is
not entirely utopian to think their achievement could be more broadly
emulated.
A package of proposals is required. In a world constantly scarred by
acts of violence, it is necessary both to find ways of removing the causes
of conflict as speedily and as thoroughly as possible, and to institute
rules for responding to the brutality that continues to occur. Pragmatic
19. Achieving that kind of world might also enable us to address a third source of
confl ict—the shadow of the past, with its recurrent, often escalating, violence between par-
ticu lar groups.
naturalists do not think there are magic wands to be found and waved,
but there may be coordinated proposals that would both take steps to-
ward the diminution and eventual eradication of violence and supply
directives for coping with the imperfect precursors of the goal state. Its
optimism derives from the hope that we can learn about the causes of
social conflict and find strategies for enlarging human sympathies.
§62. Maintaining Equality

Pragmatic naturalism commends equality. Yet a familiar concern is so
far unaddressed. At least from the Enlightenment on, critics of egali-
tarianism have argued that, however one might institute equality of
wealth, the distribution would be transitory.20 Differences of talent, in-
clination, and effort would quickly make some people richer than oth-
ers, and the resultant differences would be transmitted across the gen-
erations, typically in amplified form. Maintaining even approximate
equality of wealth would require intrusions that would limit human
freedom.
The egalitarian conception of the good does not demand equality of
wealth, merely a distribution of resources allowing everyone a serious,
and approximately equal, opportunity for a worthwhile life. Neverthe-
less, since that opportunity will rest on vastly improved systems of edu-
cation and other forms of social care, it will depend on drawing from
people who amass the most. If these people are permitted to pass on
large amounts to their children (or to whomsoever they designate as their
heirs) there will be a diminution of equality from generation to genera-
tion, just as the original argument suggested. If they are not, their free-
dom to do what they like with their property is limited. So the challenge
persists, even when we shift from egalitarianism about wealth to the
egalitarian conception of the good.
To address it, we should scrutinize our institution of property. Ordi-
nary thought operates with a very simple distinction: either there is pri-
vate property, and owners have an absolute right to do what they want
20. Hume makes the point eloquently; Inquiry Concerning the Principles of Morals (In-
dianapolis, IN: Hackett, 1983), 28.
with what they own, or property is held in common, and only the group
is permitted to make decisions about how it is used.21 Nobody should
endorse any such dichotomy. Conceptions of property lie on a contin-
uum. At one pole are those conceptions that subordinate all decisions
about the use of an item of property to the group (impossible in anything
but the very simplest society); at the other are conceptions that sub-
ordinate all decisions about the use of an item of property to the individual
(subject to legal constraints: owning a gun does not give you the right to
shoot somebody). Between these poles lie a host of positions that specify
two kinds of contexts, one where the individual owner’s will holds sway,
another where the community judgment is decisive.
An institution of property consists in a set of rules about how one may
acquire it, and a set of rules about what one may then do with it, without
intervention from others. In the Anglo-Saxon world, the dominant insti-
tution derives from a seminal discussion by Locke.22 Primarily focused
on the example of land, Locke famously proposed we acquire property
by “mixing our labor” with it—that is, making use of it in some way—
and we are entitled to acquire anything previously unclaimed by others,
subject to two provisos: first, we take no more than we can make use of,
and second, we leave as much and as good for others. Once the property
is acquired, however, we are entitled to do what we want with the entities
we own, given that we do not break any law by doing so. In particular,
we are allowed to give our property to anyone we want, so there can be
unlimited transfer across generations. The stage is set for arguments that
egalitarian distributions must be transitory.
This conception is incoherent because of the asymmetry between its
account of the acquisition of property and its relaxed attitude toward
property transfer. When property is acquired, the provisos constrain.
That accords with egalitarianism: initial distributions of property
must allow all to have approximately equal shares. Transfer does not
21. This oversimplified conception underlies the polarized rhetoric that emerges in
some political discussions: advocates of communal property think of (private) property as
theft; contemporary conservatives sometimes view taxation as a form of theft.
22. Second Treatise on Government (Indianapolis, IN: Hackett, 1980). For contempo-
rary exposition and defense, see Robert Nozick, Anarchy, State, and Utopia (New York:
Basic Books, 1974).
have to meet these constraints. How can the asymmetry be justified?

If the wants and needs of others are relevant in the original context,
why do they become irrelevant later? Why do people alive at the time
of acquisition receive the protection of the provisos, while latecomers
do not? Can temporal position properly affect your life prospects in this
way? 23
Imagine a world in which generations of property owners are discrete
and synchronized. In each generation, the elders own all the property
and support their dependents. When they die, property is transferred
to one of the dependents, who takes over the role of elder. Initially, the
world is divided into equal portions, and the first elders each control
one of these. Their successes are unequal, and, in the next generation,
the inherited portions diverge from one another. So it goes, until after
many generations, some are extremely rich, while others have no ability
to meet the most basic needs. In the later intergenerational transfers,
property is passed on without leaving as much and as good for others.
For those who inherit from impoverished elders, the situation is exactly
the same as it would have been for people in the original situation, if
there had been no requirement to allow for the needs of everyone. If
it was appropriate for that original situation to debar the acquisition of
property that shortchanged others, why should one now allow for
the transfer of property that leaves some, perhaps many, in want? Are the
sins of the fathers (presumably profl igacy and sloth) visited upon the
children?
Any such doctrine of original sin is a serious error. Later generations
are not accountable for what earlier people have done, and the chances of
a worthwhile life should not be held hostage to the vagaries of history.
So: restore the symmetry. Intergenerational transfers are permitted only
to the extent that they provide for the heirs what they can use and leave
as much and as good for others. The most straightforward mechanism
for doing this is a rigorous system of taxation setting limits to what
can be transferred across generations. Tax revenues support the public
structures needed to give all young people serious chances for a worth-
while life.
23. In discussing these points, I have been helped by conversations with Rohan Sud.
More than a century ago, Mill argued for a proposal of this kind.24
His reasoning was double sided. In addition to the more obvious con-
cern that children born to poor parents should be provided for, he wor-
ried about the effects on the recipients of dynastic wealth.25 Instead of
offering enhanced opportunities for choosing and pursuing a person’s
own life project, acquiring a very large fortune can prove limiting, sap-
ping the desire to exert oneself. Whether that psychological generaliza-
tion holds, there is an important point behind it. If what is fundamen-
tally valuable is not wealth per se, but the opportunities wealth brings,
history reveals many worthwhile human lives that began in far from wealthy
circumstances, as well as children of the extremely rich who lost their
way. Pending more detailed analysis, it is not obvious that leaving one’s
children a fortune is more likely to bring them a worthwhile life than
providing them with the means (protection, nurture, material necessities,
educational opportunities) to discover their own way.
Does the proposal wrongly curtail human freedom? “What I have
earned is mine, and I have the right to do what I like with it. When that
right is abridged, I suffer a loss of freedom.” The protest begs a crucial
question. For the issue is what it should mean to say that something is
“mine,” and one cannot choose by fiat from a continuum of possible posi-
tions. If it would be ethically progressive to introduce a modified concep-
tion of property, amending the asymmetry infecting Locke’s proposal, the
claim of right lapses.
Behind the protest stands one plausible thought. If we should aim for
a world in which all people enjoy a serious and approximately equal
opportunity for leading worthwhile lives, it is important to consider
whether the modification, one that appears to enhance the opportunity
for some, does so by interfering with the life projects of others. Is the re-
vised conception of property hostile to certain ideas about the worth-
while life, which it eliminates as possibilities for people?
What might these views about worthwhile lives be? Here are the cen-
tral possibilities:
24. Principles of Political Economy (Toronto: University of Toronto Press, 1970), Works,
2:218–27, 3:755–57.
25. Principles of Political Economy, Works, 2:221.
1. A life can be worthwhile through its enjoyment of rare and

expensive material goods.
2. A life can be worthwhile through the effort to accumulate material
goods.
No doubt, enjoyment (use, viewing, consumption, etc.) of material goods

can be part of a worthwhile life. Nevertheless, we should make a twofold
distinction: some material goods are rare and correspondingly expen-
sive, although they have counterparts that are common and relatively
cheap; some material goods are enjoyed privately, while others can be
shared among a large group of people. In light of this distinction, we can
separate scenarios that accord with the first vision of the worthwhile life.
All of these share the feature that what makes the life worthwhile is the
consumption of the goods—the ownership itself—for this is the differ-
ence between the two perspectives. (If the enjoyment does not lie in
consumption, the good will persist and be available for enjoyment by
others.) Imagine, then, a life centered on consumption of rare and ex-
pensive material goods that can be enjoyed only privately—caviar and
fine champagne, for example. The list of such goods cannot be extended
sufficiently far to defend this form of existence as valuable. On the other
hand, if one starts with rare and expensive goods whose “consumption”
(or, in many instances, contemplation) would more plausibly make for a
worthwhile life, there would be no need to make the enjoyment private.
Imagine the life of the connoisseur, who lives in an architectural master-
piece, on whose walls hang masterpieces of world art, and whose eve-
nings are devoted to hearing great music played by the most gifted per-
formers. Would a decision to make these goods public seriously detract
from the value of his or her life? 26 Passive consumption of even the most
beautiful things, whether natural products or human artifacts, is not
obviously sufficient for a valuable life, but, even if one concedes this
point to the first vision, claiming the enjoyment must be private cannot
be defended.
26. It is important to be mindful here of Mill’s point about the need for solitude (Works,
3:756). Public resources can be shared without introducing such dense crowds that enjoy-
ment of them becomes impossible.
The second vision sees value accruing not from the toys one has, but
from the activity of accumulating them (already a positive step). Lives of
this sort are not threatened in any way by the envisaged modification of
the conception of property. For the duration of a person’s existence, the
activity of accumulation is allowed to go undiminished (subject to what-
ever taxes are levied before the intergenerational transfer). If there is to
be any frustration of the accumulator’s plan, it must result from the cen-
trality of a further desire: to give what has been amassed to a younger
heir. Assuming the desire is directed toward the perceived interests of
the heir, the thought must be that bequeathing the wealth increases the
opportunities for the heir to live a worthwhile life. Yet, under the egali-
tarian conception, serious chances are already available. Only if the
life envisaged for the heir were one of consumption—after the pattern of
vision 1—would the inheritance be necessary. It would be strangely
schizophrenic for someone to hold the active work of accumulation
worthwhile for himself, whereas a life of passive consumption would be
suitable for another. If the life envisaged as valuable were one of active
accumulation, it would be counterproductive to provide great wealth for
the young.
Vision 2 is more promising as an account of a worthwhile life than is
vision 1 because it emphasizes the work of achievement rather than the
material goods garnered. Material goods are important bases for most
activities around which valuable lives are pursued, but those required
are typically not the expensive and the rare. A world in which all were
given serious (and equal) opportunities for living valuable lives would be
one in which anxieties about accumulating wealth for one’s heirs would
evaporate.
§63. The Challenges of Technology

How might pragmatic naturalism be applied to the novel options tech-
nology has made possible? Consider an area of human intervention
in which secularism and an emphasis on equality of opportunity for a
worthwhile life would make an evident difference. Biomedical technol-
ogy has offered abilities to carry out in vitro fertilization, to obtain de-
tails about the genomes of adults, children, and embryos, and to clone
mammals—and these opportunities have sparked debates about legiti-

mate ways of using the new techniques. Contemporary societies face the
following questions:
1. Is it permissible to produce early-stage human embryos for

purposes of developing treatments that might aid people with
degenerative diseases?
2. Is it permissible for prospective parents to obtain genetic data
about an embryo or fetus and to terminate a pregnancy based on
what they learn?
3. Is it permissible for parents to produce a child by deploying the
techniques of mammalian cloning?
A striking feature of the discussions about these questions, particularly

in the United States, but to a lesser extent elsewhere, has been the ten-
dency of large groups of people to give blanket answers, based on their
readings of religious texts.
The passages of scripture usually cited supply no direct answers to
these questions: the issues lie beyond the conceptual horizons of their
authors. People inspired by those passages believe a human zygote is
inviolable: it is wrong to destroy a zygote or to manipulate it, or even to
bring it into being in a deliberate way (as would be done by reproductive
cloning). Religious people prefer less abstract, more homely, language.
The embryo is, from the moment of conception, a “human being,” it has
an immortal soul and is sacred; to decide on what features it is to have is
to arrogate powers belonging only to God.
Pragmatic naturalism regards claims like these as having no place in
ethical discussion. They depend on very particular ways of interpreting
the religious texts, and alternative ways of reading are offered by other
devout people, for whom the texts are canonical. Even supposing the
interpretations are well grounded, the texts have no ethical authority. If
there is a basis for crediting the texts with insight, it must lie in the fact
that the ideas attributed to the nonexistent transcendent authority con-
tain elements that would be taken up by progressive ethical traditions.
In other words, the passages could obtain authority only on the basis of a
prior ethical analysis.
The appeal to scripture stops the conversation just where it should

begin. We do better to start from what we know. So we would arrive at
the following:
1. A blastomere is an early-stage embryo produced after several

divisions from the original zygote; further cell divisions are
required before gastrulation, the stage at which the pre-pattern of
the central ner vous system is laid down. Is it permissible to
produce blastomeres in order to find ways of generating stem
cells for research into degenerative diseases?
2. It is possible to obtain DNA samples from a fetus at the fourth
month (and refinements of technique may make this possible
significantly earlier), and to determine whether or not that fetus
possesses genetic markers associated with known probabilities,
in the usual human environments, with various phenotypic
properties. With respect to what properties, and what probabili-
ties, if any, is it permissible for parents equipped with this
knowledge to terminate the pregnancy?
3. Mammals can be cloned by inserting the DNA from a chosen cell
into an enucleated egg, stimulating the product to begin cell
division, and inserting the early-stage embryo into the uterus of
another mammal of the same species. When this is done, there is
a known frequency (quite low at present) with which the preg-
nancy will proceed to term, producing a mammal with the
nuclear DNA originally inserted. Under what conditions, if any,
is it permissible for human beings to bring about a child in this
way? 27
Given these descriptions, an initial appeal to pragmatic naturalism’s

conception of the good yields preliminary answers.
With respect to question 1, the commitment to providing equal op-
portunity for worthwhile lives is expressed in helping people whose
27. In each of these formulations, I give only the bare bones of the scientific account of
the options made available by biotechnology. The intention is to draw attention to impor-
tant features without drowning readers in technical details.
degenerative diseases, especially if they strike early in their lives, threaten

their life projects. The chances of giving aid are obtained through bring-
ing into being clusters of cells that are, in one clear sense, human beings:
these clusters of cells are organisms, and the species to which they be-
long is Homo sapiens.28 The blastomeres are not, however, organisms to
which one can ascribe the capacity for having any kind of life project,
or even organisms capable of feeling pleasure or pain—they are a major
embryonic stage away from having even the pre-pattern of a ner vous sys-
tem. There is no violation of the conception of the good in using them
for the specified purpose, and the purpose itself accords with that concep-
tion of the good. While there is no sense in which the proposed interven-
tion ignores or overrides the desires of the organisms manipulated (they
lack any desires), neglecting the aspirations of those who suffer from
neural degeneration (for example) appears as a form of altruism failure.
The second question invites a more complex answer. With respect to
some knowledge about a fetal genome, termination of the pregnancy seems
evidently permissible (perhaps even mandatory); with respect to others,
it appears just as evidently impermissible, and a range of instances de-
mand further exploration and discussion. Consider first the most devas-
tating and excruciating developmental abnormalities—Lesch-Nyhan
syndrome and Hurler syndrome can serve as examples. In these in-
stances, massive retardation is inevitable, and the child will have a short
life, with no capacity for forming a sense of how that life is to be pursued.
Furthermore, pain will be a necessary part of it (in Lesch-Nyhan, this
will stem from the intense urge to self-mutilate), unless the drugs admin-
istered render the child comatose. However the conception of a worth-
while life is articulated, these children lack the opportunity of having
such a life. Here, then, pragmatic naturalism recommends sparing them
pain by terminating the pregnancy as soon as possible. Contrast that
example with that of cases in which the genetic test reveals a condition
that does not interfere with the opportunity for a worthwhile life—female
28. Similarly, the mutant embryos used by Christiane Nüsslein-Volhard in her brilliant
investigations of the genetics of development in fruit fl ies belonged to the species Drosoph-
ila melanogaster. For an account of this work, see Peter Lawrence, The Making of a Fly
(Oxford: Blackwell, 1992).
sex, say, or relatively short stature.29 Here, any wish to terminate the
pregnancy would have to be defended against charges that it embodies
a frivolous view of what matters in human lives, and, in general, we can
expect the defense to prove inadequate. Hence, pragmatic naturalism
frowns on termination in these instances. Finally, there are genetically
identifiable conditions with respect to which the probability of worth-
while life is significantly diminished. Down syndrome (strictly speaking
a chromosomal abnormality) covers a wide range of possible courses of
life. On the one hand are children (and occasionally adults) whose mental
retardation is severe and whose physiological problems cause difficulty
and pain; on the other are people who achieve significant things. Although
great strides have been made in providing environments allowing people
with Down syndrome to develop more fully, we still lack knowledge of
the distribution of phenotypes across the kinds of environments we can
provide. Prospective parents cannot even calculate the chances that
their child with Down syndrome will have the opportunity for a worth-
while life. Under this situation of extreme uncertainty, it is hard to blame
them for choosing termination—even though, in the future, we might ar-
rive at more refined understandings that would change our perspective.
With respect to the third question, it is important to ask why repro-
ductive cloning might appeal. We can distinguish two types of cases: there
are circumstances in which cloning would provide a unique opportunity
for two people to conceive a child with a close biological relationship to
both of them;30 alternatively, someone (or some pair of people) may want
a child with the genetic material of an admired individual. With respect
to the latter case, the conception of the good tells against the use of re-
productive cloning. If children should be provided with opportunities
for a worthwhile life, and if having a worthwhile life involves making free
decisions about what is most central, there are serious concerns about
parental pressures that appear to be implicit in the idea of having a child
29. A caveat is necessary here. There are plainly environments in which characteristics
like these do interfere with the possibility of having a worthwhile life. The primary ethical
imperative in such instances is to do something about the crippling environments.
30. Consider, for example, two lesbians who have no biological relatives who propose to
have a child by inserting the nuclear DNA of one into an enucleated ovum from the other,
and reintroducing the embryo into the second woman’s uterus.
with this splendid DNA.31 To want a child with the phenotype of some-
one one admires is already to set one’s own ideas about what matters
ahead of those the child might choose for him- or herself—the same sort
of coercion James Mill inflicted on his son. By contrast, when reproduc-
tive cloning would be a unique means to biological continuity, matters
cannot be settled so simply. It is worth asking whether people should be
so wedded to biological continuity with those they rear that they resort
to such technologically delicate ways of achieving it. Under conditions
where mammalian cloning requires the use of a very large number of ova
to obtain a single birth, a couple with a desire for children should choose
some more reliable method, even if biological continuity is sacrificed.
These are only preliminary answers to issues about the uses of bio-
technology. Answers ought to be constrained by the broader commitment
to egalitarianism. Sophisticated forms of reproductive technology are
currently available to a minority of the human population, and we should
ask if the costs of making them available to all would be worth the benefit.
Even if one were to conclude that some instances of reproductive cloning
would be permissible, providing opportunities to all those who would
want them might be incompatible with satisfying more urgent medical
needs: biomedical research might use its resources more profitably. The
nonaffluent world has higher priorities. Conversely, there might be in-
creased pressure to engage in genetic testing and early termination of
pregnancies because providing appropriate environments for children
born with various disabilities could not be carried through on too large a
scale—without using genetic tests to reduce the population of the dis-
abled, maintaining good environments for children with particular gene-
tic conditions would draw resources away from meeting the needs of many
other children. Adding these points about the background commitments
of the view of the good does not end the discussion but merely reveals
reasons for possible modification of the initial answers.
Yet a principal concern about those answers, whether in their original
form or modified, is that something is lost in substituting the bare bio-
31. Prospective parents who aim to replicate someone they admire also have radically
misguided views about genetic contributions and are victims of a crudely mistaken genetic
determinism.
logical characterizations for those offered by religious opponents of bio-

technological interventions. Although we should not accept the appeal
to scriptural texts, or the tendentious descriptions religious people offer,
it is not obvious that talk of blastomeres and zygotes is ethically ade-
quate.32 The bare scientific language, one might charge, cultivates an
important kind of ethical blindness.33 Simple appeals to the scriptures,
conversation stoppers founded in myths, need to go. Yet we also need
concepts enabling us to recognize human worth that do not lead to prac-
tices that ignore some groups of people and their desires, treating them
as “commodities” or as “vermin.” The deliberation pragmatic natural-
ism recommends should be alert to this potential problem. In countries
where repulsive eugenic programs have been carried out, many people—
thoroughly secular people—urge caution with respect to the uses of bio-
technology.34 They should be part of the conversation.
How exactly? Clinical language permits interventions—and it is im-
portant to ask if people who accepted those permissions would also be
led to conclusions that, from their present perspective, they would com-
pletely disavow. Can one find ways of making sharp and clear distinc-
tions between using blastomeres for medical benefit and engaging in the
eugenic experiments of Nazi medicine? I believe one can, but the issue
should be thoroughly and carefully analyzed, attending to possible aspects
of the “worth” or “sacredness” of early human life that too simple and
crudely secular formulations overlook. Empirical issues about human
32. This point is made with great eloquence by Ronald Dworkin in Life’s Dominion
(New York: Knopf, 1993). Dworkin’s discussion of issues about the “ends of life” in that
book is an exemplar of the wide-ranging canvassing of different points of view that I see as
the proper form of ethical conversation. Very few writers on ethical topics, whether reli-
gious or philosophical, have offered so sensitive and insightful a treatment of any contem-
porary ethical question.
33. The point here is akin to one affecting the “two-sphere” approach to trade. Sus-
pending maxims fostering altruism toward others within a particu lar sphere (between
buyer and seller) might erode responsiveness to others more broadly. Talking clinically
about embryos might similarly promote a more callous attitude toward (for example) chil-
dren with severe disabilities. Once again, there are psychological facts to be explored.
34. This is particularly true in Germany. The German translation of my book on ethical
consequences of genomic research (The Lives to Come [New York: Simon and Schuster,
1996]; Genetik und Ethik [München: Luchterhand]) aroused far more controversy than its
Anglo-American original.
attitudes under alternative types of socialization lurk here—would people

who systematically thought in terms of the biological concepts be inclined
to insensitivity toward a broader class of human beings? Behind the literal
conclusion of the would-be conversation stoppers, in the emphasis on the
“sacred” character of human life—may lie an ethical insight. For, even
when that language is thoroughly detached from the idea of a creative
deity, it may have normative force, supporting attitudes toward human
life progressive ethical traditions want to cherish.
Conclusion
§64. Summing Up
Tens of thousands of years ago, our remote ancestors began the ethical
project. They introduced socially embedded normative guidance in re-
sponse to the tensions and difficulties of life together in small groups.
They were equipped with dispositions to psychological altruism that
enabled them to live together, but the limits of those dispositions pre-
vented them from living together smoothly and easily. Out of their nor-
mative ventures have emerged some precepts we are not likely ever to
abandon, so long, at least, as we make ethical progress, the vague gener-
alizations that embody ethical truths. Besides those core themes, we
have also inherited a conception of the good that includes confl icting
elements, as well as providing for us a far richer conception of human life
than any the first ethical pioneers could have apprehended. Our ethical
task is to decide how to go on.
The three preceding chapters recommend renewing the project by
emulating the early phases and expunging some of the distortions intro-
duced in later transitions. Pragmatic naturalism’s normative stance con-
sists in an egalitarian conception of the good, focusing on equal opportu-
nity for a worthwhile life, and a method for ethical discussion in terms of
409
mutual engagement within a comprehensive population; both proposals

advocate disentangling our ethical practices from myths about super-
natural beings. Neither religion nor philosophy can pronounce with au-
thority. Ethics is something people work out together, and, in the end, the
only authority is that of their conversation. The final chapter has offered,
in a highly preliminary way, some suggestions—in the spirit of philo-
sophical midwifery—about how we might continue that conversation.
The greatest philosophers of the past focused on problems salient in
their societies, framing the issues in light of what they took themselves to
know. Philosophy today should address, not technical questions spun
off from investigations that no longer touch human lives, but the deepest
challenges of our contemporary predicament.1 Because the totality of
what we know is so hard to survey, our task is daunting—and may well
require cooperative interactions among scholars attuned to different
realms of expertise. We might, however, be encouraged to think of such
cooperative inquiries as the continuation of an enterprise, the ethical
project, that has occupied us for fifty thousand years, and that has been
a major part of what makes us human.
1. See Dewey, Democracy and Education (New York: Free Press, 1966), 328; The Quest
for Certainty, vol. 4 of The Later Works (Carbondale, IL: University of Southern Illinois
Press1984), 204.
Acknowledgments
I entered ethics by the back door. In the early 1980s, as I tried to become
clear about the pitfalls of human sociobiology (“pop sociobiology,” as I
uncharitably called it), it seemed important to assess the famous claim
that the time had come for ethics to be removed from the hands of the
philosophers and “biologicized.” The final chapter of my book Vaulting
Ambition criticized sociobiological proposals to provide ethics with a
biological basis. Yet, as I reflected, I was not entirely satisfied with my
own response. The negative arguments seemed sound, but I imagined
the complaint of an aggrieved biologist: “You have found fault with ef-
forts to reconcile ethics with a Darwinian picture of life, but do you have
any positive view to offer? Is any of the available philosophical concep-
tions satisfactory?” I had no good answer to this complaint, and, around
1985, I set myself the task of finding one.
The present volume is my attempt to complete the task. It has taken a
long time, and a very large number of people have helped me. My views
have evolved in ways I would not have anticipated. Along the way I have
found myself increasingly at odds both with efforts to carry forward the
program Darwin outlined in The Descent of Man, and with the standard
perspectives from which contemporary philosophers pursue ethical and
metaethical questions. I have often worried that the book I have tried
so many times to write would please nobody.
411
412 Acknowledgments
My first efforts were thoroughly Darwinian, inspired by the brilliant

work on altruism begun by William Hamilton, Robert Trivers, and
Robert Axelrod. At early stages, I was impressed with the insightful
books of Allan Gibbard and Simon Blackburn, and I envisaged extend-
ing the theoretical framework for understanding the evolution of psycho-
logical altruism and connecting it with the kinds of metaethical consid-
erations Gibbard and Blackburn advanced. My work was aided by
conversations with Brian Skyrms, to whom I am indebted for his interest
and encouragement, and it culminated in some articles of the early 1990s.
If Skyrms, Gibbard, and Blackburn do not figure largely in this book, it
is not because my respect for their ideas has diminished, but rather be-
cause my investigations have led me to different questions and different
ways of conceptualizing the issues.
By the mid-1990s, my study of primate social life had convinced me
that, while we probably shared some ethically pertinent psychological
capacities with our evolutionary cousins, an ability to regulate conduct
by explicit rules, formulated in discussions among members of a small
group, had transformed the possibilities for living together. I came to
view ethics as an evolving practice, founded on limited altruistic dispo-
sitions that were effectively expanded by activities of rule giving and
governance. Since 1996, I have tried out versions of this approach in
many forums, giving lectures or series of lectures. I am most grateful to
the members of many audiences, whose questions, criticisms, and com-
ments have helped my ideas evolve: prominent among them are Michael
Baurmann, Richard Bernstein, Akeel Bilgrami, Martin Carrier, Nancy
Cartwright, Kate Elgin, Sam Freeman, Bob Goodin, Paul Guyer, Stuart
Hampshire, Dick Jeffrey, Edie Jeffrey, Mark Johnston, Arthur Kuflik,
David Lewis, Steffi Lewis, Bill Lycan, Karen Neander, Fred Neuhouser,
Richard Rorty, Alex Rosenberg, Carol Rovane, Geoff Sayre-McCord, Jerry
Schneewind, Jack Smart, Kim Sterelny, Sharon Street, Thomas Sturm,
Pat Suppes, Bas van Fraassen, Achille Varzi, Jeremy Waldron, David
Wiggins, Torsten Wilholt, Meredith Williams, Michael Williams, David
Wong, and Allen Wood.
During my years at University of California, San Diego, I learned
from many colleagues and students. Patricia and Paul Churchland offered
good advice about my models of psychological altruism; John Batali was
Acknowledgments 413
a wonderful interlocutor and collaborator; George Mandler provided

pointers to psychological discussions I would otherwise have missed.
As the project evolved, I was greatly aided by conversations with Jessica
Pfeifer, Gila Sher, and Evan Tiffany, and especially by the expert guidance
of Dick Arneson and David Brink.
The move to Columbia brought me a new set of colleagues and stu-
dents. Sidney Morgenbesser was a principal influence and was respon-
sible for leading me to Dewey. My Columbia friends and colleagues have
been constantly helpful and supportive, and I am most grateful to them
for creating such a wonderful forum for exchanging ideas.
An important influence on the ideas of this book has come from my
involvement in Columbia’s famous core course, Contemporary Civiliza-
tion, in which I have had the opportunity to discuss ethical and politi-
cal texts with students from a range of disciplines.
A first full draft of the present book was written during a sabbatical
year in 2007–8, spent in Berlin. I learned much from conversations at the
Wissenschaftskolleg with Catriona MacCallum, Randy Nesse, Bob Perl-
man, Sascha Somek, and especially Moira Gatens and Candace Vogler.
During that year, I was affiliated with the Max Planck Institut für
Wissenschaftsgeschichte, spending the year in Lorraine Daston’s de-
partment (Abteilung II). I am most grateful to Raine for her warm hospi-
tality, for the extraordinarily cooperative atmosphere she creates, for
organizing a workshop on my manuscript in progress, and for innumer-
able insights and suggestions, offered on many formal and informal oc-
casions. Thanks also to supportive friends and astute advisers: Michael
Gordin, Maria Kronfeldner, Erika Milam, Tania Munz, Adrian Piper,
and Thomas Sturm.
During that sabbatical year, I received much warm hospitality and valu-
able feedback. I would like to thank Michael Baurmann, Mario Branhorst,
Harvey Brown, Martin Carrier, Nancy Cartwright, Christel Fricke, Ulrich
Gähde, Oren Harman, Paul Hoyningen-Huehne, Matthias Kaiser, Evelyn
Fox Keller, Anton Leist, Tim Lewens, Chrys Mantzavinos, Kirsten Meyer,
Felix Mühlholzer, Susan Neiman, Chris Peacocke, Suzann-Viola Ren-
ninger, Peter Schaber, Thomas Schmidt, Bettina Schöne-Seifert, Tatjana
Tarkian, and Gereon Wolters. Chrys and Mario also gave me extensive
written comments that have prompted significant changes.
414 Acknowledgments
During the many years I have been thinking about this project, there
have been several people who have helped me on a number of occasions.
Conversations with Allan Gibbard have invariably been wonderfully
instructive, and I am delighted that he decided to spend a sabbatical
leave in New York City. Isaac Levi has shaped my thinking in many
ways, most likely more than either of us could enumerate. Interactions
with Frans de Waal, both on the occasion of his Tanner Lectures at
Princeton and during a visit to Emory have enlightened me greatly about
crucial facets of primate behavior. Chris Peacocke has been a splendidly
open-minded interlocutor and a constructive critic. Continuing discus-
sions with John Dupre, Michael Rothschild, and Elliott Sober have been
illuminating and always enjoyable.
My debts to my students at Columbia, especially those who have taken
my seminars on evolution, altruism, and ethics, are extensive. Particular
thanks go to current and former graduate students: Dan Cloud, Laura
Franklin-Hall, Michael Fuerstein, Jon Lawhead, Katie McIntyre, Heather
Ohanesen, Herb Roseman, and Matt Slater; also to some truly brilliant
undergraduates: Lauren Biggs, Leora Kelman, Jonathan Manes, Howard
Nye, Michael Roberto, Sam Rothschild, and Rohan Sud. Many thanks
to Jon Lawhead for preparing the index.
The penultimate draft, prepared in 2008–9, was somewhat longer
than its predecessor—and considerably longer than the final version. I
am extremely grateful to two perceptive and constructive readers for
Harvard University Press who gave me excellent advice—in particular to
make it shorter. It helped that their thoughts coincided completely with
those of the person who has always been my best reader, Patricia Kitcher,
to whom I owe more than I can say.
Lindsay Waters has been a helpful and supportive editor. He has had
to wait a long time, and I appreciate his patience.
My thinking about these topics began before our sons were teenagers,
and that is reflected in the nickname the book in progress still bears.
Charles continues to ask me how “Nice Monkeys” is going—even though
I have told him many times that it is no longer much about nonhuman
animals, that the relevant nonhuman animals are apes, not monkeys,
and that part of the point is that they are not particularly nice. For many
years, I have been entertained, enlivened, challenged, and informed
Acknowledgments 415
through their increasingly sophisticated conversation—and in recent

years by that of Sue-Yun Ahn as well. It has been wonderful watching
them grow. From Andrew, now a doctor, I have learned much about
medical care in New York City—and have been moved by his compas-
sion for his patients, many of them indigent and most baffled by an alien,
monolingual English, environment. He embodies many of the qualities
I have tried to articulate and defend, and I dedicate this book to him
with love and admiration.
Index
abortion. See bioethics archeology, 5, 11, 116

agriculture, 305, 375–376, 389 aristocracy, 141, 174, 180, 328
alienation, 35, 168 Aristotle, 154, 247, 326
altruism: psychological, 5, 17, 19– 20, Arrow, Kenneth, 312
23– 24, 26, 30– 31, 35, 39, 42–51, 53, atomic physics, 182–184, 272
56–58, 63– 67, 71– 72, 82–83, 86–88, atomism in ethics, 210–212, 363
100, 105, 111, 131–136, 226– 229, 232, autonomy, 207, 255, 325
340, 344, 368, 394, 40; failures of, 6– 7, awe, 81, 84, 92, 113, 131, 201, 226, 231
9, 73– 74, 86, 91, 93, 98, 100, 103–104, Axelrod, Robert, 52
107, 111–112, 124, 126, 128, 131–132, 135, Ayer, A.J., 203, 280
137, 143, 150–151, 153, 217, 222– 231,
233– 240, 242– 244, 247– 248, 250, 259,
262, 273– 274, 276, 279, 294– 295, Babylon, 117–119, 145, 216
297– 299, 303– 305, 307, 309, 311, 327, Barkow, James, 36
339– 340, 346, 349– 351, 355, 360, 365, Batali, John, 56
368, 380; biological, 18–19, 37, 49; Beauvoir, Simone de, 147
behavioral, 19, 23, 45–47, 56, 75– 76, 79, Bentham, Jeremy, 289, 293, 314
82, 86, 94, 131–132, 226, 231– 232; vs. Berlin, Isaiah, 249
egoism, 19– 20, 23– 24, 31, 34– 36, 40, Bible (Judeo-Christian): Scriptures, 4, 154,
44–45, 105, 135–136, 230, 274– 276; 158, 160–161, 165; Deuteronomy, Book
golden-rule, 24, 32, 34, 37, 56, 144, of, 119; Exodus, Book of, 119; Genesis,
222– 223, 346; physiological basis, Book of, 130; Jacob, 130; Joseph, 130,
26– 27, 39, 47–49, 199, 309; dimensions 154; Moses, 141, 144, 174; and the
of, 31– 34, 44, 69, 95, 144, 344– 345; practice of slavery, 143, 154–155; Acts,
reciprocal, 50–58, 64– 65, 135; quasi, Book of, 144; Luke, Book of, 144;
56; higher-order, 135–137, 217, 232, 291, Matthew, Book of, 144; Gospels, 144,
327, 339, 345, 371; and asceticism, 227, 155, 166, 214, 236, 250; Pentateuch, 154;
289 Ham, 154–155, 160; Leviticus, Book of,
animals, ethical status of, 180–183, 167
306–310 bioethics, 401–408
anthropocentrism, 66, 310 biology: ontogeny vs. phylogeny, 330–336;
anthropology, 3, 5, 11, 18, 26, 68, 94, 96, cell biology, 402–403
114, 116, 138, 156, 186 Boehm, Christoph, 18, 96, 116
Aquinas, Thomas, 231 Bohr, Niels, 272
417
418 Index
boundaries, 120, 190, 214, 303, 306, 309 compromise, 241, 244, 260, 272, 300,
Boyd, Robert, 53, 87, 109, 112 336
Brownian motion, 182 computer simulation, 52, 56, 63
brutality, 31, 47, 60, 118, 143, 161, 166, 225, conscience, 40–41, 92– 93, 112
229, 278, 307, 395 consensus building, 175, 334–335,
340–344, 348–360
consequentialism, dynamic, 288–294,
calculus, 234, 293 311–313
callousness, 46, 393 constructivism about truth, 190–192,
Canaan, 154, 167 201–203, 245
Carcopino, Jerome, 174 contractarianism, 273, 279, 344
Cardozo, Benjamin, 120 conversation. See discussion; mutual
caregiving practices, 38, 152, 385–386, engagement
389–390 cooperation: discriminating, 12, 55–57,
causation: between affective and emotional 88–89, 174; noncooperation, 52–53, 85,
states, 27–30, 77–78, 80–84; of ethical 89; nondiscriminating, 55
regression, 177–178; in functionalism, Crusades (Christian), 144
219–220; role in evaluating ethical facts, Cummins, Robert, 219
291–294; as a way of demarcating ethical
communities, 302–305
Cheney, Dorothy, 36, 58 Darwin, Charles: Darwinian evolution, 12,
children, and altruistic behavior, 40–42, 36, 90, 107–109, 111–112, 114, 123, 138,
132 213, 219–220, 237–238; coevolution of
citizenship, 106, 128, 140, 145–147, 149, 162, genes and culture, 108–109; social
165–166, 174, 177, 212, 375, 378, 382, 390 Darwinism, 255–256, 258
civilians, 392–393 Dawkins, Richard, 18, 326
cloning. See bioethics deliberation about norms, 111–115, 130–135,
coalition game, 57– 66, 88, 96– 97, 228 143, 259–261, 286, 308–313, 331–332,
cognition: role mediating ethical judg- 340–345, 372–384
ments, 25–30, 37–39, 75–80, 198–200; in Dennett, Daniel, 114
animals, 39–45, 50, 56, 78–80 deontology, 289, 294, 298, 300, 302, 314,
Coke, Sir Edward, 146–147 325, 328
collective: construction of ethics, 2, Descartes, Rene, 205
191–192, 204–205; use of resources, de Waal, Frans, 9, 17, 30, 43–44, 50, 57,
123–124, 318–323; acceptance of an 59– 60, 66, 69–70, 87, 189
ethical code, 234–237, 256 Dewey, John, 3, 114, 120, 179, 207, 251, 264,
colonial, 109, 154 266, 304–305, 327, 342, 347
colonies, 43, 69–70, 146, 153–156, 352 Dickens, Charles, 353
combat, 142, 175, 249, 392–393 discussion: ideal, 96–100, 104–108, 111–112,
commitment, 74–75, 79, 81, 96, 121, 151, 339–341; among groups, 270–272,
195, 200, 211–212, 233–234, 247, 256, 281–282, 285–288, 296–298, 302–329,
280, 324, 331, 352, 373, 380, 388, 331–338, 342–354, 356–360, 363–369,
390–391 394–396; contemporary, 370–373,
communal resources, 125, 129–131, 242, 397 377–378, 383–385, 388–391, 394–396,
compassion, 140, 143, 160, 174, 214 402–404, 406–410
competition: for resources, 60– 62, 248, disobedience, 81–82, 93, 111, 279
255, 350, 375, 390; cultural, 107–115, dispositions, 36–38, 63– 65, 67, 69–74,
128–131, 169–170, 259 78, 87–88, 91, 93, 98–100, 102, 108, 131,
Index 419
145, 161, 179, 183, 185, 191, 198, 204, 217, 226–227, 230, 234–235, 243, 247,
228, 232, 263, 278–279, 303, 335–336, 295, 302, 315, 356–360
340 expertise (ethical), 8, 285–287
divine commander, 111–115, 119–121, explanations, 12–13; how possibly, 60–74,
127–128, 163, 166–170, 207–208, 78–80, 87– 94, 99–103, 104–107, 111–115,
230–231, 241, 260, 289, 294–298 122–131, 135–137, 216n8, 223–229,
divinity, 2, 4, 8, 113, 115, 119, 121, 128, 131, 236–237, 239–242; 242–245, 307–308;
148–150, 163, 165–170, 207–208, how actually, 105–106, 117–119, 173–178,
230–231, 260, 285, 289, 298, 311, 313, 213–229, 294–296, 374–381; evolution-
328, 372–373, 395 ary, 107–110, 114, 213–214, 237–240
division of labor, 117–118, 122–131, 151n29,
239–241, 243–245, 294–295, 327–329,
350–352 Faustian bargains, 41
dominance, 59, 69–71, 228 Fehr, Ernst, 85, 101
Dummett, Michael, 192 feminism, 146–148, 152, 390
Foot, Philippa, 247, 278
forgiveness, 120, 144–145
early friendships, 64– 66 Fossil record, 5, 18
ecology, 59, 319 Foucault, Michel, 296
economics, 23, 50, 65, 85, 312, 391 Freud, Sigmund, 34–35, 92, 95, 163
education, 146, 149, 193–197, 317–319, 324, friendship, early friendships, 64– 66
359, 371, 375–377 function, etiological approach to, 218, 220
egalitarian commitments, 96– 97, 106, functional confl ict, 241–245, 247–249, 252,
115–116, 122–123, 127, 217, 248, 297, 321, 260–262, 281–282, 286–288, 296,
324–325, 351–352, 359, 364–366, 373, 333–334, 337–353, 355–356, 360–369.
377, 389–390, 393–397 See also progress; pluralism (ethical)
egoism, 19–20, 23–24, 31–32, 34, 36, 40,
44–45, 105, 136, 230, 274–276
Egypt, 11, 113, 118–119, 130, 216–217 Galileo, 148, 181, 189
elitism, 314–315, 328 generosity, 121, 136, 141
emotions, 36–37, 49, 76–85, 93– 95, genetic determinism, 27, 50, 114, 406
98–102, 198–204; nature of, 25–32, genetic testing, 380, 402–406
78–79, 198–201 Gibbard, Alan, 96, 203–204
empathy, 31, 34, 144, 250–251, 345, 371 Gilgamesh, 118
enforcer (divine), 111, 113–115, 119, 127, 131, Goldman, Alvin, 209
169, 217, 226, 230–231, 241, 260, 274, Goodall, Jane, 42
294, 297–298 Goodman, Nelson, 3, 258
environmentalism, 305, 310, 321 Gould, Stephen, 10
equality, 11, 85, 96, 115, 157, 178, 248, 285, government: autocratic, 148; democratic,
287, 289, 291, 293–295, 297, 299, 301, 146, 188–189, 265
303, 305, 307, 309, 311, 313, 315, 317–319,
321, 323–325, 327, 329, 339, 341, 364,
380, 385, 387, 389, 396 habits, 97, 331–332, 381–385
ethical point of view, 80–85, 106, 231–232 Hacking, Ian, 183
eugenics. See bioethics Hamilton, William, 18, 49
evolution. See Darwin, Charles Hamlet, 3, 92, 94
experiments: psychological, 45–46, 85–86, Hammurabi, code of, 118–119
101–102; of living, 104–137, 150–153, 213, Harman, Gilbert, 139, 181
420 Index
health: effect on shaping ethical systems, Kierkegaard, Søren, 122

91n28, 101, 112, 137, 150–151, 316; as an Kinsey, Alfred, 164
ethical issue, 296–297, 304–305, 318, kinship, 45, 71, 120, 223, 259
374–381 Kuhn, Thomas, 2, 147, 180, 189, 333,
hedonism, 313, 328, 363 349
Hitler, 393 !Kung San, 96, 115–116
Hobbes: Leviathan, 60, 95; state of nature,
154
Hofstadter, Richard, 256 Lecky, W.E.H., 143–144, 174, 214
holism, 292–293, 314, 316, 363 Levi, Isaac, 101, 245, 393
Homer: Iliad, 141–142; Odyssey, 141–142, Lewis, David, 168
174; as a model of ancient ethics, Lewontin, Richard, 10
141–142; contrast with Solon, 142–143 lex talionis, 140, 186, 194, 197, 202, 215,
hominid, 5, 9, 17–18, 31, 35–36, 47, 59, 61, 234–235
64, 66– 69, 71, 73–74, 80, 87–88, 90, 92, liberation, 2, 13, 29, 81, 93, 111, 163, 173,
96– 98, 103, 116–117, 194, 222, 237–238, 183, 279, 325, 337, 367
251, 275, 279, 297, 304, 337, 367, liberty, 107, 147, 150, 154, 164, 264,
393–394 292–293, 305, 315, 327
honesty, 246–247, 261–262, 266, 271, Lipit-Ishtar code, 118–119
299–302, 358–359, 391–392 Locke, John, 397
Hume, David: Enquiry Concerning the Lowance, M., 154–155, 157–158
Principles of Morals, 99, 213, 274, 308; Lumsden, Charles, 108
problem of induction, 255, 257–258
Hüyük, Çatal, 116–117
machiavellianism, 19, 23, 26, 30–31, 36–40,
42, 44, 50, 53, 56, 65, 69–70, 75–77, 79,
institutions (evolution of), 120–122, 128, 82, 86, 136
137, 146–148, 248–249, 282n32, MacKinnon, Catherine, 146–147
295–296, 372–374, 382–385; private malthusian population control, 320,
property, 120–131, 396–401; marriage, 395
137, 150–151. See also slavery Marx, Karl, 296
internalization of ethical norms, 87, 92– 96, Mathematics: comparison to ethics,
106 192–193, 203, 205–206; foundationalism,
intuition, 179–181, 183, 205–206, 334–336 205
irrationality, 272–273, 275 McDowell, John, 198
Mendel, Gregor, 110
Mesopotamian civilization, 11, 113, 118–119,
James, William, 3, 92, 176, 185, 384 144, 217, 234
Jefferson, Thomas, 157 Metaphysics, 83, 166, 201, 263
Jex-Blake, Sophie, 147 Mill, John Stuart, 107, 129, 143, 147–152,
Jones, Matthew, 213 164, 263–264, 289, 292–293, 302,
justice, 73, 105, 119, 139, 141, 201, 251, 258, 305, 314–316, 325, 327, 351,
274, 312, 334 399–400
mirroring: primitive, 346–347, 350;
extended, 347, 350, 354–357, 364
Kant, Immanuel: Groundwork of the mutual engagement, 136, 340–349,
Metaphysics of Morals, 83, 166, 201; 356–365, 368, 370–373, 376–381,
Categorical Imperative, 206, 245 384
Index 421
Nagel, Thomas, 85, 335 74, 79–81, 90–91, 96–98, 100, 102, 104,
naturalism: pragmatic naturalism (general 112–114, 216–218, 221–223, 225, 228–229,
discussion), 3–13, 225, 246, 256–258, 231, 236–238, 242–243, 251, 258, 261,
268–270, 281–282, 317–324, 362–369; 275, 278–279, 287, 294, 296, 316, 380,
nonnaturalism, 8, 269–271, 273, 276, 393–395; chimpanzee, 11, 17–18, 36,
279; counternaturalistism, 281 42–44, 47, 57, 59–60, 64–74, 87–88, 90,
Nazi, 139, 177, 255 222, 228, 232, 251, 278–279, 393–394;
Neurath, Otto, 279–280 bonobo, 17–18, 44, 47, 57, 59–60, 64–65,
neuroscience, mirror neurons, 28–30, 43, 68, 74; baboons, 36, 60; Arnhem
346 chimpanzee colony, 43–44, 49, 69–70,
Newton, 205, 213 77–78, 189; gorilla, 58; gibbon, 58, 144;
Nietzsche, Fredrich, 92, 95, 143, 211, orangutans, 58, 60
276–278, 282, 360–361 primatology, 5, 17, 36, 57, 60
nihilism, 269–270 primitive vs. civilized, 242–245, 248
noncognitivism, 30, 201–205 Prisoner’s Dilemma: defection strategy,
normative: guidance, 67– 69, 74–87, 51–55, 67– 68, 83, 87–88, 162, 379–380,
90–103, 111–112, 123–125, 131–137, 394; tit-for-tat strategy, 52–53; iterated,
221–223, 230–234, 409–410; conclusions, 52–54, 56, 58, 64; optional, 54–58,
191–192, 253–259, 338 60– 65, 88– 90, 126–127. See also
Nozick, Robert, 397 coalition game
progress: ethical, 6–8, 81, 99–100, 138–170,
173–206, 209–213, 338–342, 391–395,
obedience, 84, 99, 108–109, 111, 160, 167, 402–407, 409–410; technological, 7,
231, 264, 289 218–220, 238–242, 288; atomism about,
objectivity in ethics, 139–140, 209–212, 210–212, 363
243–245, 267, 312 properties, ethical properties as disposi-
tions, 198, 204, 263
Parfit, Derek, 322

paternalism, 133–135, 149–153, 227, quakers, 146, 158, 184
301 Quine, W.V.O., 254, 280
peacemaking practices, 57, 69–70, 73, 87,
116, 221–222
Peirce, C.S., 176, 210, 246 Railton, Peter, 35
philosophy (role in ethical project), 8– 9, rationalism, 361–362
286–287, 370–373, 410 rationality, 272–279
Plato: Euthyphro, 128, 166; Republic, 128, Rawls, John, 105, 156, 201, 223, 245,
274, 367 257–258, 312–313, 334–335, 344
pluralism (ethical), 131, 210–213, 245–249, realism (ethical), 186–208
281–282, 306, 318, 337–338, 355, reductionism, 271–272, 292, 314
367–368, reflective equilibrium, 257–259, 334–342,
pragmatism naturalism. See naturalism 342–343, 349–352
pre-ethical societies, 5, 31, 44–45, 134 relativism, 138–143
prehistoric society: neolithic society, 105, reliabilism, 209, 258
119, 179; paleolithic era, 122–130, religion: and ethical practice, 4– 6, 106,
216– 217, 236– 238, 307–308, 310– 311 111, 115, 207, 285, 298, 331, 358, 372,
primates: hominid ancestors, 3, 5, 7–8, 402; christianity, 140, 143–145,
10–12, 17–18, 35–36, 47, 54, 57, 59, 69, 71, 154–156, 160, 165, 168, 174; evangelical,
422 Index
144, 297; judaism, 145, 165, 176, 206, 230, 232, 264, 278– 279, 333, 359,
254– 255, 299, 301 371
resentment, 26, 134, 198, 204, 308, 324 sociobiology, 9–10, 36, 87, 101, 108,
Richerson, Peter, 87, 109, 112 255–256, 258
Robinson, Jenefer, 27 sociopathy, 279
roles (evolution of), 68–71, 89, 106–110, spirituality, 154, 158, 160, 184–185, 313–314
115–116, 125–131, 146–151, 216–217, stem cells, 1. See also bioethics
239–245, 247–249, 295, 331–332,
382–391
Roman empire, 143–145, 174–175 Tarski, 189, 246
Roseman, Herbert, 63 Taylor, Harriet, 147–152, 293
Rousseau, J., 61– 63, 149 technology: progress in, 7, 218–220,
Rovane, Carol, 139 238–242, 288; ethics as a social
technology, 221–241, 262, 394; ethical
challenges of, 401–408
Sabini, John, 46, 101, 331
Scanlon, T.M., 201, 223, 344
Schelling, Thomas, 73 unseen enforcer, 1. See also divine
Schopenhauer, Arthur, 17 commander
secularism, 4–5, 81, 121, 161–162, 287, utilitarianism, 272, 290–294, 311–314,
325–328, 344, 372–374, 395, 402–408 322–323
Sen, Amartya, 312, 319 Utopia, 84, 316–318, 324–325, 339, 395, 397
sexual behavior: polygamy, 143; hetero-
sexuality, 162–163, 165; homosexuality,
162–165, 174; incest, 255 value (dimensions of), 323–324
Seyfarth, Robert, 36, 58 violence: rape, 140, 234; warfare, 142, 174,
Singer, Peter, 183, 214–215, 307, 335 391–392, 395; coercion, 146, 162, 185,
skepticism (ethical), 12–13, 35–56, 65– 69, 226, 228, 296, 317, 327–328, 359–360,
106, 247, 270–273, 279–282, 287, 318–321 363; genocide, 169, 392
slavery: abolition, 143, 153, 157–158, 161,
184, 187, 200, 214; and christianity, 143,
154–155, 160–161 Wilson, E.O., 39, 108, 256
Smith, Adam, 17, 25, 83, 121, 123–124, 156, Wittgenstein, Ludwig, 99
346, 351 Wollstonecraft, Mary, 147, 149–150, 152,
Smith, John Maynard, 53, 108 163, 179–182, 184, 193–197, 202, 240, 250,
Sober, Elliott, 39, 256 257
social: cohesion, 9, 97, 287–288, 292, women, rights of, 140, 145–153, 179–180,
334–336, 341–342, 353, 362, 365–367; 193–196, 212, 214, 236, 250, 256–257,
asociality, 55, 61; blindness, 176–177, 293, 318–319, 358, 385–389
211–212, 335; fragmentation, 228, 356; Woolf, Virginia, 147, 150, 152, 390
nonsociality, 328 Woolman, John, 158–161, 179–182, 184,
socialization, 93– 94, 97, 101, 107, 126, 193–197, 200, 202, 240, 250, 257
130–132, 148, 151–153, 182, 199, Wrangham, Richard, 58– 60

Philip Kitcher The Ethical Project 2011

Uploaded by

Copyright:

Available Formats

Philip Kitcher The Ethical Project 2011

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Philip Kitcher The Ethical Project 2011

Uploaded by

Copyright:

Available Formats

What is the book about?

What is the book about?

What topics does it cover?

What topics does it cover?

The Ethical Project

The Ethical Project

Harvard University Press

Library of Congress Cataloging-in-Publication Data

Kitcher, Philip, 1947-

3. Experiments of Living 104

II. A Metaethical Perspective

§42. Troublesome Characters 269

III. A Normative Stance

§1. The Shape of Things to Come

1. “Recent Work on the Principles of Mathematics,” International Monthly, 4, 1901, 84.

mathematicians to reach agreement on conclusions that endure fosters

The position to be elaborated—pragmatic naturalism, to give it a

methods of rigorous inquiry. Appeals to divine will, to a realm of values,

10. James, Pragmatism, 97.

people received authoritative information about how they should live,

altruism failures—remains primary. Challenging the enduring impor-

§2. Methodological Preliminaries

ethical project is not simply the unfolding of previously existent altruis-

ing them in light of the evidence, it is adopted as describing the actual

distinction. I have attempted to be clear about the kinds of explanations

The Springs of Sympathy

§3. Psychological Altruism: Basics

psychological capacities of the members of those groups. Fossil evidence,

Biological altruism requires no perceptive or cognitive abilities. Even

A acts psychologically altruistically with respect to B in C just in

expressed in terms of the mode of division: egoists give nothing, self-

vSoc = wEgovSol + wAltvBen

§4. The Varieties of Altruistic Reactions

a physiological response connected to judgments about what has been

This concession does not entail any of a number of conclusions some-

1. A feels an emotion different from the emotion A would have felt in

§5. Some Dimensions of Altruism

sess Anything Remotely Resembling a ‘Theory of Mind,’ ” Philosophical Transactions of

of altruism is represented by how much of the food you are willing to

vSoc = wEgovSol + wAltvBen

who is hopelessly misguided about what the beneficiary wants: almost

cion about completely impartial altruism.19 Moreover, in a world with

§6. Maternal Concern

supposes the propriety of attributing wants and intentions apparently

22. For details, see Kitcher, “Evolution of Human Altruism.”

The best version of skepticism invokes psychological variables corre-

offspring.” Even the most liberal cognitive ethologist is likely to wonder

Hence, the skeptic proposes, mothers do the impressive things they

psychological comfort as trivial compared with the value of saving the

§7. Broader Forms of Altruism?

24. Goodall, Chimpanzees of Gombe, 357, 386.

There was no realistic possibility of her pronounced efforts on behalf of

the hypothesis that, in some of these cases at least, we find genuine

who are allegedly being “trained” or to function as an effective “prison

community. Behavioral altruism of this sort cannot be found in the societ-

§8. Possibilities of Evolutionary Explanation

28. For details, see Kitcher, “Evolution of Human Altruism.”

allows for psychological altruism as one mechanism for helping behavior

Bee. If their helping behavior were based on calculation, it would be