Philosophy of Mind – Reading Reflection

Based on the two chapters attached answer the following questions:

1. What is the issue the author is concerned with and the main point the authors are trying to make about that issue? What are the reasons the authors give for thinking that this point is worth serious consideration?

Don't use plagiarized sources. Get Your Custom Essay on

Just from $13/Page

Order Essay

2. How do the assigned readings relate to other readings we have discussed (or will discuss)? What do they agree or disagree about?

3. Identify an example from the news, controversial issue, or personal experience that the reading(s) made you think about. How does it relate to the reading and/or illuminate the issues the reading raises?

4. What do you find interesting or confusing about the reading(s)?

5. The question in the chapter on Moral Motivation is about whether there is a form of distinctively moral motivation, that motivates people to act morally. Of the four models of moral motivation presented (instrumentalism, cognitivism, sentimentalism, personalism), which seemed initially most plausible and why? How does the neuropsychology research presented in the paper affect your assessment of the plausibility of that (and other) models?

6. When we say that human moral psychology “evolved” what do you think we (should) mean by that? What interesting lessons can we draw, from the analysis of moral evolution, about human morality and moral practices and judgments, or about our understanding of the reasons why people act morally (or fail to do so)?

Evolution of Moralitys 1

Page 1 of 46

PRINTED FROM OXFORD SCHOLARSHIP ONLINE (oxford.universitypressscholarship.com). (c) Copyright Oxford University Press,
2021. All Rights Reserved. An individual user may print out a PDF of a single chapter of a monograph in OSO for personal use.
Subscriber: University of Central Florida; date: 03 February 2021

The Moral Psychology Handbook
John M. Doris and The Moral Psychology Research Group

Print publication date: 2010
Print ISBN-13: 9780199582143
Published to Oxford Scholarship Online: September 2010
DOI: 10.1093/acprof:oso/9780199582143.001.0001

Evolution of Moralitys 1
Edouard Machery (Contributor Webpage)
Ron Mallon

DOI:10.1093/acprof:oso/9780199582143.003.0002

Abstract and Keywords

This chapter examines whether morality really evolved, as many philosophers,
psychologists, anthropologists, and biologists claim. It distinguishes three
possible versions of this claim and reviews the evidence in support of each. It
concludes that two versions of the claim that morality evolved are relatively well
supported, but that they are unlikely to have significant philosophical
consequences, while the stronger version, which is of real interest to
philosophers, is in fact empirically unsupported.

Keywords: fairness, adaptation, moral emotions, guilt, shame, norms, moral, conventional distinction,
cultural variation, cooperation

Biology provides a broad source of information about humans that has no
substitute. It clarifies long‐standing paradoxes. It shows that some things
have indeed been missing from the debates about morality, and that they
have been missing because the process of organic evolution that gave rise
to all forms of life has been left out of the discussions.

(Alexander, 1987: xvii)

Walking in Darwin’s footsteps, numerous philosophers, psychologists, anthropologists,
and biologists have turned toward evolutionary theory to provide a scientific
understanding of morality.2 In spite of their differences, these thinkers concur on the
provocative claim that morality is an evolved part of human nature, much like a
tendency to weave nets is an evolved part of spiders’ nature.

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143

http://www.hps.pitt.edu/profile/Machery.php

https://www.universitypressscholarship.com/search?f_0=keywords&q_0=fairness

https://www.universitypressscholarship.com/search?f_0=keywords&q_0=adaptation

https://www.universitypressscholarship.com/search?f_0=keywords&q_0=moral emotions

https://www.universitypressscholarship.com/search?f_0=keywords&q_0=guilt

https://www.universitypressscholarship.com/search?f_0=keywords&q_0=shame

https://www.universitypressscholarship.com/search?f_0=keywords&q_0=norms

https://www.universitypressscholarship.com/search?f_0=keywords&q_0=moral

https://www.universitypressscholarship.com/search?f_0=keywords&q_0=conventional distinction

https://www.universitypressscholarship.com/search?f_0=keywords&q_0=cultural variation

https://www.universitypressscholarship.com/search?f_0=keywords&q_0=cooperation

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-3

Evolution of Moralitys 1

Page 2 of 46

This claim is supposed to have far‐reaching implications in moral philosophy
(e.g. Gibbard, 1990; D’Arms, ms). Proponents of evolutionary ethics have often
attempted to justify specific moral norms by appealing to the evolution of
morality (e.g. Spencer, 1892; Richards, 1986, 1989; Rottschaefer & Martinsen,
1990; Rottschaefer, 1991, 1998; Casebeer, 2003a).3 The claim that morality
evolved has also been used as a premise for various skeptical arguments about
(p.4) morality (Ruse, 1986; Woolcock, 2000; Joyce, 2000, 2006; Street, 2006,
2008; for critical discussion, see, e.g., Sober, 1994; Copp, 2008).

While it matters philosophically whether or not morality is a product of
evolution, we find ourselves agreeing with Darwall, Gibbard, and Railton’s
complaint that “more careful and empirically informed work on the nature or
history or function of morality is needed ( . . . ) [p]erhaps unsurprisingly, very
little such work has been done even by some of those who have recommended it
most firmly” (1992: 34). Fifteen years after they expressed this complaint in
their well‐known article “Toward fin de siècle ethics: Some trends,” it remains
unclear whether, and in which sense, morality evolved. Our goal in this chapter
is to answer these questions. Specifically, we propose to clarify the claim that
morality evolved by distinguishing three possible versions of this claim and to
review the evidence in support of each. We conclude that two versions of the
claim that morality evolved are relatively well supported, but that they are
unlikely to yield significant philosophical payoffs, while the stronger version,
which is of real interest to philosophers, is in fact empirically unsupported.

Here is how we proceed. In Section 1, we examine a First interpretation of the
claim that morality evolved—one on which some components of moral
psychology have evolved. We argue that this claim is uncontroversial although it
can be very difficult to show that some particular components of moral
psychology really evolved. In Section 2, we turn to a second interpretation of the
claim that morality evolved, the claim that normative cognition—that is, the
capacity to grasp norms and to make normative judgments—is a product of
evolution. We argue that normative cognition might well have evolved, and that
it may even be an adaptation. Finally, we turn to the philosophically most
interesting interpretation of the claim that morality evolved. In Section 3, we set
out the view that moral cognition, understood as a special sort of normative
cognition, is the product of evolution, and we argue that the evidence adduced in
support of the view is unpersuasive.4 We conclude by expressing our skepticism
about the philosophical implications that can be drawn from the literature on the
evolution of morality. (p.5)

1. The Evolution of Components of Moral Psychology

1.1. The Project

As noted in the introduction, the claim that morality evolved can be interpreted
in at least three different ways. The first interpretation asserts that specific
components (e.g. emotions, dispositions, rule‐based reasoning systems, or

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-70

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-151

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-128

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-129

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-137

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-135

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-136

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-30

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-138

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-167

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-92

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-93

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-157

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-158

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-147

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-33

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-45

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-div1-3

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-div1-4

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-div1-5

Evolution of Moralitys 1

Page 3 of 46

concepts) of moral psychology or specific behaviors typically associated with
morality evolved. Some evolutionary theorists ask whether some of these
components or behaviors evolved, whether they are adaptations, how they could
have contributed to fitness, and whether they evolved exclusively in the hominid
taxon or in other taxa.

Frans de Waal’s work is a good illustration of this approach (e.g. de Waal, 1996;
Preston & de Waal, 2002; see also Darwin, 1871; Bekoff, 2004). He is interested
in whether some of the emotions, dispositions, and cognitive competences that
underlie moral behaviors—e.g. empathy and the recognition of norms—are
present in our closest extant relatives, the apes, as well as in more distant
relatives, such as old‐world and new‐world monkeys. Thus, when he defines his
project at the beginning of Good Natured, he asks, “Do animals show behavior
that parallels the benevolence as well as the rules and regulations of human
moral conduct? If so, what motivates them to act this way? And do they realize
how their behaviors affect others?” (1996: 3).

This interpretation of the claim that morality evolved strikes us as not at all
contentious although specific hypotheses about the evolution of particular
components of moral psychology may be controversial. It is highly plausible that
some moral emotions have an evolutionary history because many emotions have
a long evolutionary history (e.g. Fessler & Haley, 2003). And the cognitive
architecture of morality also relies on various components of social cognition,
many of which also have a long evolutionary history (e.g. Fessler, 1999; Stone,
2006).

Although the idea is fairly uncontroversial, showing that a specific component of
moral psychology evolved is difficult. In the remainder of this section, we focus
on what is perhaps the main difficulty: before looking for (p.6) the homologues5

of human moral traits6 in other species, such as chimpanzees, researchers
should establish that these traits are good candidates for being evolved traits.

1.2. Fairness in Non‐Human Primates?

De Waal has long argued that many important components of moral psychology,
such as the sense of fairness and numerous fairness‐related emotions, e.g.
gratitude (Brosnan & de Waal, 2002) and inequity aversion (Brosnan & de Waal,
2003; Brosnan, 2006), are homologous to psychological systems in other
primates.7 Here, we focus critically on de Waal’s claim that there is evidence for
a precursor of the human sense of fairness among female brown capuchins
(Brosnan & de Waal, 2003; for related results with chimpanzees, see Brosnan,
Schiff, & de Waal, 2005; and for dogs, see Rangea, Horna, Viranyi, & Hubera,
2009). Our goal is not to challenge the idea that many components of our moral
psychology (psychological systems, emotions, etc.) evolved: as noted above, this
claim strikes us as non‐controversial. Rather, focusing on the example of the
sense of fairness, our goal is to illustrate how difficult it is to show that some

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-162

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-122

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-46

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-7

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-162

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-61

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-58

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-156

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-23

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-24

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-22

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-24

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-27

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-127

Evolution of Moralitys 1

Page 4 of 46

particular component evolved because some traits that might seem to be good
candidates for being evolved traits might, on further examination, turn out to be
poor ones.

Brosnan and de Waal’s experimental design is clever. Capuchins, which have
been trained to exchange coins for foods, are put in two adjacent cages. They
are given a coin and have to give it back in order to receive a piece of food,
which is visible in a transparent bowl in front of them. In one condition, the two
capuchins are given a similar recompense, a piece of cucumber. In a second
condition, one monkey receives a piece of cucumber, while the second monkey
receives a piece of grape (a highly valued food). In a third condition, one monkey
receives a piece of cucumber, while the second monkey is given a piece of grape
without having to exchange it for a coin. Brosnan and de Waal measure the rate
of rejection by monkeys, i.e. the number of cases where the monkeys do not
exchange the coin or throw it. The results are surprising: (p.7) female
capuchins reject at a much higher rate the piece of cucumber when the other
capuchin is given a grape for a coin and at an even higher rate when the other
capuchin is given a grape for free.8

Brosnan and de Waal argue that this is tentative evidence for expectations about
fair distributions of food, that is, for norms of fair distribution, as well as
evidence for social emotions similar and homologous to human moral outrage.
They write (2003: 299):

People judge fairness based both on the distribution of gains and on the
possible alternatives to a given outcome. Capuchin monkeys, too, seem to
measure reward in relative terms, comparing their own rewards with those
available, and their own efforts with those of others. They respond
negatively to previously acceptable rewards if a partner gets a better deal.
Although our data cannot elucidate the precise motivations underlying
these responses, one possibility is that monkeys, similarly to humans, are
guided by social emotions. These emotions, known as ‘passions’ by
economists, guide human reactions to the efforts, gains, losses and
attitudes of others.

And, in a related paper (2004: 140), they add:
[C]apuchin monkeys react negatively when another individual gets a better
reward for the same or less effort on a specific task. This finding suggests
that precursors to inequity aversion are present in animals from which our
lineage split millions of years ago.

We are skeptical, and we now argue that it is unlikely that capuchins obey a
norm of fair distribution of windfall gains that is homologous with any human
fairness norm. Let us emphasize that we are not denying that the sense of
fairness—the tendency to find some actions fair and others unfair—has plausibly

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-24

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-25

Evolution of Moralitys 1

Page 5 of 46

evolved. Our claim is more specific: we question whether Brosnan and de Waal’s
work provides evidence that a specific norm of fairness—the equal distribution
of windfall profits—is a homologue present among capuchins and humans. We
then use the example of Brosnan and de Waal’s research to draw some
cautionary conclusions about the search for homologues of the components of
human morality.

First, Brosnan and de Waal (2003) found no effect for male capuchins (but see
van Wolkenten et al., 2007). This is curious if Brosnan and de Waal have really
identified a homologue of a human norm of fair distribution of windfall gains.
Among humans, there is some variation in how males and females (p.8) behave
in similar situations (e.g. Andreoni & Vesterlund, 2001; Solnick, 2001). However,
in an economic game called “the dictator game,” both males and females are
disposed to reject low offers (Solnick, 2001), suggesting that both get upset
when windfall gains are shared unequally.9

In addition, Henrich (2004) has noted two problems with Brosnan and de Waal’s
proposal (but see Brosnan & de Waal’s [2004] reply). First, in similar conditions,
humans tend to react very differently from female capuchins. When they are
offered a deal that they judge to be unfair, humans in many cultures reject this
deal, when such a rejection hurts the person who offered the deal (Henrich et
al., 2004). However, when rejecting the deal does not hurt the person who
offered it, which is a situation analogous to the second and third conditions in
Brosnan and de Waal’s experiment, people tend to accept the deal, in sharp
contrast with capuchins (Bolton & Zwick, 1995).

One could argue on behalf of Brosnan and de Waal that Henrich’s first objection
is unconvincing. Henrich is certainly correct that humans do not behave
similarly to capuchin monkeys. However, it is plausible that in situations that are
analogous to Brosnan and de Waal’s experiments, humans in many cultures feel
annoyed and angry. But because they are able to control their anger and to act
in their best interest, humans accept the offer, when rejecting the offer would
not hurt its author. By contrast, capuchins are not able to control their anger
and are thus unable to act in their best interest. It would be easy to test the
hypothesis that in situations that are similar to Brosnan and de Waal’s conditions
2 and 3, humans and capuchins react similarly in that they both feel anger. In
particular, focusing on humans, one could examine whether there is any facial
micro‐expression of anger (micro‐expressions are facial expressions of emotions
that last only a fraction of second because the agent tries to suppress or control
her emotion). One could also examine whether the brain areas involved in
negative emotions (particularly the insula) and in executive control (the
dorsolateral prefrontal cortex and the anterior cingulate cortex) are activated in
these situations.

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-24

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-166

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-5

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-148

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-83

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-25

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-85

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-12

Evolution of Moralitys 1

Page 6 of 46

More troubling is Henrich’s second criticism. Henrich and colleagues have
documented that there is much cross‐cultural normative diversity in the norms
bearing on the distribution of windfall gains (Henrich et al., 2004, 2005). For
instance, Americans believe that a fair distribution of such gains consists in
splitting them equally. By contrast, in a few small‐scale societies, such (p.9) as
the Machiguengas of the Peruvian Amazon, people seem to expect the
beneficiaries of windfall gains to keep the gain for themselves.

Of course, by itself, variation, including cultural variation, does not show that a
trait (i.e., in the present case, the norm of splitting windfall gains equally) has
not evolved. To begin with, different adaptations can be selected for in different
human populations. Moreover, a trait can also be designed so as to take different
forms in different environments, including different social environments (see,
e.g., Draper & Belsky, 1990). Finally, a given adaptation often varies across
environments because the environment in which organisms develop influences
its development.

However, in this specific case, cultural variation suggests that the norm about
the fair allocation of windfall gains was not selected for. If this norm is really
present in capuchins (as de Waal and colleagues would have it), then it is very
ancient: it had already evolved 30 million years ago, before the platyrrhines (the
phylum to which capuchins belong) and the catarrhines (the phylum to which
humans belong) split. If it is that ancient, then it should plausibly be species‐
typical, exactly like vision is. However, the cross‐cultural research suggests that
it varies across cultures, undermining the hypothesis that the norm of splitting
windfall gains equally is an evolved trait that is homologous in capuchins and
humans. Rather than being a trait homologous to old‐world monkeys and
humans, what is fair in the kind of situations considered by Brosnan and de Waal
or by Henrich and colleagues is determined by the culture‐specific norms
governing economic interactions.

Henrich’s second comment illustrates what is maybe the most important
difficulty that accompanies attempts to discover homologues of human moral
traits. Suppose that one is interested, as de Waal or Bekoff are, in finding
homologues of some of the traits (emotions, norms, concepts, etc.) that
constitute human moral cognition (and not in establishing that these traits are
themselves evolved). Because two traits are homologues only if they evolved
from a common ancestor trait, such a research project assumes that the relevant
components of human moral cognition have evolved or, at least, that they are
good candidates for being evolved traits. Thus, before looking for homologues of
a given component of human moral cognition, it would seem important to ensure
that there are no strong reasons to doubt that this component really evolved.10

The existence of a trait in only a few cultures, its emergence in (p.10) some
recent, historical times, or its acquisition by means of a domain‐general learning
mechanism are strong reasons to doubt that this trait evolved. Thus the cross‐

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-85

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-86

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-48

Evolution of Moralitys 1

Page 7 of 46

cultural variation of how windfall gains should be shared suggests that the norm
of splitting windfall gains equally is unlikely to be an evolved trait. It is then
pointless to look for homologues of this norm.

1.3. Summary: The Evolution of Psychological Components of Moral Psychology

Researchers often focus on some components of moral psychology, such as the
norm of fairness. Then they attempt to determine the evolutionary history of
these components, by studying whether other species, particularly other
primates, also possess the relevant traits. We have argued that this first
interpretation is uncontroversial: some, and perhaps many, components of moral
psychology evolved. At the same time, hypotheses about the evolution of specific
components are difficult to establish, in part because some traits that might
seem to be good candidates for having evolved may, on further examination, turn
out to be poor candidates. Looking for homologues of the components of moral
psychology requires careful attention to a range of data from multiple fields of
scientific inquiry to ensure that there are no strong reasons to doubt that these
components evolved. Cultural psychology and anthropology are needed to
establish that this trait is not culturally local while developmental psychology is
needed to show that it is not acquired by means of a domain‐general learning
mechanism. In this section, we illustrated this difficulty by discussing Brosnan
and de Waal’s claim to have found a homologue of the human fairness norm
about windfall gains.

As noted in the introduction, the claim that morality evolved is often supposed to
have far‐reaching implications in moral philosophy, but little attention has been
dedicated to examining whether and in which sense morality evolved. Now, we
have just argued that, under at least one interpretation, it is uncontroversial that
it did: some components of moral cognition evolved. So, one might ask, what
follows from the evolution of morality in this sense? As we shall now see, very
little.

It is important to distinguish three strategies for answering this question. First,
one might ask whether anything of interest in moral philosophy follows from the
claim that some components of moral cognition evolved. We believe that the
answer is probably negative since we do not see how the argument would go,
and indeed we know of no philosopher who argued to significant philosophical
conclusion from this premise. Second, one might attempt to derive moral
conclusions from the evolution of specific components of moral cognition. For
instance, D’Arms (ms) argues that research on the evolution (p.11) of “self‐
righteous anger”—the anger directed at people who are not angered by others’
moral violations—has moral consequences. Arguments of this kind can take one
of the following two forms. First, one could propose to derive moral norms about
dispositions to act, character traits, etc. from facts about the functions of these
dispositions and character traits.11 For instance, Casebeer contends that “moral
facts are reducible to functional facts” (2003b: 67). Although we do not have the

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-31

Evolution of Moralitys 1

Page 8 of 46

space to discuss this first kind of argument at length, we are not sanguine about
it. Although some normative propositions might be reducible to functional
propositions—e.g. the claim that an organ is working as it should might be
reducible to the claim that it fulfilled its function—we doubt that this can be
done from moral propositions without falling prey to some version of the open
question argument (see Joyce, 2006 for further criticism of Casebeer). The
second type of argument is illustrated by D’Arms’s discussion of the normative
consequences of the evolution of self‐righteous anger. D’Arms correctly notes
that research on the evolution of a morally relevant trait can improve our
knowledge about what this trait is or what it does. For instance, research on the
evolution of self‐righteous anger improves our understanding of the effect of
self‐righteous anger on the social stability of norms. And what a trait is or does
is surely relevant to whether one should morally have this trait. Although this
kind of argument is the most promising way of deriving moral consequences
from some evolutionary findings about a specific component of morality, it is
noteworthy that these consequences are not derived from the fact that this
component evolved, but rather from what it is or what it does. So, just like the
two strategies discussed above, this argumentative strategy does not establish
that moral consequences follow from the evolution of the components of moral
cognition.

2. The Evolution of Normative Cognition

We now turn to the second interpretation of the claim that morality evolved.
Researchers interested in this second interpretation focus on normative
cognition in general: they contend that normative cognition evolved (and often,
that it is an adaptation). In this section, we explain this claim in more detail, and
we argue that there is a small, but suggestive, body of evidence that normative
cognition is an adaptation. (p.12)

2.1. Normative Cognition

Although the nature of norms is a controversial topic in the social sciences (e.g.
McAdams, 1997), we offer an informal account that should be acceptable to
many social scientists. As we shall understand them, norms are attitudes toward
types of actions, emotions, thoughts, or other traits. These norms are typically
shared by many members of a given group and regulate people’s behaviors,
thoughts, emotions, characters, and so on. Their content essentially involves
deontic concepts, such as SHOULD or OUGHT. Such norms can prescribe or
forbid a thought, behavior, or any other characteristic, and may be associated
with a disposition to punish those individuals who do not comply with the norms.

Normative cognition is underwritten by a complex cognitive architecture. People
learn and assimilate, explicitly and implicitly, numerous norms; they are
motivated to comply with them; and they typically expect others to comply with
them. Emotions are also a key component of this cognitive architecture. Several
negative emotions are triggered by norm violations (Haidt, 2003; Fessler &

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-93

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-112

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-75

Evolution of Moralitys 1

Page 9 of 46

Haley, 2003). Norm violators are likely to feel shame or guilt (depending on
which emotion is emphasized in their culture).12 Victims of norm violations and
third parties are likely to feel anger or disgust toward norm violators. These
emotions motivate behavior: the anticipation of feeling ashamed and guilty
motivates avoiding the violation of norms, shame and guilt motivate reparative
behavior, and anger motivates punishment (e.g. Fehr & Gächter, 2002; Haidt &
Sabini, 2000). Disgust causes third parties to distance themselves from norm
violators, which results in the loss of cooperative opportunities for the norm
violators. Anticipatory fear of shame or guilt often motivates norm compliance
(Fessler, 2007).13 In addition to these negative emotions, positive emotions are
caused by norm compliance. People feel elevation when others endure some cost
to comply with certain norms (Haidt, 2003).

It is remarkable, however, there has been little systematic work on human
normative cognition. One exception is Chandra Sripada and Stephen Stich’s
(2006) article. Sripada and Stich argue for the existence of two cognitive
systems subserving the psychology of norms: an acquisition mechanism and an
implementation mechanism. The function of the acquisition mechanism is to
learn the norms that are prevalent in one’s culture, while the function (p.13) of
the implementation mechanism is to store representations of these norms, to
produce some intrinsic desires to comply with them, and to motivate people to
punish norm violators. While their hypothesis is consistent with the existence of
innate representations of norms, Sripada and Stich speculate that the
implementation mechanism does not store any innate representation of norms.
Rather, children, and sometimes adults, need to learn the prevalent norms of
their social community.

2.2. How to Study the Evolution of Normative Cognition?

Many researchers’ work on the evolution of morality is best understood as being
about the evolution of normative cognition in general, since they do not single
out a specific kind of norms (i.e. moral norms).14

Before going any further, it is worth noting that there are many ways to
investigate the evolution of a trait. It is particularly useful to distinguish two
related claims. To claim that a trait evolved is simply to claim that the trait has a
phylogenetic history, and one project would be to inquire into this history.15 That
is, one can study what changes took place in the psychology of our primate
ancestors during the evolution of normative cognition (just as one can study the
evolution of the human eye by identifying the changes that took place during the
evolution of the mammalian eye). A stronger claim is that normative cognition
constitutes an adaptation. An adaptation is a specific sort of evolved trait—i.e. a
trait whose evolution is the result of natural selection. Since not all products of
evolution are adaptations, someone who conjectures that normative cognition is
an evolved trait can also examine whether it is an adaptation, the by‐product of
another adaptation, or an evolutionary accident. In addition, if one proposes that

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-61

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-56

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-76

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-60

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-75

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-154

Evolution of Moralitys 1

Page 10 of 46

normative cognition is an adaptation, one should consider what its evolutionary
function might be—that is, what selective forces might have driven its evolution.

2.3. Evidence that Normative Cognition is an Adaptation

Sociological and psychological evidence suggests that normative cognition is an
adaptation. We consider these two types of evidence in turn (for further
evidence, see Cummins, 1996b).

Norms, either informal or formal, are ancient: the historical record has no trace
of a society without norms. Furthermore, norms are universal (although (p.14)
the content of norms varies tremendously across cultures). Small‐scale societies
are typically regulated by informal norms, while large‐scale societies are
typically regulated by informal and formal norms. All known societies also have
policing mechanisms that ensure people’s compliance with the prevalent norms
(Brown, 1991). These policing mechanisms naturally vary across cultures. In
some societies, but not in all, policing is the socially sanctioned role of a
dedicated group of individuals (e.g. policemen, Iran’s “moral” police [a branch of
the Islamic Revolutionary Guard], etc.). In addition, in all societies, informal
social practices contribute to ensure people’s compliance with the prevalent
norms (Boehm, 1999). These include gossip (Dunbar, 1996) and various forms of
ostracism (Brown, 1991). Finally, as noted by Sripada and Stich (2006), norms
permeate people’s life: few behaviors and decisions are immune to the influence
of some norm or other.

The antiquity and universality of norms is evidence that normative cognition
evolved. When a trait is ancient and universal, it is either because it can be
easily acquired by individual learning or by social learning, or because a
developmental system is designed to ensure its regular development. In the
latter case, but not in the former case, the universality and antiquity of a trait is
evidence that it evolved. Ancient and universal traits that are not evolved, such
as the belief that the sun rises every morning, are easy to acquire from one’s
physical and social environment (Dennett, 1995). Since it is difficult to see how
one could acquire the capacity for normative attitudes toward thoughts,
behaviors, and other traits—i.e. a capacity for norms—from one’s environment
(in contrast to acquiring specific norms, which can obviously be learned), it is
plausible that normative cognition evolved.

Turning from sociological to psychological considerations, evidence suggests
that people are endowed with a reasoning capacity that is specific to the domain
of norms. While people reason poorly about non normative matters, they are
adept at reasoning about normative matters (for review, see Cosmides & Tooby,
2005). Both Western and non‐literate Shuar Amazonian subjects easily
determine in which situations deontic conditionals, such as “If you eat
mongongo nut (described as an aphrodisiac in the cover story), then you must
have a tattoo on your chest” (described as a mark denoting married status), are

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-42

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-28

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-11

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-51

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-28

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-154

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-47

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-35

Evolution of Moralitys 1

Page 11 of 46

violated, while they are surprisingly poor at determining in which situations
indicative conditionals, such as (“If there is a red bird in the drawing on top,
then there is an orange on the drawing below”), are false (Cosmides, 1989;
Sugiyama, Tooby, & Cosmides, 2002). Although the interpretation of these
findings remains somewhat controversial (e.g. Sperber, Cara, & Girotto, 1995),
they suggest to us that people are distinctively adept at detecting norm
violation. (p.15)

Furthermore, just like adults, young children are much better at reasoning about
the violations of deontic conditionals than about the falsity of indicative
conditionals (Cummins, 1996a; Harris & Núñez, 1996). For instance, Cummins
(1996a) showed 3‐year‐old children some toy mice and told them that some, but
not all, could squeak. She also told them that some squeaky mice were inside the
house, while others were outside. Finally, she told children that a cat was
hunting mice outside the house, but only when they squeaked. Half of the
children were told that Queen Minnie Mouse had told the mice, “It’s not safe
outside for the squeaky mouse, so all squeaky mice are in the house.” Those
children were then asked to say which mice must be examined to see whether
Minnie Mouse was right. The other half was told that Queen Minnie Mouse had
told the mice, “It’s not safe outside for the squeaky mouse, so all squeaky mice
must stay in the house.” Those children were then asked to say which mice must
be examined to see whether Minnie Mouse’s rule has been broken. While almost
65% of 3‐year‐olds answered correctly the second question, only 30% of them
answered correctly the first question. These findings suggest that the capacity to
reason about norms develops early (as early as children’s fourth year) and in a
distinctive manner (since it seems independent from children’s capacity to
reason about conditionals in general).

The existence of a cognitive system that seems dedicated specifically to produce
good reasoning about norms from an early age on provides some suggestive
evidence that normative cognition is an adaptation.16 Generally, the functional
specificity of a trait is (defeasible) evidence that it is an adaptation.
Furthermore, the fact that a trait develops early and that its development is
distinctive—it is independent from the development of other traits—suggests
that natural selection acted on its developmental pathway. The early
development of a psychological trait suggests that it is not acquired as a result
of our domain‐general learning capacity; the distinctive development of a
psychological trait suggests that it is not acquired as a by‐product of the
acquisition of another psychological capacity (for further discussion, see
Machery, forthcoming). Thus, evidence tentatively suggests not only that
normative cognition is an evolved trait, but also that it is an adaptation.

The findings just considered provide suggestive (though inconclusive) evidence
that normative cognition is an adaptation. It is instructive to anticipate (p.16)
Section 3 and to compare these findings with the body of evidence typically

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-34

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-159

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-152

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-41

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-79

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-41

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-div1-5

Evolution of Moralitys 1

Page 12 of 46

adduced to support the claim that moral cognition, conceived as a specific kind
of normative cognition, evolved. While researchers often claim that moral
cognition, conceived as a specific kind of normative cognition, is universal, we
shall argue in the next section that the evidence for this claim is lacking. By
contrast, the evidence for the antiquity and universality of norms is extremely
solid. We shall also challenge the claim that a key component of moral cognition
—i.e. grasping the distinction between properly moral norms and conventional
norms—develops early and reliably. By contrast, although the research on adults’
and children’s capacity to reason with deontic conditionals is not entirely
uncontroversial, it is on safer ground. In this case, psychologists have indeed
typically not challenged the claim that people reason better with deontic
conditionals than with indicative conditions; rather, they have focused on how
this difference is to be explained.

2.4. “How‐Possible” Models of the Selection of Normative Cognition

In addition to this small, but suggestive, body of evidence that normative
cognition in general is an adaptation, several models show how normative
cognition could have been selected for during the evolution of hominids.17 These
“how‐possible” models (Brandon, 1990) do not establish how normative
cognition actually evolved: evidence is lacking to answer this question. But these
models show, first, that the hypothesis that normative cognition was selected for
is consistent with our knowledge of evolution; second, the selection of normative
cognition in evolutionary models is robust: in several possible evolutionary
situations—those represented by the how‐possible models—normative cognition
would have been selected for.18

Since the 1980s, Robert Boyd, Peter Richerson, and their colleagues have
developed a series of models explaining how norms can be stable in a
community. Here, we present two of their models informally. In a well‐known
model, Boyd and Richerson (1992) have shown that punishment (actions
inflicting a cost on norm violators) can stabilize any norm, including norms that
prescribe costly behaviors such as cooperation. Suppose for an instant that
punishment is cost‐free—the punisher does not endure any cost when she
punishes. By violating a norm, norm violators might get some (p.17) benefit or
might avoid some cost, when norm compliance is costly. However, because they
are punished, violators suffer a cost and do less well than those who comply with
norms, but avoid punishing (“lazy norm compliers”) and those who comply with
norms and enforce them (“punishers”). If successful behaviors tend to become
common in a population (maybe because they are imitated by others), then
compliance with norms will prevail. Thus punishment can stabilize norms.
Importantly, in this model, norm compliance does not depend on the content of
the norms, only on the punishment of norm violators. Thus different norms might
be stabilized in different societies, consistent with the diversity of norms across
cultures.

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-19

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-16

Evolution of Moralitys 1

Page 13 of 46

But, of course, punishment is not cost‐free, although, in humans, it might be low‐
cost because of the development of weapons and of language (which allows
people to hurt others by gossiping negatively about them). Because punishment
is not cost‐free, lazy norm compliers do better than punishers. Thus compliance
without punishment might become more common in a population at the expense
of compliance with punishment (Boyd & Richerson, 1992). However, if lazy norm
compliers become more common, norm violators will in turn increase in
frequency, because they will be less often punished. This will prevent the
stabilization of norms. How, then, is norm compliance obtained?

This problem has been addressed in various ways. One could first suggest that
punishment itself is a norm, and that lazy norm compliers get punished when
they fail to punish norm violators, a type of punishment called
“metapunishment” or “second‐order punishment” (Boyd & Richerson, 1992).
However, this suggestion only pushes the problem one step further, because
metapunishment is itself costly.

Henrich and Boyd (2001) have proposed an alternative solution (for a different
model, see Boyd et al., 2003). In their model, behaviors are transmitted
culturally, but biased by conformism and prestige. Conformist bias means that
common behaviors are more likely to be transmitted than rare behaviors, while
prestige bias means that high‐payoff behaviors are more likely to be transmitted
than low‐payoff behaviors. Conformism favors the cultural transmission of the
prevalent behaviors, whatever these are, while prestige‐biased transmission can
undermine norm compliance, punishment, and metapunishment, because these
can be costly. Suppose now that, in a population, everybody complies with the
norms, but fails to punish norm violators (everybody is a lazy norm complier). An
intruder who fails to comply with the norms (a norm violator) would be better off
than these lazy norm compliers, and prestige bias would tend to lead others to
become norm violators, providing that this bias was stronger than the
countervailing bias to conform with the more common (p.18) compliant
behavior. So, where conformism is weak, norm violation will become common.

Consider now a second case. Everybody complies with the norms and punishes
violators (everybody is a punisher). An intruder norm violator would not be
better off than the common punishers, because she would be punished. But an
intruder lazy norm complier would be better off than the common punishers,
since she would not get punished (since she complies with the norms) and would
avoid the cost of punishing others. By contrast, punishers would pay the cost of
punishing the other punishers who would fail by mistake to comply with the
prevalent norms. The extent of a lazy norm complier’s advantage over the
punishers depends on how costly it is to punish and on how often punishers fail
to comply by accident with the prevalent norms.19 If this advantage is large
enough to offset the advantage conformism gives to punishers’ common behavior
(that is, to offset the fact that due to people’s conformism, common behaviors

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-16

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-84

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-17

Evolution of Moralitys 1

Page 14 of 46

are more likely to be imitated by others than rare behavior), compliance with the
norms without punishing would become common. If lazy norm compliers replace
the punishers in a population, the norm violators will ultimately invade this
population.

Now, consider a third case. Everybody complies with norms, punishes violators,
and punishes non‐punishers (everybody is a metapunisher). An intruder lazy
norm complier would not be better off than the common metapunishers, because
she would be punished for non‐punishing the metapunishers’ accidental norm
violations. But an intruder who would comply with the prevalent norms, punish
violators, but fail to punish those who fail to punish (a lazy punisher) would be
better off than the common metapunishers because she would not be punished
(she complies with the norms) and because she would avoid the cost of
punishing the failure to punish. By contrast, metapunishers would pay the cost
of punishing those metapunishers who would fail by mistake to punish the
metapunishers who by mistake violate a norm. However, the advantage of a lazy
punisher over the metapunishers is smaller than the advantage of a lazy norm
complier over the punishers, for the former advantage depends on two mistakes,
i.e. a metapunisher failing by accident to comply with a prevalent norm and
another metapunisher failing by accident to punish the accidental non‐
compliance. The lazy punisher’s advantage is thus less likely to offset the
advantage conformism gives to the common metapunishing behavior. Of course,
the same argument applies at further orders of punishment. Thus, even if
conformism is weak, it can stabilize punishment at some order (p.19) of
punishment. If punishment is stable, then norm compliance is itself stable. Thus
Henrich and Boyd show that with a small amount of conformism that stabilizes
metapunishment (or some higher‐order punishment), costly norm compliance is
stable: compliance with the prevalent norms, even at one’s own cost, is more
likely to be culturally transmitted than non‐compliance with these norms.

Now, suppose that, as this and other models suggest is possible, cultural
transmission stabilized norms during human evolution. Because norm violators
were punished for violating the prevalent norms and because lazy norm
compliers were punished for not punishing norm violators, both norm violators
and lazy norm compliers incurred costs that punishers (who comply with the
norms and punish) did not incur. Our ancestors who were good at learning the
prevalent norms and who were motivated to comply with them and to punish
norm violators might thus have had a fitness advantage over people who learned
badly the prevalent norms or had a weak (if any) motivation to comply with
them. Thus natural selection might have favored some important elements of the
architecture of normative cognition—a disposition to learn prevalent norms, a
disposition to comply with norms, and a disposition to punish norm violators.

Evolution of Moralitys 1

Page 15 of 46

2.4. Summary: The Evolution of Normativity

In this section, we have focused on a second interpretation of the claim that
morality evolved: normative cognition—the capacity to grasp and apply norms—
evolved. A small body of evidence suggests that normative cognition evolved by
natural selection. Furthermore, several how‐possible models show how
normative cognition could have been selected for during the evolution of the
human species.

Importantly, this conclusion is cold comfort to those philosophers who want to
get some philosophical mileage out of evolutionary findings. This is particularly
clear when one focuses on the argument that the evolution of morality would
undermine the authority of moral norms (e.g. Ruse, 1986; Joyce, 2006). Suppose
that this argument from the evolution of morality is meant to hang on the
reading of the claim that morality evolved considered in this section: normative
cognition in general evolved. While this argument would then rest on a premise
that is supported by a small, but convincing body of evidence, it would have very
troubling consequences. If the evolution of normative cognition really
undermines the authority of moral norms, then it should also undermine the
authority of any kind of norms (including epistemic norms), for there is no
reason why only the authority of moral norms would be undermined by the
evolution of the capacity to grasp norms tout court. (p.20) The unpalatable
nature of this conclusion would plausibly give grounds for concluding that the
argument from the evolution of morality is flawed. The upshot should be clear: if
the claim that the evolution of morality undermines the authority of morality is
to be plausible at all, it has to be based on an interpretation of the claim that
morality evolved different from the one considered in this section. In our view, it
is no accident that when philosophers have attempted to derive philosophical
implications from the hypothesis that morality evolved, they have typically
focused on a third reading of this hypothesis. We now turn to this reading.

3. The Evolution of Moral Normativity

3.1. The Project

Researchers who endorse the third version of the claim that morality evolved
start by drawing a distinction among different types of normative cognition and
by singling out one specific type of normative cognition, which they call
“morality.” They then proceed to argue for the evolution of this type of normative
cognition. Consider each step of this project.

3.1.1. Morality as a Type of Normativity

As we saw in Section 2, normative cognition includes the capacity to grasp
norms, to make normative judgments, and to be motivated to act according to
the norms that one endorses.20 Norms have to do with the regulation of people’s
actions, emotions, thoughts, or other traits. They specify what kinds of
behaviors, emotions, thoughts, or other characteristics are mandatory,
permissible, or recommended.21 In turn, normative judgments consist in judging

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-138

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-93

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-div1-4

Evolution of Moralitys 1

Page 16 of 46

that behaviors, emotions, and thoughts (one’s own or others’) are
unconditionally mandatory, permissible, or recommended.22

Turn now to the claim that moral cognition is a distinctive type of normative
cognition. The basic idea is that moral norms are a distinct type of (p.21) norm
and that related entities like moral judgments, moral motivations, and moral
behaviors and thoughts are similarly distinct. For the sake of simplicity, we focus
our discussion here especially on norms and normative judgments. There are
many kinds of normative judgments, and moral judgments are only one of them.
Other kinds of normative judgments might include judgments about what is
rational (e.g. “you shouldn’t believe that, given what else you believe”),
aesthetically appropriate (“one should never wear green pants with a yellow
shirt”), prudent (“if you want to live a long life, you should wear your seatbelt”),
and conventionally expected (“if you are satisfied with the service, the tip should
be at least 20%”). The first interpretation of the claim that morality is the
product of evolution rests on the idea that moral judgments provide a distinctive
means of regulating or evaluating actions, emotions, intentions, or character.

Richard Joyce’s (2006) book The Evolution of Morality provides a particularly
clear illustration of the kind of research considered here (see also Ruse, 1986;
D’Arms, 2000; Joyce, 2008a, b). Focusing on moral judgments, he proposes that
seven properties distinguish moral judgments and moral behaviors from other
kinds of normative judgments:

• Moral judgments (as public utterances) are often ways of
expressing conative attitudes, such as approval, contempt, or, more
generally, subscription to standards; moral judgments nevertheless
also express beliefs; i.e., they are assertions.
• Moral judgments pertaining to action purport to be deliberative
considerations irrespective of the interests/ends of those to whom
they are directed; thus they are not pieces of prudential advice.
• Moral judgments purport to be inescapable; there is no “opting
out.”
• Moral judgments purport to transcend human conventions.
• Moral judgments centrally govern interpersonal relations; they
seem designed to combat rampant individualism in particular.
• Moral judgments imply notions of desert and justice (a system of
“punishments and rewards”).
• For creatures like us, the emotion of guilt (or “a moral conscience”)
is an important mechanism for regulating one’s moral conduct. (2006:
70–71)

He adds that “so long as a kind of value system satisfies enough of the above, then it
counts as a moral system” (71).

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-93

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-138

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-43

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-94

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-95

Evolution of Moralitys 1

Page 17 of 46

While Joyce clearly intends his list to function not as a checklist of necessary and
sufficient conditions, but rather as something like a cluster concept, it is worth
emphasizing that his claim is substantive and provocative precisely because of
the rich characterization of moral judgments that he offers. That is, (p.22)
according to his account, moral judgments have many distinctive properties that
differentiate them from other sorts of normative judgments. In expressing
skepticism about whether the capacity to make moral judgments is a product of
evolution, we mean specifically to doubt Joyce’s view and others like it. We do
not doubt that there exists some thin description of the class of moral judgments
that could be offered such that, under this description, the capacity to make
moral judgments would be the product of evolution.23 We deny the claim that
when moral judgments are richly described, the capacity to make them is a
product of evolution.

3.1.2. Morality as an Evolved Trait

After characterizing distinctively moral cognition—for example, after
characterizing moral norms as a distinct kind of norm or moral judgments as a
distinct kind of normative judgment—researchers conjecture that such
distinctively moral cognition is an evolved trait. Remember that in Section 2 we
distinguished two related ways of studying the evolution of a trait. One could
simply claim that morality, as a specific kind of normative cognition, evolved.
One would then study what changes took place in the psychology of our primate
ancestors during the evolution of the grasp of distinctively moral norms and of
the capacity to make moral judgments. Since, as noted in Section 2.2, not all
products of evolution are adaptations, someone who conjectures that the
capacity to grasp moral norms and the capacity to make moral judgments are
evolved traits can also examine whether they constitute an adaptation (as
Dennett, 1995, Kitcher, 1998, and Joyce, 2006 have claimed), whether it is a by‐
product of another adaptation, or whether it is an evolutionary accident
(Williams, 1988). In addition, if one proposes that the grasp of moral norms and
the capacity to make moral judgments constitute an adaptation, one should
consider what its evolutionary function might be—that is, what selective forces
might have driven its evolution.

Joyce, for example, suggests that the capacity to make moral judgments is a
specifically human adaptation for motivating us to act in a prosocial way. In
essence, moral judgments provided our ancestors with compelling reasons to act
in ways that typically favor others and that can be detrimental to themselves
(see also Dennett, 1995). As a result, moral judgments reliably caused prosocial
behavior. Moreover, loosely following Robert Frank (1988), Joyce contends that
because moral judgments can be linguistically expressed, they signal to (p.23)
others that we are committed to act in a prosocial way. The capacity to make
moral judgments was favored by natural selection because reliable prosocial
behavior and the signaling of one’s dispositions to act in a prosocial way were

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-div1-4

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-div2-5

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-47

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-98

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-93

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-164

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-47

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-67

Evolution of Moralitys 1

Page 18 of 46

favored during the evolution of the human species, possibly because prosocial
behaviors were reciprocated.

Joyce is not the only researcher to claim that the capacity to make moral
judgments, understood as a distinct type of normative judgment, is an
adaptation. In Moral Minds, Marc Hauser (2006) contends that, like the
language faculty, the moral faculty is a distinct psychological adaptation,
although he has little to say about its evolutionary function (2006: xvii):

The central idea of this book is simple: we evolved a moral instinct, a
capacity that naturally grows within each child, designed to generate rapid
judgments about what is morally right or wrong based on an unconscious
grammar of action.24

In contrast to these projects, we see little reason to believe that the grasp of
distinctively moral norms and the capacity to make moral judgments, understood as a
specific kind of normative judgments, evolved at all, and so we doubt that they
constitute an adaptation, a by‐product, or an evolutionary accident. We conjecture that
in this respect, the capacity to grasp moral norms and the capacity to make moral
judgments might be similar to chess or handwriting. The capacities to play chess and
to write involve various evolved cognitive traits (e.g. visual recognition and
memorization of rules for the former), but they did not evolve. Similarly, we conjecture
that the capacity to grasp moral norms and the capacity to make moral judgments
involve various evolved cognitive traits (including, as we proposed in Section 2, a
disposition to grasp norms in general), but they themselves did not evolve. In any case,
as we shall argue now, none of the available evidence suggests that they did.
In the remainder of Section 3, we look critically at two different forms of
argument for the claim that moral cognition evolved:

(1) There are plausible adaptationist models that predict its selection.
Thus the grasp of moral norms and the capacity to make moral judgments
are likely to be an adaptation and, a fortiori, to have evolved.
(2) The universality and innateness of the capacity to grasp moral norms
and to make moral judgments is evidence that it is an evolved trait.

(p.24)
3.2. Adaptationist Models of the Evolution of Morality

It is common to argue that a trait is an adaptation by showing that there are
plausible adaptationist models that predict the selection of this trait. For
instance, some sociobiologists have provided this kind of argument in support of
the hypothesis that female orgasm is an adaptation: they have argued this
because it is plausible that, among our ancestors, those females who were able
to have orgasms were more motivated to have sex and, as a result, were more
likely to have descendants than those females who had no orgasm (for review
and criticism, see Lloyd, 2005). This kind of argument has been severely
criticized in the philosophy of biology because it involves telling just‐so stories
that cannot be supported by any available evidence (Gould & Lewontin, 1979;

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-81

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-93

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-div1-4

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-div1-5

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-108

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-72

Evolution of Moralitys 1

Page 19 of 46

Figure 1.1. Reciprocal altruism

Note: c and c′ stand for the costs of the
altruistic actions for the agents A and B, b
and b′ for the benefits bestowed upon the
beneficiaries of the actions, A and B (c < b′, b > c′), and t and t′ for the times of the
actions.

Kitcher, 1985).25 Here, we do not discuss the value of this kind of argument in
general. Rather, we criticize specific adaptationist models of the evolution of
morality. We argue that these models do not support the claim that moral
cognition, conceived as a distinct kind of normative cognition, was selected for.
We first consider models that appeal to reciprocal altruism, then we consider
models that are based on indirect reciprocity, before finally raising a general
problem for all current adaptationist models of the evolution of morality.

Many adaptationist models of the evolution of morality appeal to a specific
evolutionary mechanism, reciprocal altruism (Figure 1.1). The notion of
reciprocal altruism was developed by evolutionary biologist Robert Trivers as a
possible explanation of altruistic behavior among non‐related organisms
(Trivers, 1971).26 The idea goes roughly as follows: a gene G for an altruistic
trait T is favored by natural selection if T benefits discriminatively recipients
who are likely to reciprocate in the future. Reciprocation is delayed and may be
of a different kind (as happens, e.g., when chimps exchange grooming for
political support, a phenomenon reported in Foster et al., 2009). To use a toy
example, a gene G for sharing food is favored by natural selection if the bearer
of G shares food with individuals that are likely, at some point in the future, to
reciprocate by acting in a way that increases the fitness of the bearer (p.25)

of G. According to Triver’s model,
three conditions have to be met
for reciprocal altruism to explain
the evolution of altruism. First, for
the two individuals involved in a
reciprocal interaction, the cost of
being altruistic (in units of fitness)
must be lower than the benefit
received from the reciprocator’s
altruism. The cost of sharing food
for the bearer of G must be lower
than the benefit taken from the
reciprocation at a later time.
Second, the benefit of altruism
must be withheld from those
individuals who did not
reciprocate in the past—usually
called “cheaters.” The bearer of G
should refrain from sharing food with those individuals who did not reciprocate in the
past. Third, individuals must interact repeatedly, so that the cost to cheaters of
foregone protracted reciprocal interactions (in units of fitness) is greater than the
benefit cheaters take from non‐reciprocating. The bearer of G must share food with
individuals with whom she is likely to interact often, so that not benefiting from food
sharing over a long period of time is more costly than avoiding the cost of
reciprocating a past benefit.

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-97

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-figureGroup-1

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-160

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-66

Evolution of Moralitys 1

Page 20 of 46

It is likely that reciprocal altruism fails to explain a range of human altruistic
behaviors. As we saw, Trivers’s idea is that if interactions between two
individuals last for long enough, traits that benefit discriminatively those
individuals that are likely to reciprocate will be favored by natural selection.
This hypothesis fails to explain why people are disposed to benefit individuals
with whom they will probably have no further interaction. For instance, people
regularly tip waiters while on vacations, even though they will have no further
interactions with them.

To explain away this difficulty, one might suggest that our ancestors lived in
close‐knit societies, where interactions were typically repeated. As a result, we
evolved cooperative dispositions that do not distinguish between interactions
that are likely and interactions that are unlikely to be repeated. In substance, we
evolved to treat every interaction as if it was likely to be repeated (e.g. Johnson,
Stopka, & Knights, 2003). This reply won’t do, however (Fehr & (p.26) Henrich,
2003). Anecdotal reports and experimental evidence in behavioral economics
show that people effortlessly distinguish between interactions that are unlikely
to be repeated and long‐term cooperative situations, and that they behave
differently in these two types of situations. For instance, in experimental
contexts, people tend to be less generous and helpful in the first type of
situations than in the later type of situations (e.g. Gächter & Falk, 2002). Thus it
does not seem to be the case that we behave cooperatively in some non‐repeated
interactions (e.g. when we tip strangers) because we evolved to treat every
interaction as if it is likely to be repeated.27

Moreover, it is unclear whether our ancestors really lived in small and close‐knit
communities. Reciprocal altruism can explain the selection of altruistic
behaviors only if our ancestors were able to discriminate between those
individuals who failed to reciprocate altruistic acts (“cheaters”) and those who
did reciprocate, that is, only if they were able to remember who did what. The
larger the group in which our ancestors belonged and the more fluid
membership in residential units was,28 the less likely it is that this condition was
met. Paleoanthropological evidence suggests that at least for the last 50,000
years, our ancestors have lived in large groups of several thousands of members
—too large for them to have been able to remember who did what (see
Richerson & Boyd, 1998, 1999 for a detailed review of the evidence).
Furthermore, as Richerson and Boyd write (1999: 254), “Foraging societies are
simple by comparison with modern societies, but even the simplest
contemporary hunting and gathering peoples, like !Kung San and the peoples of
Central Australia, link residential units of a few tens of people to create societies
of a few hundred to a few thousand people.” Migrations and long‐distance
economic exchanges have also characterized the life of our ancestors for maybe
several hundreds of thousands of years (McBrearty & Brooks, 2000). Finally, in

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-91

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-57

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-69

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-130

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-131

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-113

Evolution of Moralitys 1

Page 21 of 46

Figure 1.2. Indirect reciprocity

Note: c and c′ stand for the costs of the
altruistic actions for the agents A and D,
b and b′ for the benefits bestowed upon
the beneficiaries of the actions, B and A
(c < b′), and t and t′ for the times of the actions.

many modern hunter‐gatherer societies, membership in residential units is fluid
(see, e.g., Hill, 2003 on the Ache; Smith, 2004 on the Hadza).

Clearly, these findings do not establish beyond doubt that reciprocal altruism
could not explain the evolution of altruism. Morality could have evolved before
our ancestors lived in large and fluid groups. Furthermore, even if our ancestors
lived in large and fluid groups, they might have interacted altruistically only with
a small number of group members. Nonetheless, the body of evidence about the
size and fluidity of the social groups that have been common during (p.27)

part of the evolution of our species
casts at least some doubt on the
importance of reciprocal altruism
for understanding the evolution of
altruism.
A better reply to the charge that
reciprocal altruism fails to
explain a range of human
altruistic behaviors consists in
extending the notion of
reciprocal altruism.
Evolutionary biologist Richard
Alexander (1987) has done
precisely this with the notion of
indirect reciprocity.29 Roughly,
the idea is that a trait that
benefits another individual will
be favored by natural selection
if the possession of this trait
increases the probability of
benefiting from the altruism of
a third party. One way to characterize indirect reciprocity is in terms of
reputation. The possessor of the altruistic trait increases her reputation and
people with a good reputation are the target of others’ altruism (see Figure 1.2).

Prominent researchers, including Trivers himself, have proposed that while
originally developed to explain altruism in a large range of species (including
some species of fish), reciprocal altruism and indirect reciprocity also explain
the evolution of morality in humans. Alexander puts it succinctly (1987: 77):
“Moral systems are systems of indirect reciprocity.”30 ,31 (p.28)

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-87

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-146

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-3

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-figureGroup-2

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-3

Evolution of Moralitys 1

Page 22 of 46

Figure 1.3. Public goods

Note: c stands for the cost of the
altruistic action for the agent A, b for the
benefit bestowed upon the beneficiaries
of the actions, B and C, and t for the time
of the action.

In spite of its impressive
pedigree, the hypothesis that
reciprocal altruism or indirect
reciprocity selected for moral
cognition, as a distinct kind of
normative cognition, faces
serious challenges. Following
Sripada (2005), we highlight
three shortcomings of this
hypothesis. First, reciprocal
altruism and indirect reciprocity
are supposed to explain the
evolution of behaviors in
pairwise interactions, while
moral behavior does not always
take place in the context of
pairwise interactions. For
instance, morally sanctioned altruistic behaviors often benefit a large number of
people, as is illustrated by the sacrifice of soldiers for their country.

One could propose to extend Trivers’s reciprocal altruism to interactions that do
not take place in pairs. This kind of situation is typically modeled by means of a
public‐goods game (aka, n‐person prisoner’s dilemma) (Figure 1.3).

Thus one could examine whether a trait that benefits the other members of the
group conditional on the altruism of all the other members (or of a specific
proportion of these group members) could be favored by natural selection, if
interactions between the members of this group lasted long enough.32 However,
Boyd and Richerson (1988) have shown that natural selection is unlikely to favor
this kind of trait (for critical discussion, see Johnson, Price, & Takezawa, 2008).
Their model presents a dilemma. On the one hand, if altruists were to benefit
others only when every other member of the group they belong to behaves
altruistically, they would be unlikely to ever behave altruistically when this
group is large and they would not reap the benefits of long‐term cooperation.
Their fitness would then not be higher than the fitness of non‐altruists. On the
other hand, if altruists were to behave altruistically when most (in contrast to
all) members of their group behave altruistically, (p.29) they would then
behave altruistically even when their group included some non‐altruists. The
fitness of altruists would then be lower than the fitness of non‐altruists. Thus, on
both horns of this dilemma, altruists have a lower fitness than non‐altruists.
Consequently, reciprocal altruism cannot explain the evolution of altruism in
large groups.

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-153

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-figureGroup-3

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-15

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-90

Evolution of Moralitys 1

Page 23 of 46

Second, reciprocal altruism and indirect reciprocity are designed to explain the
evolution of altruism (in a biological sense: acting in a way that increases others’
fitness) while many moral norms have little to do with altruism because they do
not regulate social interactions. For example, there are various apparently moral
prohibitions against eating particular types of food. But it is very unclear how
one can extend these two evolutionary mechanisms to account for the evolution
of moral norms—like food taboos—that are not related to altruism.33

Third, neither reciprocal altruism nor indirect reciprocity can account for the
link between morality and punishment. If reciprocal altruism or indirect
reciprocity really explained the evolution of morality, people would be disposed
to exclude from cooperation agents who violate moral norms. However, when an
agent commits morally reprehensible actions, rather than merely terminating
cooperation with this agent, people are disposed to punish her.34 A large body of
evidence in psychology and in behavioral economics highlights the link between
punishment and morality. For instance, Haidt and Sabini (2000) showed subjects
several videos of unjust actions. When asked to decide between various endings,
subjects were more satisfied when the agent suffered for her action than when
the victim pardoned the agent (see Fehr & Fischbacher, 2004 for consistent
behavioral evidence).

To summarize, because the main adaptationist models of the evolution of
morality appeal to direct or indirect reciprocity, they seem badly tailored to
account for three key properties of moral cognition: moral norms do not
exclusively (nor even primarily) bear on pairwise interactions; many moral
norms have nothing to do with altruism; and violations of norms are punished.
(p.30) For these three reasons, we view these models with skepticism. Their
existence provides little support to the claim that moral cognition evolved.

3.3. Distinctive Properties of Evolved Traits

In addition to providing adaptationist models, it is also common to argue for the
evolution of a trait by showing that this trait possesses some properties that are
distinctive either of evolved traits or of adaptations. Here we focus successively
on three alleged properties: moral norms are a cultural universal; complex moral
norms are acquired in spite of impoverished stimuli; and very young children
reason about norms and have harm‐related emotions.

3.3.1. Universality of Moral Norms

It is often said that there is no culture without a system of moral norms. Joyce
writes (2006: 134): “[m]orality (by which I mean the tendency to make moral
judgments) exists in all human societies we have ever heard of.” Similarly,
Hauser contends that “[t]he principles [of the moral faculty] constitute the
universal moral grammar, a signature of the species” (2006: 53). And the
universality of moral norms is taken to be evidence that it evolved.35

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-76

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-55

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-93

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-81

Evolution of Moralitys 1

Page 24 of 46

To evaluate this argument properly, it is important to keep in mind that the claim
that morality is present in every culture is not just the assertion that one finds
norms in every culture. Rather, it asserts that in every culture, one finds norms
that possess the properties that distinguish moral norms from other kinds of
norms according to the characterization used by a given evolutionary
researcher. Thus, when Joyce asserts that morality is present in every culture, he
is claiming that in every known culture, one finds norms that have most of the
seven properties he uses to single out moral norms from other kinds of norms
(see above).

But why should we believe that moral norms are present in all known cultures?
Researchers often bluntly assert this claim (e.g. Dwyer, 2006: 237) or illustrate it
with ancient or exotic codes of norms. Thus Joyce refers to the norms in the
Egyptian Book of the Dead and in the Mesopotamian epic of Gilgamesh (2006:
134–135). We find this casual use of the anthropological and historical literature
problematic. The problem is not so much that these researchers have not
examined a large number of cultures and historical periods to substantiate the
claim that moral norms are ancient and pancultural. Rather, the problem is that
because they fail to clearly distinguish norms from moral (p.31) norms, the
evidence they allude to merely supports the claim that norms are universal, but
not the much more controversial claim that moral norms are universal (see
Stich, 2008 for a similar point).

What anthropology and history show beyond reasonable doubt is that all known
cultures have norms. In all cultures, some actions are prohibited while others
are mandatory and some character traits are disvalued while others valued.
Sanctions and rewards for behaviors and personal attributes are good
candidates for being cultural universals. But because moral norms are conceived
as a distinct type of norm and moral judgments as a distinct kind of normative
judgment, the universality of norms should not be confused with the universality
of moral norms. For a given researcher (e.g. Hauser, Dwyer or Joyce) to support
the hypothesis that moral norms are universal thus requires far more than citing
exotic and ancient codes. It requires him or her to show that such codes amount
to an expression of morality, that is, to show that the norms in these codes
possess the properties that distinguish moral norms from other kinds of norms
according to the researcher’s rich characterization of moral norms and
judgments. However, to our knowledge, the relevant research has not been
done. Furthermore, the richer the characterization of moral norms and
judgments is, the less likely it is that norms in other cultures will count as moral
norms and thus that moral norms will be a universal.

Let’s illustrate this point with an example. In the sixth century, the Catholic
Church prohibited Christians from being buried with their wealth and it
recommended (but did not require) that part of one’s wealth be given to the
Church (Duby, 1996: 56–58). This is a clear example of a norm. But the existence

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-53

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-93

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-155

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-50

Evolution of Moralitys 1

Page 25 of 46

of this norm in Europe fifteen centuries ago provides no support whatsoever for
the hypothesis that moral norms are universal since it is unclear whether it is a
moral norm—that is, since it is unclear whether it possesses the properties that
distinguish moral norms. Instead of merely noting that this norm exists, Joyce
and other researchers would have to show that it possesses the properties that
(for them) distinguish moral cognition from other kinds of normative cognition.
This is a very difficult task, and it is far from obvious whether the norms found in
other cultures possess the properties that characterize moral norms according
to their rich characterization.36 (p.32)

What might explain philosophers’, psychologists’, and anthropologists’ confusion
between the universality of norms and the universality of moral norms is the fact
that they find in various cultures and times some norms that are somewhat
similar to the norms they themselves view as moral. For instance, many cultures
have norms against harming others, such as prohibition against in‐group harm.
But this fact does not show that all cultures have a system of moral norms, for,
again, it is unclear whether these norms are moral norms in all the cultures in
which they hold.

3.3.2. The Moral/Conventional Distinction

The second piece of evidence adduced to support the hypothesis that morality
evolved relies on the research on the moral/conventional distinction by
developmental psychologist Elliot Turiel (Dwyer, 2006: 239–242; Joyce, 2006:
134–137; Hauser, 2006: 291).37 In substance, Turiel and colleagues argue that
very early on, and panculturally, children distinguish two types of norms, called
“moral norms” and “conventional norms.” Moral norms are those norms that are
judged to hold independently from the authority of any individual or institution,
that are judged to be universally applicable, that are justified by appeal to the
harm done to others, to their rights, or to justice, and whose violations are
judged to be serious. Conventional norms are those norms whose force depends
on authority, that are judged to be only locally applicable, that are justified by
reference to convention, and whose violations are judged to be less serious than
the violations of moral norms.

Joyce concludes that “[t]hese results from developmental psychology strongly
suggest that the tendency to make moral judgments is innate” (2006: 137). The
argument from the universality and early development of the so‐called moral/
conventional distinction to the evolution of morality is best viewed as a poverty
of the stimulus argument (Dwyer, 1999, 2006; Mikhail, 2000). According to this
type of argument, developed most famously by Chomsky (1975), the fact that a
trait, such as the capacity to speak a language, develops reliably, while the
environmental stimuli are variable and impoverished, is evidence that this trait
is innate (for discussion, see Cowie, 1999; Laurence & Margolis, 2001; Pullum &
Scholz, 2002). Innateness is then often taken to be evidence that the trait under
consideration is the product of evolution or sometimes that it is an adaptation.38

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-53

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-93

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-81

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-93

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-52

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-53

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-114

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-32

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-39

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-104

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-125

Evolution of Moralitys 1

Page 26 of 46

It is tempting to apply this form of argument (p.33) to the distinction between
moral norms and conventional norms. Turiel and others have argued that even
very young children grasp the distinction between moral norms and other kinds
of norms. It is dubious whether young children could have learned this
distinction because the evidence needed to learn is often missing and because,
when it is not missing, it is unreliable. For instance, Dwyer notes that caregivers’
reactions to norms violations are unlikely to distinguish the two types of norms
because, as she puts it (2006: 240), “(s)ome parents get just as hot under the
collar about conventional transgressions as they do about moral transgressions.”
Furthermore, explicit moral instruction would not distinguish between different
kinds of norms because moral norms are not linguistically distinguished from
other norms: consider, e.g., “You ought to put your fork on the left of your plate”
and “You ought to keep your promises.” One might then conclude that the
distinction between moral norms and conventional norms is innate.

The poverty of the stimulus argument about the so‐called moral/conventional
distinction is unsound. The research on this distinction has a long and
respectable history, and it is received wisdom in much of contemporary
psychology. Recently, however, a growing body of evidence has emerged that
challenges the research tradition on the distinction between moral and
conventional norms in psychology (Gabennesch, 1990; Haidt et al., 1993; Kelly et
al., 2007).39

First, as Gabennesch (1990) has convincingly argued, the common wisdom—
endorsed by Dwyer and others—that very early on, children view some norms as
social conventions is poorly supported by the evidence. Carter and Patterson
(1982) found that half of their second‐ and fourth‐grader subjects judged that
table manners (e.g. eating with one’s fingers) were not variable across cultures
and that they were authority‐independent. Similarly, Shweder and colleagues
(1987: 35) concluded that among American children under 10, “there [was] not a
single practice in [their] study that is viewed predominantly in conventional
terms” (see Gabennesch, 1990 for many other references). Because many
children do not understand early on that some norms are conventional, the
poverty of the stimulus argument about the moral/conventional distinction
mischaracterizes the nature of the moral knowledge of young children.
Furthermore, because people come to understand slowly and rather late (p.34)
in their adolescence that some norms are mere social conventions, the amount
of evidence children and adolescents might rely on to come to understand this
distinction (whatever it amounts to—see below) is much less impoverished than
is assumed by the poverty of the stimulus argument.

More important, the research on the so‐called moral/conventional distinction
assumes that being authority‐independent, being universal, being associated
with serious violations, and being justified by appeal to harm, justice, or rights
form a cluster of co‐occurring properties (Kelly et al., 2007). That is, it is

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-53

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-68

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-77

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-96

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-68

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-29

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-141

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-68

https://oxford.universitypressscholarship.com/view/10.1093/acprof:oso/9780199582143.001.0001/acprof-9780199582143-chapter-2#acprof-9780199582143-bibItem-96

Evolution of Moralitys 1

Page 27 of 46

assumed that when a norm is judged to be authority‐independent, it is also
judged to be universal, it is justified by appeal to harm, justice, or rights, and the
violations of this norm are judged to be serious. By contrast, when a norm is
judged to be authority‐dependent, it is not judged to be generalizable to other
cultures, it is justified by appeal to conventions, and the violations of this norm
are judged to be less serious. However, research shows that this assumption is
unsubstantiated. People judge some actions (e.g. having sex with a dead chicken
before eating it or cleaning the bathroom with the national flag) to be serious
and authority‐independent, but do not justify the relevant norms by appeal to
harm, justice, or rights (Haidt et al., 1993). People are also sometimes reluctant
to generalize to other cultures norms that are justified by appeal to harms, and
they sometimes view these norms as authority‐dependent (Kelly et al., 2007; but
see Sousa, 2009; Sousa, Holbrook, & Piazza, 2009 for discussion). It thus
appears that the four properties assumed to set apart moral norms may not form
a cluster of co‐occurring properties at all.40 If this is the case, it is unclear what
the claim that early on children distinguish moral norms from conventional
norms amounts to, putting into jeopardy the poverty of the stimulus argument
about the moral/conventional distinction.

3.3.3. Developmental Evidence

In addition to the two alleged pieces of evidence discussed above (the presence
of moral norms in every culture and the early development of the moral/
conventional distinction), other aspects of human psychological development
have been mentioned as evidence for the evolution of morality. However, as we
now argue, they do not constitute evidence that moral cognition, understood
again as a specific kind of normative cognition, evolved, rather than evidence
that normative cognition evolved.

We have seen in Section 2 that children understand deontic conditionals, such as
“It’s not safe outside for the squeaky mouse, so all squeaky mice must (p.35)
stay in the house” much earlier and much better than indicative conditionals,
such as “It’s not safe outside for the squeaky mouse, so all squeaky mice are in
the house” (Cummins, 1996a; Harris & Núñez, 1996). One might be tempted to
argue that the capacity to identify the violation of deontic conditionals early on
provides evidence for the evolution of morality (Joyce, 2006; Dwyer, 2006).

This is certainly a very interesting finding, one that suggests that normative
cognition might be the product of evolution by natural selection (as noted in
Section 2). However, it is unclear how this finding is supposed to support the
idea that moral cognition proper, understood as a specific kind of normative
cognition, rather than normative cognition in general, evolved, since the norms
in the stories presented to children by Cummins and by Harris and Núñez were
not moral. More generally, deontic conditionals are not exclusively used in
specifically moral reasoning.