Content uploaded by John Shook
Author content
All content in this area was uploaded by John Shook on Jun 16, 2016
Content may be subject to copyright.
161
Biodata of John R. Shook , author of “ The Design of Morality .”
John R. Shook , Ph.D . is a scholar and professor living in Washington, DC. He is
Director of Education and Senior Research Fellow of the Center for Inquiry.
He also is a visiting assistant professor of science education at the University
at Buffalo and associate fellow at the Center for Neurotechnology Studies in
the Potomac Institute for Policy Studies in Arlington, Virginia. From 2000 to
2006, he was a professor of philosophy at Oklahoma State University. Among his
books are A Companion to Pragmatism (2005), Ectogenesis (2006), The Future of
Naturalism (2009), and The God Debates (2010).
E-mail: jshook@pragmatism.org
L. Swan et al. (eds.), Origin(s) of Design in Nature,
Cellular Origin, Life in Extreme Habitats and Astrobiology 23, 161–190
DOI 10.1007/978-94-007-4156-0_9, © Springer Science+Business Media B.V. 2012
163
THE DESIGN OF MORALITY
JOHN R. SHOOK1,2
1University at Buffalo , Buffalo , NY 14260 , USA
2George Mason University , Fairfax , VA , USA
1. Morality in Nature
Scienti fi c research into our primate cousins and our own human capacities are
providing insights into the long history behind the design of morality and our
facility with conducting and modifying morality (see, e.g., de Waal, 2009 ) . An
account of the origins of human morality must recognize three main phases: a
long period of protomorality evolving among the primates; another long period
of protomorality growing into morality in the several Homo species as brain size
increased; and a relatively brief period down to present times when Homo sapiens
has been self-consciously modifying morality. Both natural evolution and cultural
evolution have been involved in intertwined ways to design morality, and both will
continue to shape morality. A naturalistic account of human morality must be
both backward-looking and forward-looking. This account looks far back in time
when evolution was forming the building blocks of morality and then morality
itself without any conscious designing, and then it shows why humans would look
forward to take some measure of deliberate control of their inherited morality
utilizing only the cognitive resources naturally available to them.
An internally consistent and gradual evolutionary account of these three
main phases is needed. Any abrupt break, some strange leap or dramatic shift,
presents a severe problem for a naturalistic understanding of the development of
morality. For example, if earlier hominid species are thought to have had nothing
like morality, only competing in a fi erce “sel fi sh” struggle where most kindnesses
and cooperations are punished in the long run by natural selection, it becomes
hard to explain how Homo sapiens would be practicing a far more social morality
instead. Some extraneous factor must be abruptly introduced to account
for that leap, and speculative theories abound. Perhaps there was a suddenly
evolved capacity for universal empathy or for a separate rationality untainted by
emotion or ego. Even less naturalistic are proposals about how humans must
utilize a spiritual free will uncaused by anything physical or must consult divine
revelations descended from above. No great leap is needed, however, if primate
behavior displays the building blocks of morality and if the differences between
hominid protomorality and human morality are mediated by gradual stages. Nor
would there be a forced choice between crediting nature or nurture for morality.
The long-standing debate over whether human morality is mostly the product of
164 JOHN R. SHOOK
natural evolution or cultural evolution presumes a dividing line during our hominid
ancestry where morality could be placed primarily on one side or the other.
Morality’s design is likely more complex than that nature-cultural debate admits.
An objective investigation of morality is needed in the course of pursuing
an evolutionary account of human morality. For the sciences to be able to study
human morality, it must have some place in the natural world, available for
observation and experimentation. Not surprisingly, the most obvious place to
investigate human morality is to examine the conduct of natural humans. By
regarding humans as entirely natural entities with biological properties and
cultural capacities, the sciences have plenty of morality within view for study.
Let “moral naturalism” stand for the view that there is a natural phenomenon
properly called morality that exists in human societies and that everything about
morality’s operations or evaluations is open to scienti fi c scrutiny. Moral naturalism
holds that human morality can be objectively studied by the several social and
natural sciences and that nothing about human morality must elude scienti fi c
understanding. This is not the same enterprise as using science to answer our
questions about what the morally right thing is to do. As a descriptive effort only,
moral naturalism is actually a quite modest and limited enterprise. “What is
morality?” is not the same question as “What is morally right?” There is a dis-
tinction between studying a practice and having expertise in that practice. For
example, we can observe how agriculture works without acquiring farming
expertise, and we can observe how people build all kinds of bridges without
ourselves having to answer questions about the best way to build a bridge.
Although often taken synonymously, it is useful to distinguish moral natu-
ralism from ethical naturalism. Morality refers to related kinds of actual human
conduct, while ethics refers to philosophical questions that arise about morality.
Ethics is the philosophical domain that deals with metalevel issues about ways for
determining the applicability of moral norms and terms, questions about the
appropriate methods for judging and answering moral problems, and concerns
over whether one correct morality can be determined. Adding the naturalistic
outlook to morality and then to ethics yields two quite distinct fi elds of study.
Moral naturalism takes humans to be doing morality naturally, without any
nonnatural features (such as a spiritual soul, free will, or pure reason) involved
in the process. Ethical naturalism goes much further than moral naturalism, by
regarding all metalevel issues about morality as issues to be resolved naturalisti-
cally, and typically includes views that moral rightness and personal goodness are
reducible to nonmoral features of nature, that things like moral values and moral
facts refer only to natural entities known by science, and that there are true pro-
positions of morality that are made objectively true by nature alone regardless
of any human cognition or consensus. Moral naturalism, as the term is used here,
is entirely unaffected by the validity or invalidity of ethical naturalism or moral
realism. Moral naturalism and ethical naturalism share a respect for science and
agree that science should take the lead when investigating morality. Indeed, deter-
mining the degree of ethical naturalism’s validity largely rests on fi rst carrying out
165
THE DESIGN OF MORALITY
the program of moral naturalism as far as it can go. No moral naturalist needs
to fi rst decide whether any part of ethical naturalism is valid. Indeed, moral
naturalism remains useful even if ethical naturalism is invalid. There are good
prospects for ethical naturalism all the same. Moral naturalism has made great
strides, and there are sound varieties of naturalism capable of grappling with
philosophical ethics (see, e.g., Darwall et al., 1997 ; Foot, 2001 ; Wong, 2006 ;
Flanagan et al., 2008 ) .
Regardless of ethical naturalism’s prospects, we are here only concerned
with moral naturalism. Ethical naturalism’s interest in discerning moral goodness,
value, and truth naturalistically is not shared by moral naturalism. For example,
only confusion arises if it is fi rst supposed that there are moral truths, or some-
thing like one “true” morality, and that moral naturalism should determine or
ground true morality. The project of moral naturalism may not lead to a victory
for ethical naturalism. Moral naturalism is compatible in theory with moral
relativism, for example – the many varieties of moral societies and their deep
disagreements indicate that there may be no way to determine a uniquely correct
morality. Still, moral naturalism need not imply any simplistic relativism, since
there may yet be better and worse moralities even if there is no perfect morality
(see Harris, 2010 ) . Moral naturalism may even reveal that even our “best” human
moralities are hardly as good as we think they are because much seemingly moral
conduct is actually motivated by sel fi sh concerns, either consciously or uncon-
sciously. All the same, the very fact that humans spend so much time deliberating
about good and better moralities, discerning whether each other’s conduct is
genuinely moral or not, and encouraging each other to be more moral proves that
morality is no illusion even if it cannot be so prevalent. Humans put a great deal
of effort into morality and, by extension, into ethical inquiry. Moral naturalism’s
ability to explain how humans now regard morality as variable and modi fi able
does show why ethical theorizing bothers to seek moral reform and agreement
about moral correctness. We envision possible better moralities because we
already understand what it is like to redesign inherited moralities.
Because societies do think about what the best morality may be, and whether
there ought to be one morality for all, and because societies frequently try to
impose one morality on everyone, moral naturalism also gets confused with
moral universalism. Despite widely shared hopes, moral naturalism is not a
method for determining one uniquely correct morality. Conversely, moral univer-
salism is a bad starting point for undertaking moral naturalism. Those in thrall of
one universal morality tend to denigrate as either immoral or nonmoral any
mode of human conduct that strays too far from the one correct morality,
whatever that may be. Universal moralizers are not alone in this inability to
recognize different moral systems. Since we all have been encultured into at least
one moral system, any of us can feel susceptible to this inability when we read
accounts of the moralities of societies which endorse such things as paternalism
or slavery. “That’s not any kind of morality,” we feel compelled to say, and we
may not be assuaged by reminders that those practices are considered moral by
166 JOHN R. SHOOK
those other societies. Relativism is not intuitive for morally trained people; those
other “immoral” societies equally judge our morals be immoral. But scienti fi c
objectivity is needed here. Moral naturalism takes its subject matter to be those
modes of conduct which peoples themselves have regarded as ful fi lling morality,
regardless of whether those morals are presently regarded as moral conduct by us.
Morals have and will forever vary widely across humanity, and moral naturalism
attempts to naturalistically account for all of them.
Moral naturalism limits its interest to the actual moral conduct of humans,
but this is a wide fi eld, since all human societies use morality. Unlike many ethical
naturalisms and some other ethical theories, moral naturalism does not limit the
meaning of “morality” to only some allegedly correct propositions of morality, as
if societies unaware of “the” correct morality therefore fail to have any morality
at all. Moral naturalism treats morality rather like food production or child
rearing: these are things which all human societies do, even if some societies may
do them somewhat differently or even better than others. While setting aside the
question of one “correct” morality, moral naturalism is primarily concerned with
morality as practiced by humans and hominid ancestors, and it uses a de fi nition of
morality as humanity now practices it (this de fi nition is coming in the next section).
While moral naturalism cannot begin from any single moral system, it must start
from a unitary notion of what “morality” is, in order to specify the subject matter
for inquiry. Similarly, any study of agriculture starts from a preliminary de fi nition
of what “agriculture” consists of, without also premising criteria for some “best”
kind of agriculture. It will not serve to leave open what “morality” might be, sup-
posing that empirical inquiry will settle the matter down the road. After all, moral
naturalism primarily aims at accounting for what humans currently are capable
of doing and not accounting for something we do not do. Although there are
researchers inquiring into other species from dolphins to wolves who claim that
those species have “moralities” of their own (de Waal,
1996, 2006 ; Bekoff and
Pierce, 2009 ) , their conclusions about nonhuman moralities are just that conclu-
sions about nonhuman moralities. Moral naturalism inquires into the homologous
features among human and hominid moralities, regarding them as sharing a
common ancestry and remaining grounded in anthropology. It can only recognize
near morality or components of human morality in other species. While it may
ultimately be a good idea to extend moral relativism to other social mammals and
cetaceans, moral naturalism aims at explaining morality as humans have come to
practice it.
2. Human Morality
A rough de fi nition of human morality would start from the way that morality is
practiced. Morality is naturally embodied in the ways that human individuals
voluntarily and habitually conduct themselves in accord with understood norms
regulating social interactions and related deeds of social concern, by not only
167
THE DESIGN OF MORALITY
regulating their own behavior but also by participating in the needed enforcements
of moral norms and by teaching these norms and the modes of enforcement to
those who need moral education. Morality is primarily designed to regulate social
relationships. Moral rules about what a person does in private have their ultimate
basis in what society deems as un fi t for relationships: disgusting, vulgar, disabling,
sacrilegious, or dishonorable deeds that offend society and render a person un fi t
for some social relationship. Enforcement and education are essential to morality
because morality is embodied in the voluntary habits of understood normative
conduct. The noticeable way that people frequently avoid or violate norms of
conduct reinforces this crucial point about morality: while morality must be to
some degree habitual, it must also be voluntary, and hence violable and irregular.
By “habitual” we do not mean instinctive, re fl exive, or robotic – there is nothing
necessary or fi xed about learned human habits. Indeed, precisely because acquired
habits inculcated by cultural training have only a modest way of guiding our con-
duct, so much enforcement and education is needed throughout life. On the other
side, “habitual” should not be taken to always mean rule following – people can
acquire habits by guided imitation and not memorization of express rules, and
people usually practice habits without re fl ecting on any rules governing their
habits. Some cultures may get around to expressing and teaching expected moral
habits and social roles with explicit rules, and many need not. Self-conscious rule
following is not essential to morality but only comprehension of what normally
ought to be done in a particular situation.
A morality, like any normative practice, is largely internalized. When people
are conducting themselves morally, they are following moral norms not because
they feel a sudden urge to be kind, or they are forced or coerced to do so, or
because they strategically regard obedience as simply a means to obtain what they
really want, but rather because habitual respect for moral norms and other people
are among their own important motivating values. Neither sympathy to needs,
obedience to commands, nor compliance with expectations is suf fi cient to consti-
tute morality (although those three factors can enhance moral conduct). A person
is not being moral by feeling motivated to help because a sudden discomfort
of sympathy or pity has been aroused. A person is not being moral simply by
obeying a command because it is backed by threats of punishment that she wants
to avoid. A person is not being moral by complying with a rule of conduct
because compliance is the best way to get what he wants anyways. A person is
behaving morally when they willingly conform to a moral norm because this
person’s respect for that norm is a suf fi cient and effective motivation to habitually
want to conform.
Moralists frequently demand that genuine morality must spring from a
purely altruistic desire to help another regardless of any estimation of duty.
Perhaps the idea that morality should be like the characteristic altruism of close
kinship has perpetuated this intense moralism. Finding so little angelic purity
across humanity, disappointed moralists are heard to complain that people have
little or no morality, but such cynicism arises from looking in the wrong place for
168 JOHN R. SHOOK
morality. Philosophical ethicists frequently demand that genuine morality occurs
when that person’s respect for moral obligations provides an overriding and
compelling reason for complying regardless of that person’s emotions, desires, or
values. But we are not undertaking philosophical ethics here, either. Moral natu-
ralism need not postulate anything like a detached rationality capable of dictating
conduct quite apart from emotions, desires, or values. This is fortunate, since
there may be no such detached thing in human psychology (see, e.g., Gazzaniga,
2005 ; Greene, 2008 ) . Nor is moral naturalism premised on moral realism or
on any claims that truth or rightness attaches to the moral judgments people
make. Moral naturalism is unaffected by the alleged prevalence of vast moral
error (as claimed by many such as Joyce, 2001 ; Lillehammer, 2003 ) , and morality
has not been “debunked,” so moral naturalism still has a subject matter. Whether
morality is actually what people suppose it to be is a concern for philosophical
ethics, not moral naturalism. Despite the fond dreams of some moralists or
ethicists, neither pious altruism, pure reason, nor perfect judgment has been
powering human morality, even if we may presently want to modify morality to
become more altruistic or rational.
Moral conduct, when it occurs, is primarily motivated by a perceived duty.
There may be other motivations to conform as well: nonconformity may bring
unwanted punishment; conformity may be a means to get what one really wants; or
conformity may assuage one’s uncomfortable emotions. However, when a person
suf fi ciently respects a norm, that person conforms even where nonconformity may
bring no punishment or personal bene fi ts, and even if no compassionate emotions
are dominant. There are degrees of respect for moral norms, and enforcement and
education are common means to increase respect. Enforcement and education rely
on the deeper morality-building motivations of caring for others, avoiding harms
from others, or getting bene fi ts from others, but encultured morality results in
motivating habits not reducible to any combination of these more basic and
evolutionarily older motivators. Because human societies can promulgate a wide
variety of moral habits, we may say that in a sense, morality is socially conven-
tional, but only in one sense. While morality persists in human culture largely
because humans do the training, that does not mean that humans must regard their
moralities as conventional. Indeed, many societies teach that their own morality
is the only morality, and some additionally teach that their morality is grounded
on nonhuman matters like nature’s ways or a god’s wishes. A well-trained moral
individual is not likely to regard moral duty as based merely on what society
wants – the moral individual is more likely to regard what is morally right as
enjoying a foundation independent from humanity. Similarly, although the
practices of mathematics persist because human societies promulgate them,
mathematics practitioners are unlikely to regard the truths of mathematics as
dependent on what society happens to promulgate. Socially designed morality is
functioning well for people when they do not regard it merely as locally and
conventionally designed. We must not make too much of this looming paradox
when we are pursuing moral naturalism. It is a peculiar problem for ethics, and
169
THE DESIGN OF MORALITY
not moral naturalism, to help reconcile the seeming independence of general
moral duty with its actual dependence on local moral education.
Morality is designed to function somewhat differently from other socially
normative practices. In human societies, morality can be distinguished from two
similar social practices concerning relationships that extend beyond the range of
the family: the norms of political laws on the one side and the norms of etiquette
on the other. Political laws backed by government force can secure widespread
and uniform obedience from the people, but entirely voluntary respect for those
laws may be weak or far from universal. Laws backed by effective threats do their
proper work of preventing harms and violence by appealing to an individual’s
basic desires to avoid public shame and harsh punishment. Civil etiquette is
commonly quite voluntary, but it can vary so widely among individuals in the
same society, and so irregularly enforced by sanctions, that speci fi c norms
accepted by all cannot be clearly identi fi ed. Nevertheless, norms of etiquette do
their proper work of promoting aid and altruism by appealing to an individual’s
basic capacities for feeling sympathy and compassion toward others. Research
on primates indicates that a protomoral sense of compassionate altruism and
respectful fairness can occasionally be observed in their social behavior, appro-
ximating what is labeled here as “etiquette” (not in the sense of re fi ned manners,
of course, but just the simple gestures of nice and fair treatment).
Morality is in the middle ground between law and etiquette and overlaps
them on each side – it concerns speci fi able norms about social relationships
and interactions for the entire society to willingly conform and mutually enforce.
In every human society, such norms are evident and powerful, although their
speci fi c scope and content varies widely across cultures. A type of social interac-
tion dictated by morality in one culture may be left up to etiquette in another;
what is assigned to law to enforce in one culture may be left up to morality in
another; and some societies may not regulate some kinds of conduct which other
societies heavily regulate using morality and law. However, morality has a distinc-
tive role in every society: it concerns those norms in a society where that society
regards them as both universally applicable and universally worthy of suf fi cient
respect. Society demonstrates its regard for morality by expecting voluntary con-
formity to its norms, expecting people to help enforce conformity where needed,
and expecting people to help instruct the young to acquire moral norms. Where
society encounters individuals unable or unwilling to conform morally, even if
they conform obediently or self-servingly, that society expends efforts to reform
that person’s attitudes and habits, or failing in that effort, reclassi fi es them into
some subnormal status (e.g., with unreformed criminals, the mentally ill, or the
cognitively disabled).
Just as morality, while having universal applicability within a society, can vary
in scope and content from society to society, morality can vary in effectiveness.
If it were somehow impossible for morality to vary in its effective impact on the
lives of society’s members, they would not even think to try to modify it. Since
morality obviously can and does have varying noticeable impacts on different
170 JOHN R. SHOOK
individuals, so where there is suf fi cient intelligence, morality can therefore become
an object of interest and an object of manipulation.
The opportunities for deliberate modi fi cation of morality are everywhere,
since individuals can acquire some intelligent control over their conduct, even
much of their habitual conduct, if they can consider their conduct as something
controllable and modi fi able over time. The story of the evolution of culture is
essentially the story of the increased capacity for humans to regard some of their
habitual conduct as modi fi able with attention and practice and as teachable
through instruction. The story of the evolution of morality, as a mode of encul-
turation, is essentially the story of the increased capacity for hominids to regard
and enforce some social norms as worthy of everyone’s willing conformity. As
objects of intelligent attention in their own right, humans then gradually came
to regard such moral norms as deliberately modi fi able and proceeded to experi-
mentally redesign the many moralities now embodied in diverse human cultures.
By analogy, early hominids developed habitual tool use over two million years
ago, but few modi fi cations to choppers, fl akes, and blades occurred until brain
size had dramatically increased; the immense proliferation and complexity of
tools associated with Homo sapiens indicate how hominids and humans gradually
took deliberate control over experimental tool construction.
3. The Function of Morality
Humans would not experimentally redesign morality unless it came to be viewed
as something not just modi fi able but as modi fi able for serving some end. An
experimental modi fi cation to something, as opposed to an accidental, sporting, or
aesthetic modi fi cation, treats it as a means for achieving envisioned consequences.
What is it like to regard morality as a means functioning to serve ends? What
would be the function of morality? We have located morality among the modes of
human conduct, where individuals willingly regulate their social interactions out
of respect for its norms and they expect everyone to do likewise. This descriptive
view of morality omits its functionality: why would humans have morality?
Couldn’t human societies do well with just familial altruism and social etiquette,
or some combination of familial love, etiquette, and law?
Morality supplies something that neither familial love, etiquette, nor law
can provide. Morality permits standardized modes of interactions which each
individual can reliably expect from all others under conditions when kinship is
absent, etiquette is doubtful, and punishment is uncertain, unwise, or too costly.
Etiquette is not standardized and not evenly enforced; indeed, norms of etiquette
are precisely those optional norms which lack overriding respect and little punish-
ment if any is attached to their enforcement. (That is why reciprocal altruism
cannot be the essence of morality but only a display of optional etiquette – see
Tullberg 2004 on differences between altruism and reciprocity.) The most impor-
tant norms of etiquette overlap with the minor norms of morality, and in modern
171
THE DESIGN OF MORALITY
human societies that range of normative conduct makes up what we call civility.
On the other side, where a society uses law, the most important codes of morality
overlap with much of law, and that range of normative conduct presently makes
up what can be called good citizenship. Law promises standardized modes of
interactions and high probabilities of punishment, but enforcing law has many
costs because it is very intrusive on individuals and requires large resources for
adjudication and punishment; that is why societies which use laws are regulating
only the most important kinds of social interactions. Morality serves to regulate
conduct across a broad array of human interactions where norms must be both
universally respected and ef fi ciently enforced.
What sorts of social interactions would bene fi t from something like moral-
ity? The obvious kinds of interactions are cooperations. Moral norms, because
they regulate everyone’s conduct in an ef fi ciently uniform manner, are highly useful
for promoting cooperation among all members of society. Where the more basic
components of morality are already in place, especially the norms of etiquette
encouraging mutual assistance and fair treatment, stable patterns of mutually
bene fi cial cooperation can emerge and grow. If individuals can be con fi dent that
mutual assistance will reliably yield suf fi cient bene fi ts, without worry that unfair
treatment might occur, they will naturally undertake cooperative projects with
some frequency. In short, the original effects of occasional friendly etiquette
can magnify into the repetitive trusting cooperation that can deliver even higher
bene fi ts for individuals in intensely social societies. Indeed, a highly social society
is precisely that society in which both simple etiquette and complex cooperation
are regularly occurring to the high bene fi t of all members across lifetimes and
over generations.
The mutual helpfulness and fair cooperation permitted by basic etiquette
can yet remain unstable and less frequent compared to other strategic modes of
social interaction going on within a society. To become the dominant mode of
social interaction, robust cooperation across societies requires more than just
etiquette. The additional assistance comes from morality’s universal and stricter
obligations. The deliberate invention and design of law was a further extension of
this same process in highly complex human societies, increasing the bene fi ts
of cooperative social environments by applying more costly regulation to the
most important behaviors threatening the proper functioning of a civil society.
The continuities between stages of moral development, proposed at the outset of
our expectations for moral naturalism, can be observed in theory and fact.
If etiquette had utility for small groups, morality would evolve that utility to larger
groups, and law extends that normative development to fully complex societies,
making higher civilization possible. A society concurrently utilizing all three
modes of etiquette, morality, and law would then maximize ef fi cient continuity:
making everything a matter of law would be too costly, as would folding all of
etiquette into morality. It may be theorized that an optimal society, therefore,
would attempt to match its expenditures for normative implementation to the
social signi fi cance of the behavior regulated. We do observe many modern human
172 JOHN R. SHOOK
societies displaying such effort to ef fi ciently utilize etiquette, morality, and law
concurrently, although societies design their own distinctive modes of assigning
expected conduct to each category.
4. Morality and Cooperation
Intuitively, cooperation and morality are a good functional match. However, we
must not hastily assume that morality only exists where intense cooperation is
ongoing, and we must not assume that cooperation requires morality. Species can
evolve intense forms of social cooperation without any morality or even social
intelligence (take ants for example) because such cooperation can be sustained by
close kinship alone. For its part, morality could theoretically serve other functions
besides cooperation. There are competitive “zero-sum” games of winners and
losers, what can be called “win-lose” games, that can be better sustained if some
moral norms are added to the rules of the game. However, the impressive value of
morality is revealed when it is added to cooperative “win-win” or “non-zero-sum”
games of mutual bene fi t. Indeed, it is hard to imagine how any sustained forms of
intense cooperation across large societies would last without support from moral
norms. All the same, the question stands: could non-zero-sum cooperation be
worth it, relatively speaking? Numerous studies of forms of reciprocity generally
suggest that this is the case (see, e.g., Trivers,
1971, 1985 ; Hirshleifer and Martinez
Coll, 1988 ; Sober and Wilson, 1998 ; Henrich et al., 2001 ; Sachs et al., 2004 ; Taylor
and Nowak, 2007 ) .
It is widely agreed that, in theory, non-zero-sum cooperation tends to delivers
greater overall bene fi ts to most individuals over time in a complex society than
any combination of competitive zero-sum games, provided that individuals have
cognitive skills suf fi cient for conducting such social cooperation. A resulting
overall advantage of cooperation over competition would practically explain why
larger groups of increasingly intelligent hominids and humans evolved to have
greater reliance on non-zero-sum cooperation than all other zero-sum interactions
combined. Cooperation for common and mutually bene fi cial ends is a pervasive
and important feature of any human society; indeed, across several hominid
species, social cooperation largely determines the welfare of any of its members.
For modern humans especially, avoiding social cooperation is the path to death;
even a “self-suf fi cient” hermit or a combative aggressor was taught survival and
fi ghting skills by others.
Where does the role of morality enter? The modest amount of social
cooperation that gradually arose in early hominid species would have bene fi tted
greatly from simple moral norms, over and above the contributions to cooperation
made by the sentiments of sympathy, compassion, and altruism. Those crucial
sentiments secure the bonds of commonality among closely related kin and a few
caring friends, but they do not function well for impartial extension to everyone.
The value of morality only increased as Homo sapiens gradually lived in larger
173
THE DESIGN OF MORALITY
tribal units and invented more complex forms of social cooperation occurring
across the boundaries of close family and friends. Where the emotions guiding
family commonality are absent, a different kind of cooperative relationship with
acquaintances can develop instead, one based on reciprocity. Given the above
sketch of what morality is and how it basically functions, we can understand
why morality is well-suited to facilitating widespread cooperation for high mutual
bene fi t. The universality and overriding respect for moral norms is a good func-
tional match with the widespread and repetitive modes of cooperation going
on within social groups. Reliance on etiquette does not vanish as reciprocal
cooperation increases, but its unsteady and varying force would enjoy a dramatic
supplementation with morality. In fact, etiquette and morality would reenforce
each other among the most cooperative members of a society; that is why distin-
guishing merely compassionate acts from genuinely moral conduct is no simple
matter. (Similarly, when modern humans invented political law, it hardly replaced
etiquette and morality, but supplemented them in an integrated fashion, so that
refraining from murder is simultaneously legal, moral, and nice.)
Where any two members of a group can both have some assurance that
signi fi cant mutual bene fi t is possible and no unfair harm is forthcoming, they
have a greater rational incentive to engage in reciprocity cooperation. Core moral
norms helpful for such cooperation would naturally include prescriptions against
coercion or harms by physical domination, deception, cheating, stealing, and unfair
treatment. Respect for morality provides needed assurances for all parties, and the
value of that assurance increases with the size of the social group. In very small
groups where everyone is closely related and quite familiar to each other, instinc-
tive emotions of sympathy and compassion could be strong enough to ensure care
and cooperation, so familial love and kind etiquette is usually suf fi ciently and no
morality is needed. But in larger groups, where individuals will encounter distant
relations (or nonfamily as well), morality adds its distinctive service to increasing
cooperation over and above the bonds of family and vagaries of etiquette. In such
societies, where individuals could bene fi t from cooperation with others known pri-
marily by reputation, moral norms increase the chances of successful cooperation.
In effect, morality would be displayed in situations where two or more individuals
know each other primarily by reputation, they have repeated opportunities to
mutually bene fi t from cooperation, and they conform to moral norms about
cooperating fairly and not harming each other in the process.
This view of morality’s functioning presupposes that individuals have
enough cognitive resources to absorb and understand prevailing moral norms in
their society, to know how to strategically comply or ignore those moral norms,
and to judge the other members’ reputations for complying with those norms as
well. There is no well-established theory about when early hominids developed all
of these cognitive resources and began doing what we know as morality. Perhaps
our primate cousins are capable of occasionally performing very simple versions
of morality. But it may prove dif fi cult to determine whether their conformity to
norms is due to sympathetic feelings, strategic aims, or actual respect for the
174 JOHN R. SHOOK
norms themselves (just as judging motives of fellow humans is not easy). In any
case, as hominids gradually came to rely on social strategies demanding intense
daily cooperation, it is reasonable to suppose that the basic primate toolkit for
protomorality underwent development too. That development led to us: we can
observe how the intelligence of Homo sapiens is highly developed for tracking the
conduct and reputations of many individuals and for enforcing and teaching
morality right along with the other aspects of complex culture.
When morality is supplemented into the manner in which individuals
cooperate, this moral cooperation is dramatically enhanced in two ways: they will
cooperate more frequently, and their cooperation will be much more ef fi cient.
Where cooperation is conducted morally, the participants will generally display a
high level of cooperativeness with most if not all other members of the society.
Speci fi cally, each member will tend to be willing to cooperate when there is an
opportunity, and each member will take an opportunity to cooperate with most
or all other members of the society as situations warrant. In simplistic terms,
where morality is robust, we would expect to observe lots of niceness and very
little choosiness. On an emotional level, robust morality is experienced as trust: a
felt con fi dence that reciprocity interactions with another will be reliably safe,
bene fi cial, and fair.
We should not overrationalize the gradual emergence of a rational morality
within social groups. It is unnecessary to depict hominids as cold calculators
thoughtfully rationalizing cooperation. The cognitive capacities needed for
morality’s emergence would not need to be more sophisticated than the “fast and
frugal heuristics” of severely bounded rationality (see Gigerenzer and Selten,
2001 ) . Like so many other evolved features of hominid brains, the evolving emo-
tions do most of the work of social intelligence anyways, and humans have inher-
ited and use them (see Haidt, 2003 ; Nichols, 2004 ) . Individuals growing up in a
relatively cooperative society will be guided by the feelings of enjoying mutual
trust. Nor is necessary to depict early hominids or Homo sapiens as utilitarians
who regard morality as a means to enhancing social welfare. Actually, individuals
committed to morality will instead generally regard morality as binding regardless
of calculations of social welfare. On a personal level, a moral person will try to
be virtuous on a daily basis, and a group of moral people will regard each other
as dutiful agents. Perhaps only when a society’s morality is under widespread
dispute or under discussion as a problem in itself would people think to inspect
morality in regard to its service to society. Whether virtue theory, deontology, or
utilitarianism has greater merits is an issue for philosophical ethics, not moral
naturalism. For its part, moral naturalism can at least account for the distinctive
ways that morally trained people would regard how morality works. Moral
psychology has the unenviable task of sorting out the actual motivators for
moral judgment from the stated justi fi cations that people will supply when asked
(see, e.g., Doris et al., 2010 ; Brinkmann, 2011 ) .
So far, a naturalistic explanation involving both natural and cultural evolution
for morality’s enhancement to routine and widespread cooperation seems possible.
175
THE DESIGN OF MORALITY
However, as already noted, there are many other strategies for conducting social
interactions besides morality. Even if we might see how morality can enhance
cooperation, individuals can still bene fi t from competitive, neglectful, unfair, or
harmful conduct toward others too. Why would morality come to have any large
role to play in hominid societies? Perhaps morality has long been just one optional
mode of social conduct among many others, and individuals could vary widely in
the extent to which they rely on morality over other strategies. Perhaps a society
that does use much morality could be successfully “in fi ltrated” by individuals
using more sel fi sh strategies. Is there good reason to conclude that individuals in
societies using lots of moral cooperation enjoy enough bene fi ts to survive or to
conclude that societies using moral cooperation might just as well drift away from
morality as time goes by?
There is no point to accounting for the widespread reliance on morality
among contemporary humans with an evolutionary story of hominid development
unless moral cooperation can work well, at least better than zero-sum interac-
tions. It is unnecessary to think that the proliferation and supremacy of nonzero
cooperation was in some sense evolutionarily inevitable (as suggested in Wright,
2000 ) , but there are good reasons to think that plenty of non-zero-sum cooperations
would be stumbled upon and repeated by suf fi ciently intelligent species, such as
our primate and hominid ancestors, giving rise to culture (a survey of reasons is
given by Boyd and Richerson, 2005 ; Nowak, 2011 ) . Are there good reasons to
think that moral cooperation delivers generally high bene fi ts to the members of a
society and delivers results more bene fi cial to all individuals within that society
than any other strategy? Further study is needed to analyze the bene fi ts of moral
cooperation in a society.
A commonly used tool for analyzing the supposed bene fi ts of reciprocity
cooperation is the prisoner’s dilemma game. Computer modeling of groups uti-
lizing various strategies shows that a small number of simple strategies, such as
“Tit For Tat,” tend to garner the most bene fi ts (though there is no provably optimal
strategy for inde fi nitely extended iterations). However, the prisoner’s dilemma
game models a very speci fi c kind of basic social interaction, which is too simplistic
to bring morality into full view. What is needed is a revised type of prisoner’s
dilemma game that models the more complex kind of social interaction where
emerging morality would be occurring.
5. The “Indirect Reciprocity” Prisoner’s Dilemma Game
The prisoner’s dilemma is a way to model simple interactions between individual
agents where there is the possibility of reciprocity, cooperation, and betrayal. This
section describes a computer program which models a tournament of players
taking part in what I have termed the “indirect reciprocity” version of the prisoner’s
dilemma game, or the IRPD game. I fi rst presented the IRPD game at the Seventh
Annual Conference on Computers and Philosophy at the University of Central
176 JOHN R. SHOOK
Florida in August 1992, and a more sophisticated version was presented at
Oklahoma State University in 2003. The basic inspiration for this project comes
from Robert Axelrod’s ( 1984 ) ground-breaking exploration of the prisoner’s
dilemma. The study of the prisoner’s dilemma has been widely perceived as an
opportunity to gain insights into the nature and origins of morality. A full expla-
nation of the prisoner’s dilemma and its importance is found in Axelrod’s book
and subsequent books on its analysis (see, e.g., Colman, 1995 ; Barash, 2003 ) .
The essential point is that while the prisoner’s dilemma (PD) is a single inter-
action, the prisoner’s dilemma game is a long series of such dilemma interactions,
where neither player knows when the series will end. A strategy is a method of
playing this PD game. The PD by itself has one optimal strategy regardless of the
opponent’s strategy: defect. The PD game by contrast need not have one optimal
strategy, for reasons involving the uncertainty regarding when the last interaction
will occur and, if groups of players are interacting, the variability of the strategies
involved. The study of the PD game primarily involves the examination of the
performance of the possible strategies available to the players, relative to the size
and characteristics of the rest of the group.
Axelrod studied how various strategies, embodied in computer programs,
performed as they competed in round-robin tournaments. His use of the PD
game involves players in a round-robin tournament who use strategies to try to
obtain the most points from PD games with the other players. The crucial aspect
to Axelrod’s methodology is that the strategies permitted by Axelrod can only
exhibit direct reciprocity. This direct reciprocity is characterized by the informa-
tion upon which a player has to base the decision whether to cooperate or defect in
a dilemma: the only information available to a player is the history of interactions
he has had with the other player. This explains the results of Axelrod’s tournament;
the winning player’s strategy, Tit For Tat, is not nearly as sophisticated as most
of the forms which moral behavior can take. Is it possible that Tit For Tat is a
morality? As far as familiar codes of morality go, it does resemble the ancient
“eye for an eye” rule of retribution: what you have done to me, I will do to you.
The important question is whether the methodology of using the PD game can
be altered so as to permit more complicated forms of behavior. The answer lies in
understanding reciprocity.
There are two kinds of reciprocity: direct and indirect. Direct reciprocity is
of the form: A helps B, B helps A. Indirect reciprocity is of the form: A helps B,
B helps C, C helps A; or A helps B, C who is observing, later helps A, A helps C.
Alexander ( 1987 ) and Boyd and Richerson ( 1989 ) have emphasized how the more
complex behaviors associated with morality are mostly dependent upon the ability
to use strategies that exhibit indirect reciprocity instead. Indirect reciprocity
requires an additional piece of information: the history of the opponent’s interac-
tions with the rest of the players in the tournament. After a discussion of direct
and indirect reciprocity, the basic scheme of the “indirect reciprocity” prisoner’s
dilemma game computer program is laid out, and preliminary results from this
program are described. The IRPD game shows how genuine moral cooperation
177
THE DESIGN OF MORALITY
can be highly bene fi cial and evolutionarily stable. The following brief account
of the IRPD computer program describes its capacity to formulate strategies,
group various strategies together, and have them interact in tournaments in a way
similar to Axelrod’s. The key departure from Axelrod’s work lies in the difference
between direct and indirect reciprocity.
In the PD game, help comes in the form of cooperation. Direct reciprocity
only requires a player to have access to information on an individual level: a record
of how the other player has treated one in past dilemmas. The modeling of indirect
reciprocity requires additional information: knowledge of the other player’s
interactions with the rest of the group. Additionally, the modeling must permit a
player to make a choice as to whether she will engage another player in a prisoner’s
dilemma interaction. No such choice is allowed in the direct reciprocity model.
For our purposes designing a basic IRPD game, we shall assume that every
player has complete and reliable information; other models of the IRPD game
need not do so. We shall also assume that no player will err in making decisions
based upon the information; other models may include the possibility of judgment
error. On this foundation, we can proceed to construct a prisoner’s dilemma
model of indirect reciprocity. What follows is but one way to model indirect
reciprocity, using discrete, agent-based modeling that tracks the performance of
identi fi able agents as they interact or avoid each other over time. Aggregative
modeling of populations undergoing interactions involving indirect reciprocity
has already been attempted (see, e.g., Nowak and Sigmund, 1998 ; Panchanathan
and Boyd, 2004 ) . However, aggregative population dynamics assumes that every
individual has an equally likely chance of interacting with every other individual.
Because indirect reciprocity permits an individual to abstain from an interaction,
as well as to engage in an interaction, full indirect reciprocity is imperfectly modeled
by aggregative modeling (see Hauert et al., 2008 for discussion of this point).
Indirect reciprocity also works best for fi nite populations in which members
possess some information about everyone’s reputation for cooperating or not.
Without an estimate of reputation, indirect reciprocity gains little for cooperators,
while much knowledge of reputations can produce stable cooperative societies
(Milinski et al., 2002 ; Nowak and Sigmund, 2005 ) .
The IRPD game described here is probably one of the simplest models for
agent-based modeling of strategies for undertaking indirect reciprocity with
all other social members in a fi nite population that include the basic features of
reputation, avoidance, cooperativeness, niceness, and choosiness. We shall fi rst
look at the structure of a strategy which can enable a player to engage in indirect
reciprocity. A look at the structure of the round-robin tournament follows.
As we know, indirect reciprocity requires that each player should have the
option of whether to have a PD interaction with another player. The judgment
that entering into a dilemma with a particular opponent is not in a player’s best
interests will result in a “shun”; the player will opt out of a PD interaction with
the other. (In the basic kind of IRPD game described here, a “shun” is the only
mode of “punishment” exacted by a strategy.) In the IRPD game, this judgment
178 JOHN R. SHOOK
will be made upon the information of how often the other player cooperates
in all the games he has played. When a player has decided to enter into a PD
interaction, the strategy must direct the move to be made. Accordingly, a strategy
will require access to three kinds of information: (a) how often the player should
cooperate, (b) how often the opponent has cooperated in all of its interactions
with all other players in the past, and (c) what standard the player should use to
evaluate other strategies.
The fi rst kind of information is placed in the form of a number ranging from
0 to 1, representing what the frequency of the player’s choice to cooperate in an
interaction will be. 1 = always cooperates, 0 = always defects, .5 = there is a 50%
chance of cooperation, etc. The name for this number is niceness. For the purpose
of modeling agents with limited cognitive and deliberative resources, we may
regard a player’s niceness more as a habit rather than a thoughtful choice upon
each interaction.
The second kind of information shall be just the opponent’s niceness. This
represents each tournament as taking place in the course of the life of a society,
after each player has established a reputation among the group. Again, for the
purpose of modeling agents with limited cognitive and deliberative resources,
we may regard a player’s understanding of an opponent’s niceness not as the
recollection of all the outcomes of the opponent’s past interactions, or the recol-
lection of what others have gossiped (no presumption of advanced language use
is made here), but rather just as an understanding of the opponent’s reputation
as far as a player can tell from some observation.
The third kind of information shall again be placed in the form of a number
ranging from 0 to 1, representing the standard used to evaluate the other player’s
niceness in order to make the decision upon whether to interact in a dilemma with
the other player (1 = will only choose to play will those having niceness of 1, 0 = will
play with anybody, .5 = will play with only those having a niceness of .5 or higher,
etc.). The name for this number is choosiness, and again this represents only a
habitual inclination of the player and not a series of re fl ective deliberations.
Since a player’s niceness and the choosiness do not change for a strategy
during a tournament, we can simply refer to these two qualities as a strategy. In
this way, we can say that the entire possible range of strategies can be represented
by the points on a one unit by one unit graph.
To get an IRPD tournament started, the decision as to which strategies shall
play in a tournament must fi rst be made. The program I wrote permitted the user
to form the group either by setting each strategy’s niceness and choosiness, or to
choose a preset group by selecting a size equal to a perfect square (4, 9, 16, 25, etc.)
The preset groups have an even distribution of qualities: with four members, the
strategies are 1,1 1,0 0,1 0,0. With nine members, the preset group looks like: 1,1
1,.5 1,0 .5,1 .5,.5 .5,0 0,1 0,.5, 0,0. With groups of 400 and more, the tournament
can be very inclusive. Each of the preset groups will have an average niceness
and choosiness of .500. Other kinds of preset groups are possible as well, with
differing distributions, but await a future version of my program.
179
THE DESIGN OF MORALITY
A round-robin tournament for the group is scheduled as follows. In one
“round,” a designated strategy will meet (have the opportunity to interact) each of
the other strategies consecutively. In the span of one “season,” a number of rounds
take place, equal to the number of the strategies in the group, so that each of the
strategies will take the role of the designated strategy in one season. Each strategy
will thus meet every other strategy twice in one season. As the computer program
proceeds through the schedule, for each scheduled meeting allows the two strate-
gies to fi rst decide if the PD interaction should take place. If either player declines
because the opponent’s niceness is not great enough (does not equal or exceed the
player’s choosiness), no PD interaction takes place and the programs proceed to
the next scheduled meeting. If both do choose to interact, then each strategy
decides whether to cooperate or defect in the PD interaction. This decision will be
based on the strategy’s niceness, by fi nding whether a cooperation would bring its
overall record (ratio of cooperations to total games played) closer to its niceness
or not. After each has chosen, the program determines the outcome of the game
by assigning points accordingly (e.g., in the standard assignment, 3 points go to
each if they both cooperated, 1 point goes to each if they both defected, or 5
points goes to the defector and none to the cooperator; the points awarded can be
preset by the user). The computer program saves a database of each player’s cumu-
lative points, for reporting at any selected stage of a tournament.
To summarize the program:
1. The number of players and their assigned strategies are chosen by the user.
2. The tournament length (the number of seasons to play) is chosen by the user.
3. The ordering of the group is scrambled.
4. In each season, the necessary number of rounds occurs.
A. Inw each round, the meetings are assigned.
B. For each scheduled meeting:
a. The program determines if the two decide to interact.
b. If they will interact, it fi nds each strategy’s chosen move.
c. The outcome of the game is found and points are awarded.
C. At the end of the season, the ordering of the group is rescrambled.
5. Each strategy’s cumulative points are recorded for later reporting.
There are some ways to complicate the IRPD game so that the player’s
environment and playing conditions more closely approximate the “real world.”
In the real world, organisms have to “pay” the costs of living from day to day,
they will someday die from natural causes, they must reproduce to continue their
genetic line, their offspring can have mutated genetic codes, and so forth. In the
game, a preset amount of points can be subtracted from the player’s total at each
scheduled meeting (this is called the “meeting cost”), so that death could occur
(total points dropping to zero or below). When a meeting cost is exacted, all strate-
gies begin with an allowance of points equal to 50 times the meeting cost. This
amount is not fi gured into the fi nal results and succeeds in preventing a strategy’s
180 JOHN R. SHOOK
chance bad luck in meeting assignments in the early stages from prematurely
killing it. In effect, this tries to compensate for the mere pseudorandomness of the
computer and simulates the natural fact that typically organisms are not born
(nor enter adulthood) on the very edge of death; they have some (however small)
stored resources.
The players who are able to accumulate points at a faster rate will tend to
survive, and then the effects on a player of competing in an evolving group can
be studied. The players can be given a preset lifespan, in terms of a maximum
number of scheduled meetings, after which “death” occurs. In the program, the
timing of death is not the same for all: the probability of death for a strategy
begins some meetings before the preset lifespan and steadily increases each meeting
afterward, so that the actual death can occur in a range around the lifespan.
In more sophisticated versions of the computer program, players can use
points to have offspring and use up points just to keep living. To “give birth,” a
player can be permitted to create a clone of itself after it has accumulated a preset
amount of points. The new player has the identical strategy as its “parent” and is
entered into the tournament, and some amount of points are subtracted from the
parent’s total. However, giving birth will not unduly harm a strategy. It is only set
back to its original amount of points and keeps its margin of safety from lack of
randomness if there is a meeting cost in a tournament. The result is that a more
successful player will have a higher birthrate. Also, at the cloning, a mutation can
be permitted, to change a new player’s strategy so that it differs from the parent’s.
In addition, the meeting cost can be variabilized according to a change in the size
of the group, so that meeting cost rises proportionately to the group size if the
group size should rise through births. This models a realistic aspect of a group
living in an environment with limited resources, where a rise in population makes
it harder for all to survive.
When these complications are put together, natural evolution is modeled.
The fi rst generation of strategies can compete, give “birth,” and die, giving rise to
the second generation, which in turn repeats the cycle. In this way, the evolution-
ary success of a strategy can be measured. Also, the composition of a group of
strategies can be evaluated over many generations for survival (does the group
grow, or does it die off) and for total overall success (the cumulative amounts of
points of all living strategies).
6. The IRPD Game Results
Some basic results have been con fi rmed through many hundreds of tournament
experiments, and they are consistent and very striking. When there is no meeting
cost exacted, the winner always has a very high niceness and a very low choosiness.
In fact, the ranking of the strategies by points earned consistently follows the
degree to which a strategy’s niceness approaches 1 and choosiness approaches 0.
The worst fi nishers are quite nasty and very choosy. For example, the strategy (1, 0)
181
THE DESIGN OF MORALITY
nearly always fi nishes in the top 5%, regardless of the size of the preset group, and
the strategy (0, 1) fi nishes in the bottom 5%. I have been able to con fi rm this for
groups containing up to several thousand strategies. Figures 1 and 2 show arrays
of the points accumulated by strategies in groups of 100 and 224 who have played
typical tournaments lasting 50 seasons.
If there is a meeting cost exacted from all strategies, four major phenomena
occur. First, some strategies will die at some point in the tournament because they
fail to accumulate points fast enough to cover the subtracted meeting costs.
Without exception, the nastier and choosier strategies die fi rst, so the meeting
cost results in a gradual rise in the average Niceness and a drop in the average
choosiness until no more strategies die. This should be expected, given knowledge
of the results of tournaments lacking a meeting cost in which the nastier and
choosier strategies garnered fewer points (gained points at a slower rate) than the
rest. In tournaments with a meeting cost, these strategies’ points would drop
closer to zero, and those whose rate of points per meeting remained less that the
meeting cost rate eventually die. The second phenomenon is observed when
the meeting cost is varied. There is a direct relationship between the size of the
meeting cost and the group’s average niceness; an inverse relationship holds
between the size of the meeting cost and the group’s average choosiness. For
example, in a tournament in which the meeting cost is set at .5 point per meeting,
the average niceness increases to around .60 and the average choosiness drops to
Niceness
legend:
500.0
Choosiness
Summary for 100 strategies and 50 seasons
0.0 0.2 0.4 0.6 0.8 1.0
0.0 0.2 0.4 0.6 0.8 1.0
4096.5 7693.0 11289.5 14886.0
Figure 1. One hundred strateg ies after 50 seasons. The worst strategies accumulate less than 4,000
points.
182 JOHN R. SHOOK
around .40. If the meeting cost is 1.0 point per meeting, the average niceness
is around .66 and the average choosiness is around .34. Figures 3 and 4 show
an array of 100 strategies playing tournaments of 500 and 1,000 seasons with a
meeting cost of .5 point.
If the meeting cost is 1.0 point per meeting, the average niceness rises to
around .66 and the average choosiness drops further to around .34. Figures 5 and 6
show an array of 100 strategies playing tournaments of 500 and 1,000 seasons
with a meeting cost of 1.0 point. Keep in mind that these are summaries of many
independent runs of tournaments so they will display inconsistencies among them.
For example, Fig. 5 shows how the strategy (.5, .1) had died in its tournament of
500 seasons, while Fig. 6 shows how the same strategy managed to survive to the
1,000 season mark in another tournament.
This variable meeting cost trend toward higher niceness and lower choosiness
does not seem to persist inde fi nitely, however, as very high meeting costs stabilize
the niceness and choosiness of the few surviving strategies in the vicinity of .50
and .40. This experimental model result conveniently corresponds to the familiar
resurgence of uncooperative sel fi shness among primates and humans under very
stressful conditions.
The third phenomenon is also seen when the meeting cost is varied: the
leaders in these tournaments are not as nice nor as indifferent as those from tour-
naments lacking a meeting cost. As the meeting cost rises and the group becomes
Summary for 225 strategies and 50 seasons
Niceness
Choosiness
0.0 0.2 0.4 0.6 0.8 1.0
0.0 0.2 0.4 0.6 0.8 1.0
legend:
500.0 8837.75 17175.50 25513.25 33851.00
Figure 2. Two hundred and twenty- fi ve players and 50 seasons. No matter the number of strategies,
the same strategies perform the worst.
183
THE DESIGN OF MORALITY
Choosiness
0.0 0.2 0.4 0.6 0.8 1.0
Niceness
Summary for 100 strategies and 500 seasons
0.0 0.2 0.4 0.6 0.8 1.0
legend:
14639.50 39114.38 63589.25 88064.12 112539.00
Figure 3. One hundred strategies after 500 seasons and .5 meeting cost. The white spaces indicate the
dead strategies.
Choosiness
0.0 0.2 0.4 0.6 0.8 1.0
Niceness
Summary for 100 strategies and 1000 seasons
0.0 0.2 0.4 0.6 0.8 1.0
legend:
35913.00 82966.25 130019.50 177072.75 224126.00
Figure 4. One hundred strategies after 1,000 seasons and .5 meeting cost. The nicest strategies
continue to perform well.
184 JOHN R. SHOOK
Choosiness
0.0 0.2 0.4 0.6 0.8 1.0
Niceness
Summary for 100 strategies and 500 seasons
0.0 0.2 0.4 0.6 0.8 1.0
legend:
14639.50 39114.38 63589.25 88064.12 112539.00
Figure 5. One hundred strategies after 500 seasons and 1.0 meeting cost. The nicest strategies are not
doing quite as well.
Choosiness
0.0 0.2 0.4 0.6 0.8 1.0
Niceness
Summary for 100 strategies and 1000 seasons
0.0 0.2 0.4 0.6 0.8 1.0
legend:
57400 81873 106346 130819 155292
Figure 6. One hundred strategies after 1,000 seasons and 1.0 meeting cost. The surviving strategies
have stabilized to compose a group that is fairly nice and not choosy.
185
THE DESIGN OF MORALITY
nicer and more indifferent, the leaders become further removed from the (1, 0)
standard. This is evidence for thinking that the nicest and least choosy strategies
do very well (relative to the group) only when there are plenty of quite nasty and
choosy strategies in the group. Another way of stating this phenomena is to note
that as the meeting cost rises, the average niceness and the leaders’ niceness both
move toward convergence, while the average choosiness and the leaders’ choosi-
ness similarly converges. Under conditions of signi fi cant environing strain, very
high niceness is not rewarded as much. Again, this result seems to also correspond
to the way that cooperativeness diminishes somewhat when life is very hard.
The fourth phenomenon is observed in tournaments with both modest
meeting costs and permitted births (available space regrettably forbids more
graphs). The group gradually evolves toward an overall average higher niceness
and an average lower choosiness. The most striking long-term stable trait of the
group is how most of the strategies cluster to converge around a fairly high nice-
ness and a fairly low choosiness.
The IRPG game exempli fi es one way to show how the members of group
who can monitor reputations and engage in voluntary cooperations will display,
all other things being equal, an evolutionary convergence upon a relatively high
niceness and a relatively low choosiness. This convergence on reciprocal coopera-
tiveness remains quite robust even under conditions of moderate strain, and
extreme strain only diminishes cooperativeness to a degree without disappearing.
These features are the marks of a society well on its way to conducting itself
morally as we have de fi ned morality in previous sections, since moral conduct
(more so than just etiquette) expects a habitually high degree of cooperativeness
and a habitually high willingness to cooperate with most if not all other members
of the group.
These remarkable traits by themselves are not suf fi cient for judging that these
strategies are fully moral in the sense speci fi ed in earlier sections. Some features
of moral practices are not modeled in this version of the IRPD game, such as
imitative and instructive education and more exacting punishments than just
shunning. However, these limitations are consistent with the way that this IRPD
game can attempt to model hominid conduct. Furthermore, adding education and
punishment to the IRPD game should produce similar overall results displaying the
advantages of nice reciprocal cooperation, since imitating better strategies and
supplementing punishment tend to enhance widespread cooperation (see Boyd
and Richerson,
1992 ; Henrich and Boyd, 2001 ; Fehr and Fischbacher, 2004 ;
Alexander, 2007 ; Marlowe, 2009 ) . The role of punishment, from simple shunning
and shaming to direct in fl iction of physical harm, is as crucial to indirect reciprocity
as it is to morality and law, as would be evolutionarily expected. Sripada (
2005 )
claims that only punishments and not reciprocity’s bene fi ts are behind morality,
but this cannot be right, since agents would not long suffer the costs of enforcing
something that itself offers few bene fi ts.
Third-party punishment is probably deeply connected to the emergence of
in-group morality. Third-party punishment would be a conspicuous display
of fi delity to the group’s welfare, but it would presumably require group sanction
186 JOHN R. SHOOK
(include group pity for the victim but not the transgressor) to produce net bene fi ts
to punishers (Okimoto and Wenzel, 2011 ) . Furthermore, precisely because the
retaliation by the punished back against punishers can diminish punishment’s
effectiveness (and hence effective reduce the situation to a zero-sum game of whom
can dominate the other – see Janssen and Bushman, 2008), genuine morality
would depend on a high degree of internalization so that just punishment is
usually passively accepted. This dependency on internalization may account for
the intuitive moral rightness of retribution. In small societies, the desire to avoid
shaming and suffering from group sanction would have to powerfully instilled
and maintained. In large societies, the development of law to reinforce important
moral norms is commensurate with the emergence of police to wield suf fi cient
force to forestall retaliation by those unrestrained by internalized morals.
7. Natural Morality
Early in hominid evolution, shunning and shaming would have been the primary
means of punishing enforcement, along with intermittent physical violence. Our
preliminary analysis of a basic IRPD game indicates that shunning would be
very effective by itself, and other indirect reciprocity studies show that physical
punishment with group sanction would only heighten enforcement and obedience.
It is therefore reasonable to fi nd both protomorality and the emergence of
universal moral habits in those groups able to engage in non-zero-sum patterns of
cooperation and to track each others’ performance with just a small amount of
information about reputation. The IRPD game successfully models something
of what early hominid groups were likely capable of doing. The dominance and
stability of such habitual social niceness does make a good practical fi t with the
sort of forms of intense and widespread cooperation our hominid ancestors were
developing.
Despite intriguing modeling of the development of moral cooperation, we
cannot forget how these capacities for reciprocal cooperation and protomorality
would still be “in-group” features emerging among members of a social group.
Even as cruelty or betrayal are subsiding within hominid tribes over hundreds of
thousands of years as social intelligence increases to decrease reputation errors,
those nasty deeds can robustly survive between tribes. Familiarity and closeness
remain essential to our evolved sense of morality, as they have remained essential
to charity. Efforts to expand the range of morality would hence require further
reductions to errors of social judgment, expansions of who shall count as part of
the “in-group,” and enhancements of the sense of closeness to others through such
things as new technologies of communication. And the fi eld of ethics has indeed
typically focused of these factors; cognitive psychology has also recommended
increasing the availability of reliable information about other people, their
reputations, and their social interactions (Pollock and Dugatkin,
1992 ; Paolucci
and Conte, 2009 ) .
187
THE DESIGN OF MORALITY
Furthermore, once humans were able to cognitively appreciate how the
norms of moral cooperation were modi fi able, they could take some control over
their habits and try to deliberate enhance morality’s effectiveness at making
cooperation even more reliably bene fi cial for all. Such experimental efforts must
have been halting and unsteady, yet not without practical value, since humans did
not abandon the effort and social morality eventually came to dominate human life.
Without having to suddenly invent compassion, niceness, fairness, cooperativeness,
civility, and trust, humans did intensify efforts to deliberately instill strong respect
for morality’s norms through operant conditionings and educational training,
so that everyone would be more likely to habitually and voluntarily comply in
all situations even if motivational feelings, strategic bene fi ts, or punishments are
not there.
Moral cooperation appears to be so consistently of higher bene fi t to all
members of a social group in the long run that no one would be smarter for
reverting to nothing but ruthless zero-sum games or outright harmful treachery.
Of course, morality is hardly the only kind of practice that would enjoy long-term
survival; we are just the sort of species in which multiple “strategies” would be
distributed across a population. The habits of morality, installed by instinct
and instilled by instruction, will always be statistically distributed: people will
occasionally be mean, short-sighted, and sel fi sh. Morality is designed to prevent,
so far as possible, lapses by individuals into hurtful conduct and unintelligent
noncooperation and to foster fi rm reforms of nonconforming members. In any
society, of course, actual conduct will only approximately track that society’s
moral expectations, but those noticeable deviances are the exceptions that prove
the normative rule.
Morality appears to be just the kind of low-information and largely habitual
practice that could stably emerge among agents capable of recognizing many
individuals, observing others’ interactions, and tracking their reputations. Social
intelligence leads to moral intelligence, and in return, morality improves society.
In essence, during moral cooperation’s long unconscious evolution in social groups,
it represented long-term wisdom for most members even if individuals could not
yet cognitively appreciate that fact. Furthermore, if indirect reciprocity morality
can be propelled by group selection over long stretches of time (as suggested by
Boyd and Richerson,
1990 ; Soltis et al., 1995 ; Nowak, 2011 ) , in-group moral
cooperation can also serve as a smart survival strategy for competing groups
of hominids. The fact that no human society in recorded history has entirely
abandoned morality of its own accord further indicates morality’s durable value.
Morality among hominids displayed a functional design by natural evolution, and
our human moralities now display the imprint of our own redesigns as well.
Moral naturalism appears to have the resources needed to account for the
origin and gradual development of the human practice of morality, so that no
great leaps in emotional, cognitive, or spiritual abilities need to be postulated.
Moral naturalism also appears to be able to account for why humans living in
large societies put so much deliberative effort in ethical redesigns of moralities
188 JOHN R. SHOOK
and their enforcement with law. During the past 10,000 years, dramatically
larger societies have been suffering from the inherited limitations of morality and
getting obsessed with ethics and law, precisely because morality is naturally so
emotional, limited to familiar in-groups, controlled by perceived reputation, and
yet so essential to the needed expansion of nonviolent encounters and reciprocal
cooperations among strangers.
8. Acknowledgments
Jeffrey Miecznikowski’s helpful expertise in biostatistics at the Roswell Park
Cancer Institute is responsible for the current coding for implementing the IRPD
game and the graphs displaying its results. This chapter is also deeply indebted to
Richard Carrier for his insightful criticisms and suggestions which resulted in
many improvements and to Liz Stillwaggon Swan who patiently encouraged every
stage of this chapter’s progress from mere notion to ful fi llment.
9. References
Alexander R (1987) The biology of moral systems. Aldine de Gruyter, New York
Alexander JM (2007) The structural evolution of morality. Cambridge University Press, Cambridge
Axelrod R (1984) The evolution of cooperation. Basic Books, New York
Barash D (2003) The survival game: how game theory explains the biology of cooperation and
competition. Henry Holt, New York
Bekoff M, Pierce J (2009) Wild justice: the moral lives of animals. University of Chicago Press,
Chicago
Boyd R, Richerson PJ (1989) The evolution of indirect reciprocity. Social Netw 11:213–236
Boyd R, Richerson P (1990) Group selection among alternative evolutionary stable strategies. J Theor
Biol 145:331–342
Boyd R, Richerson P (1992) Punishment allows the evolution of cooperation (or anything else) in
sizable groups. Ethol Sociobiol 13:171–195
Boyd R, Richerson P (2005) The origin and evolution of cultures. Oxford University Press, New York
Brinkmann S (2011) Psychology as a moral science: perspectives on normativity. Springer, New York
Carpenter J, Matthews P, Ong’Ong’a O (2004) Why punish? Social reciprocity and the enforcement of
prosocial norms. J Evol Econ 14:407–429
Casebeer WD (2003) Natural ethical facts: evolution, connectionism, and moral cognition. MIT Press,
Cambridge, MA
Colman A (1995) Game theory and its applications in the biological and social sciences, 2nd edn.
Butterworth-Heinemann, Oxford
Darwall S, Railton P, Gibbard A (eds) (1997) Moral discourse and practice: some philosophical
approaches. Oxford University Press, New York
de Waal F (1996) Good natured: the origins of right and wrong in primates and other animals. Harvard
University Press, Cambridge, MA
de Waal F (2006) Morally evolved: primate social instincts, human morality, and the rise and fall of
‘Veneer Theory’. In: Macedo S, Ober J (eds) Primates and philosophers: how morality evolved.
Princeton University Press, Princeton, pp 1–80
de Waal F (2009) The age of empathy: nature’s lessons for a kinder society. Harmony Books, New York
Doris JM et al (eds) (2010) The moral psychology handbook. Oxford University Press, Oxford
189
THE DESIGN OF MORALITY
Fehr E, Fischbacher U (2004) Third party punishment and social norms. Evol Hum Behav 25:63–87
Flanagan O, Sarkissian H, Wong D (2008) Naturalizing ethics. In: Sinnott-Armstrong W (ed) Moral
psychology, vol 1: the evolution of morality. MIT Press, Cambridge, MA, pp 1–26
Foot P (2001) Natural goodness. Oxford University Press, Oxford
Gazzaniga M (2005) The ethical brain: the science of our moral dilemmas. Harper Perennial, New York
Gigerenzer G, Selten R (eds) (2001) Bounded rationality: the adaptive toolbox. MIT Press,
Cambridge, MA
Greene J (2008) The secret joke of Kant’s soul. In: Sinnott-Armstrong W (ed) Moral psychology, vol
3: the neuroscience of morality. MIT Press, Cambridge, MA, pp 35–79
Haidt J (2003) The moral emotions. In: Davidson R et al (eds) Handbook of affective sciences. Oxford
University Press, Oxford
Harris S (2010) The moral landscape: how science can determine human values. Free Press, New York
Hauert C, Traulsen A, Nowak M, Sigmund K (2008) Public goods with punishment and abstaining in
fi nite and in fi nite populations. Biol Theor 3(2):114–122
Henrich J, Boyd R (2001) Why people punish defectors: weak conformist transmission can stabilize
costly enforcement of norms in cooperative dilemmas. J Theor Biol 208:79–89
Henrich J, Boyd R, Bowles S, Camerer C, Fehr E, Gintis H, McElreath R (2001) Cooperation,
reciprocity and punishment in fi fteen small-scale societies. Am Econ Rev 91:73–78
Hirshleifer J, Martinez Coll JC (1988) What strategies can support the evolutionary emergence of
cooperation? J Con fl ict Resol 32:367–398
Joyce R (2001) The myth of morality. Cambridge University Press, Cambridge
Lillehammer H (2003) Debunking morality: evolutionary naturalism and moral error theory. Biol
Philos 18:566–581
Marlowe FW (2009) Hadza cooperation second-party punishment, yes; third-party punishment, no.
Hum Nat 20:417–430
Milinski M, Semmann D, Krambeck HJ (2002) Reputation helps solve the ‘Tragedy of the Commons’.
Nature 415:424–426
Nichols S (2004) Sentimental rules: on the natural foundations of moral judgment. Oxford University
Press, Oxford
Nowak M (2006) Five rules for the evolution of cooperation. Science 314:1560–1563
Nowak M (2011) SuperCooperators: altruism, evolution, and why we need each other to succeed.
Simon & Schuster, New York
Nowak M, Sigmund K (1998) The dynamics of indirect reciprocity. J Theor Biol 194:561–574
Nowak M, Sigmund K (2005) Evolution of indirect cooperators. Nature 437:1291–1298
Ohtsuki H, Iwasa Y (2006) The leading eight: social norms that can maintain cooperation by indirect
reciprocity. J Theor Biol 239:435–444
Okimoto T, Wenzel M (2011) Third-party punishment and symbolic intragroup status. J Exp Social
Psychol 47:709–718
Panchanathan K, Boyd R (2004) Indirect reciprocity can stabilize cooperation without the second-order
free rider problem. Nature 432:499–502
Paolucci M, Conte R (2009) Reputation: social transmission for partner selection. In: Trajkovski G,
Collins SG (eds) Handbook of research on agent-based societies: social and cultural interactions.
Information Science Reference, Hershey
Pollock GB, Dugatkin LA (1992) Reciprocity and the evolution of reputation. J Theor Biol 159:25–37
Sachs JL, Mueller UG, Wilcox TP, Bull JJ (2004) The evolution of cooperation. Q Rev Biol 79:135–160
Sober E, Wilson DS (1998) Unto others: the evolution and psychology of unsel fi sh behavior. Harvard
University Press, Cambridge, MA
Soltis J, Boyd R, Richerson P (1995) Can group-functional behaviors evolve by cultural group
selection—an empirical test. Curr Anthropol 36:473–494
Sripada CS (2005) Punishment and the strategic structure of moral systems. Biol Philos 20:767–789
Suzuki S, Akiyama E (2007) Evolution of Indirect Reciprocity in groups of various sizes and
comparison with direct reciprocity. J Theor Biol 245:539–552
Taylor C, Nowak M (2007) Transforming the dilemma. Evolution 61:2281–2292
190 JOHN R. SHOOK
Trivers RL (1971) The evolution of reciprocal altruism. Q Rev Biol 46:35–57
Trivers RL (1985) Social evolution. Benjamin Cummings, Menlo Park
Tullberg J (2004) On indirect reciprocity: the distinction between reciprocity and altruism, and a
comment on suicide terrorism. Am J Econ Sociol 63:1193–1212
Wong D (2006) Natural moralities. Oxford University Press, New York
Wright R (2000) Nonzero: the logic of human destiny. Pantheon Books, New York
