Paper 1: Study One Literature Review

alison12

CharityandanonimityFrietas.pdf

Home >Psychology homework help >Paper 1: Study One Literature Review

Maimonides’ Ladder: States of Mutual Knowledge and the Perception of Charitability

Julian De Freitas Harvard University

Peter DeScioli Stony Brook University

Kyle A. Thomas MotiveMetrics, Palo Alto, California

Steven Pinker Harvard University

Why do people esteem anonymous charitable giving? We connect normative theories of charitability (captured in Maimonides’ Ladder of Charity) with evolutionary theories of partner choice to test predictions on how attributions of charitability are affected by states of knowledge: whether the identity of the donor or of the beneficiary is revealed to the other. Consistent with the theories, in Experiments 1–2 participants judged a double-blind gift as more charitable than one to a revealed beneficiary, which in turn was judged as more charitable than one from a revealed donor. We also found one exception: Participants judged a donor who revealed only himself as slightly less, rather than more, charitable than one who revealed both identities. Experiment 3 explains the exception as a reaction to the donor’s perceived sense of superiority and disinterest in a social relationship. Experiment 4 found that donors were judged as more charitable when the gift was shared knowledge (each aware of the other’s identity, but unsure of the other’s awareness) than when it was common knowledge (awareness of awareness). Experiment 5, which titrated anonymity against donation size, found that not even a hundredfold larger gift could compensate for the disapproval elicited by a donor revealing his identity. Experiment 6 showed that participants’ judgments of charitability flip depending on whose perspective they take: Observers disapprove of donations that they would prefer as beneficiaries. Together, these experiments provide insight into why people care about how a donor gives, not just how much.

Keywords: charity, partner choice, reciprocity, common knowledge

Supplemental materials: http://dx.doi.org/10.1037/xge0000507.supp

People tend to judge donors who give anonymously as more charitable and generous than those who give publicly. This con- ventional wisdom is the basis for an episode of the TV comedy Curb Your Enthusiasm in which Larry David donates money for a wing of a nonprofit building that is named after him and is chagrined to find that his rival, Ted Danson, donated money for the other wing anonymously while his identity leaked out, paradoxi- cally reaping him the reputational advantages of both the gift and the anonymity. In experiments (Critcher & Dunning, 2011; Lin- Healy & Small, 2013; Newman & Cain, 2014) and in real life,

people often criticize donors who seek too much credit for their beneficence, as seen in the outrage directed at two philanthropists who rescinded a $3 million gift to a zoo because the plaque showing their names was too small (Dunlap, 1997).

The perceived merit of anonymous gifts is more than an abstract issue of normative ethical judgment. In modern times, charitable institutions are increasingly charged with solving some of the world’s most complex humanitarian problems, including hunger, disease, natural disasters, economic development, and political instability. A society’s collective choices about which gifts to praise, reward, and encourage can affect the flow of resources to these urgent problems. For instance, Dan Pallotta organized fund- raising events like AIDSRides and Breast Cancer-3-Days that raised $305 million for charities. But his companies collapsed after complaints that they earned a profit. Pallotta said, wistfully, “Peo- ple continue to die as a result . . . This we call morality” (Pallotta, 2009).

Why do people care so much about a donor’s anonymity, recognition, or ulterior benefits? None of this has anything to do with how much a donation improves beneficiaries’ well-being. Why not embrace donors’ desires for recognition as a win–win opportunity to increase charitable giving?

This article was published Online First October 18, 2018. Julian De Freitas, Department of Psychology, Harvard University; Peter

DeScioli, Department of Political Science, Stony Brook University; Kyle A. Thomas, MotiveMetrics, Palo Alto, California; Steven Pinker, Depart- ment of Psychology, Harvard University.

For helpful suggestions, we thank Kurt Gray and Jason Nemirow. Correspondence concerning this article should be addressed to Julian

De Freitas, Department of Psychology, Harvard University, William James Hall 964, 33 Kirkland Street, Cambridge, MA 02138. E-mail: [email protected]

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

158

A Ladder of Charity

Deepening the puzzle is the fact that people not only distinguish anonymous from public gifts, but also appear to make finer dis- tinctions based on the mutual knowledge of the donor and bene- ficiary. A famous example comes from the 12th-century Jewish philosopher Maimonides, who laid out a ladder of charitable giving (tzedakah, literally “righteousness”). Maimonides put double-blind gifts high on the ladder and common-knowledge gifts near the bottom, interspersed with other rungs based on the donor’s motivation and the beneficiary’s benefit over the long-term. The Ladder, from most to least charitable, is laid out as follows:

1. A donation that enables the beneficiary to escape the need for charity altogether (e.g., giving a gift or interest- free loan to start a business).

2. A double-blind donation (e.g., secretly leaving a gift in a courtyard where the poor can privately retrieve it without revealing themselves).

3. An anonymous donation to a known beneficiary (e.g., leaving a gift on their doorstep).

4. A revealed donation to an unknown beneficiary (e.g., the donor walks and drops money behind them for benefi- ciaries to pick up unseen).

5. A public donation that is given spontaneously (e.g., giv- ing money in person).

6. A public donation that is solicited (e.g., granting a re- quest for money).

7. A willing but inadequate donation.

8. A grudging donation, motivated by pity or guilt.

Maimonides discussed two factors behind the Ladder (Mai- monides, 1170/1180): (a) doing good deeds for their own sake rather than for praise or rewards, and (b) minimizing the embar- rassment felt by the beneficiary. To illustrate doing charity for its own sake, Maimonides discussed an example in which donors left gifts for the needy in a secret chamber of a temple where benefi- ciaries could retrieve them in secrecy (Rung 2, double-blind). To illustrate minimizing embarrassment, Maimonides discussed an example in which sages wrapped coins in a scarf slung over their back so that the poor could pick them out without feeling embar- rassed (Rung 4, revealed donor/unknown beneficiary).

From Maimonides’s time to the present, the knowledge and motivation of the donor have figured prominently in normative theories of the inherent morality of charitable giving. In a 2018 article entitled True Generosity Involves More Than Just Giving, for example, the philosopher Christian Miller appears to channel his medieval predecessor by arguing that “virtues such as gener- osity are complicated. They involve more than just outward be- havior. A person’s underlying thoughts, feelings and motives matter, too. If those aren’t in good shape, then one cannot qualify as a generous person” (Miller, 2018a; see also Miller, 2018b). Yet Miller is aware of the unresolved paradox that these inner feelings, however salient they are to our moral judgments, do not actually

feed the hungry or heal the sick, and notes “Generosity is a neglected virtue in academic research in general, and perhaps most of all in philosophy. There have been very few articles on gener- osity in mainstream philosophy journals since 1975.” (Miller, 2018a).

His observation is true of mainstream psychology as well. Here we attempt to fill this gap by examining whether laypeople really do rank different kinds of charitable giving in a hierarchy of states of knowledge like Maimonides’ ladder, which we take as the best articulated explication of nonutilitarian factors in ascriptions of charitability. Though we make no normative claims about which acts truly deserve to be called charitable or righteous, we seek to explain people’s intuitions about charitability using evolutionary theories of partner choice and cooperation, which are the most explicit modern scientific explanations of the psychology of gen- erosity.

Reciprocity, Partner Choice, and Judgments of Charity

To an evolutionary biologist, charitable donations are a kind of altruistic behavior, defined as instances in which one organism pays a cost to benefit another (Hamilton, 1996; Trivers, 1971, 1985; Wilson, 1975). The problem of altruism has been one of the central issues in evolutionary theory since the 1960s, and was a major topic in classics such as Richard Dawkins’ The Selfish Gene (Dawkins, 1976; Williams, 1966). The reason is obvious: Natural selection would seem to disfavor acts that benefit another organ- ism at a cost to the actor, yet humans engage in many such acts of altruism, of which the quintessential example is charitable giving.

The standard explanation was offered in Robert Trivers’ (1971) groundbreaking article on reciprocal altruism. In direct reciprocity, two cooperative partners can each benefit by exchanging favors, such as grooming each other, alternating child care, or trading surplus goods such as wool for milk (Trivers, 1971). Similarly, altruism can be favored by indirect reciprocity, in which cooper- ation enhances an individual’s reputation, bringing favors from others in the future (Hardy & Van Vugt, 2006; Nowak & Sigmund, 2005; Wu, Balliet, & Van Lange, 2016). But these theories predict that any tendency toward altruistic giving should coevolve with a desire to make the gift public. This way, beneficiaries and third parties can easily track who has given, ensuring that giving pays off in the long run by being reciprocated. Yet this appears to be the exact opposite of the esteem people grant to anonymous donors.

We propose that the resolution to this paradox lies in an impor- tant corollary to Trivers’ theory which he called the problem of “subtle cheating,” but which is referred to in recent literature as partner choice. When people face a set of possible cooperative partners who vary in their disposition for generosity, there will be complementary selection pressures for observers to discern those with the greatest dispositional generosity and for potential partners to advertise honest signals of such generosity (Trivers, 1971). This creates a kind of market in which people compete for the best cooperative partners (Barclay, 2016; Baumard, André, & Sperber, 2013; Noë & Hammerstein, 1995; Trivers, 1971). Among a set of potential cooperators in a community, some are more generous than others and hence make for more profitable partners. Specif- ically, partners vary in how small a profit they are willing to settle for in an exchange, and in their willingness to incur short-term

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

159ATTRIBUTIONS OF CHARITABILITY

losses to cultivate the relationship over the long term (Trivers, 1971). In the search for good partners, an individual is aided by accurately assessing others’ underlying generosities, whereas in- accurately assessing others leads to inefficiencies such as sharing precious resources with someone who calculatingly reciprocates the bare minimum (this is what Trivers called “subtle cheating”).

The fundamental problem of partner choice may have shaped our judgments of altruism, leading people to judge as most char- itable those who deliver the largest benefits and require the least in return. This could elevate donors who give anonymously: Because anonymous donors provide benefits with little expectation of fa- vors in return, they show a generous disposition that makes them desirable and profitable cooperative partners. Of course, this en- tails the paradoxical circumstance in which the identity of an “anonymous” donor is known to at least some observers, as in the Curb Your Enthusiasm episode, and the prevalence of such cir- cumstances has long been noted, as in the quotation attributed to Oscar Wilde “The nicest feeling in the world is to do a good deed anonymously—and have somebody find out.” Public donors have also displayed generosity, but because they could gain reputation and future favors, it is not clear that they would be as generous when they have little to gain themselves.

The challenge of choosing cooperative partners is closely re- lated to the classic problem of attribution from social psychology (Gilbert & Malone, 1995; Kelley, 1967). When people observe how other individuals behave, they attribute the behavior in vary- ing degrees to the actor’s underlying disposition, to the external circumstances, or to an interaction between them. Partner choice theory predicts that this process will be particularly engaged when observers learn of actors giving resources to others, which prompts them to assess the donor’s underlying generous disposition in order to evaluate them as a potential cooperative partner. (Note that this presupposes that people meaningfully vary in charitable disposition, as opposed to generosity being determined completely by the situation or beneficiary.) Importantly, the idea that charita- bility is an attribution problem predicts that people will prioritize certain kinds of cues as reliably indicating a generous disposition. How a donor gives is a particularly informative cue, potentially even more informative than the amount of the gift, because it indicates the dispositions that prompted the donor. Individuals who are skilled “judges of character” can leverage their mental-state reasoning to attribute dispositional generosity more accurately, allowing them to choose the most rewarding partners. In short, the theory of reciprocal altruism with its partner choice corollary can explain why people attribute greater generosity to donors who provide less knowledge of their gifts.

The partner choice hypothesis for judgments of charitability is distinct from several major alternatives. First, as mentioned, tra- ditional reciprocity theories are rooted in reputation, which pro- motes cooperation rather than diminishing it, so these models straightforwardly predict that people will prefer transparent and public donations to help them keep track of cooperators—which is flatly at odds with people’s praise for anonymous donors. Second, although classic attribution theories address how people infer dispositions (Gilbert & Malone, 1995; Kelley, 1967), they do not specifically predict that these inferences are attuned to the cues that inform people’s cooperative strategies.

Finally, an evolutionary approach goes beyond simple rational choice accounts by grounding psychological theories in the evo-

lutionary processes that shaped human social behavior, whether apparently rational or irrational (Cosmides & Tooby, 1994). A simple rational choice theory predicts, for instance, that people should lack strong feelings about the charitability of celebrities they’ll never meet, and that they should judge acts of charity based on objective benefits bestowed upon the beneficiary. Indeed, a rational utilitarian should prefer public donations, because these incentivize donors to give more, maximizing the benefit to the needy. Evolutionary theories, in contrast, hold that our cognitive and emotional faculties evolved in small-scale societies in which personal favor-trading was essential to fitness and any individual who was seen or mentioned by others was likely to be encountered face-to-face at some point in the future (Krasnow, Delton, Tooby, & Cosmides, 2013). This can explain why humans today are obsessed with judging people’s characters, even of people they will never meet, and even when the person’s character is irrelevant to their stated goal, in this case philanthropy aimed at alleviating hunger, poverty, or disease.

In short, judgments of charitability might be shaped by psycho- logical systems for choosing the best cooperative partners (Curry & Chesters, 2012; Tooby & Cosmides, 1996; Trivers, 1971). This theory predicts that people are equipped with cognitive systems for detecting and keeping track of cues that indicate a disposition for generosity, including the way a donor gives.

The Partner Choice Hypothesis and States of Knowledge

We hypothesize that people’s judgments about a donor’s chari- tability indeed fall into a hierarchy, as Maimonides suggested, because our intuitions of righteousness are related to assessments of cooperative partners. Specifically, a subset of the rungs of the Ladder of Charity (those based on states of knowledge, the focus of this article) may be related to the theory of partner choice as follows:

● Rung 2: Anonymous Donor, Unknown Beneficiary

Double-blind giving is the most diagnostic evidence of dispo- sitional generosity, because the donor has removed the possibility of receiving return favors altogether.

● Rung 3: Anonymous Donor, Known Beneficiary

If the donor knows the beneficiary but not vice versa (e.g., a gift left on a doorstep), this relieves the beneficiary of an obligation to reciprocate directly. Furthermore, the donor forgoes any reputa- tional benefit that might facilitate indirect reciprocation from third parties. However, unlike Rung 2, such a donor could potentially reveal himself or herself to the beneficiary later to try to call in a favor. This vulnerability may be experienced by the beneficiary in negative emotions like guilt, obligation, and, if made public, lowliness or shame.

● Rung 4: Public Donor, Unknown Beneficiary

Such a donor (e.g., one who carries coins in a backpack for the poor to pick out) cannot obligate the beneficiary to reciprocate directly, although the beneficiary could volunteer to do so. Impor- tantly, the beneficiary could tell others about the donor’s good

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

160 DE FREITAS, DESCIOLI, THOMAS, AND PINKER

deed and improve his or her reputation, leading to indirect reci- procity from others. The possibility of receiving both direct and indirect reciprocity makes this donor seem less charitable than a donor on Rung 3, who can benefit only from direct reciprocity.

● Rung 5: Public Donor, Known Beneficiary

When the donor and beneficiary know each other, as when one places money into the other’s hand, this creates an obligation for the beneficiary to pay back the donor should the donor need help in the future.1 This is because the donation is common knowledge: The donor and beneficiary know each other, they both know that they both know this, and so forth, ad infinitum, a knowledge state that enables coordination and has been shown to have many psychological effects (see, e.g., Chwe, 2001; Thomas, De Freitas, DeScioli, & Pinker, 2016; Thomas, DeScioli, Haque, & Pinker, 2014). In addition to the possibility of direct reciprocity, the donor may improve his or her reputation and thus receive benefits from indirect reciprocity.

In sum, the more that donors can expect to receive favors in return through direct or indirect reciprocity, the less charitable they will appear, because the donors’ gift becomes less diagnostic of their underlying disposition for altruism. This evolutionary analy- sis could explain why the manner of a gift affects people’s judg- ments even when it is irrelevant to the welfare of beneficiaries.

The Current Studies

In all the studies, we present participants with a donation sce- nario and ask them to judge the charitability of donors who choose to conceal their own identity, the beneficiary’s identity, or both. Experiments 1– 4 use these methods to test whether people’s judgments follow the order of the four rungs in the Ladder that differ in states of knowledge in ways identified as relevant by the partner choice hypothesis, particularly, in how they vary in whether knowledge of the act is absent, shared, or common. Across the experiments, we also examine personal and impersonal contexts and cases in which a donor has possible motives for direct and indirect reciprocity, addressing the affective considerations that enter into distinctions among other rungs (such as Rungs 7 and 8). In Experiment 5 we demonstrate the magnitude of the anonym- ity effect by titrating how much more money a revealed donor has to give than a double-blind donor in order to be judged as equally charitable. And in Experiment 6 we test for a moral hypocrisy effect (Batson & Thompson, 2001) by assessing how judgments of charitability differ between observers and participants in a chari- table act.

Experiment 1: The Giving Hierarchy in Impersonal Charity

Experiment 1 investigates the effect of levels of mutual knowl- edge in how people judge a donor in a typical modern context for charity: helping unknown victims from a distance. Participants read that donors could choose among four ways to give: double- blind anonymous, receiving a photo of the beneficiary, sending a photo of themselves to the beneficiary, or exchanging photos with the beneficiary. These knowledge states correspond to descending rungs of Maimonides’ ladder, and of ulterior motives for altruism

according to the partner choice hypothesis. If the ladder reflects attributions of charitability, people should judge the respective donors as most to least charitable accordingly.

Method

All experiments were approved by the Harvard University-Area Committee on the Use of Human Subjects. In all experiments, we recruited participants from the United States through Amazon’s Mechanical Turk (MTurk), the online labor crowdsourcing plat- form, to complete a survey for a small payment (for characteristics of MTurk participants, see Crump, McDonnell, & Gureckis, 2013). In all experiments, sample sizes were chosen in advance to provide at least enough power (.8) to detect medium effect sizes in com- parisons between means or percentages. Before data collection began, we decided to exclude participants who incorrectly an- swered any comprehension questions (provided in the online sup- plemental materials). Previous research found that excluding par- ticipants in online studies who fail comprehension checks can reduce noise in responses due to inattention (Thomas & Clifford, 2017). In Experiment 1, 39 participants failed at least one com- prehension question, yielding a final sample of n � 358.2

Participants read about how four donors decided to give money after receiving an advertisement:

A group of people all receive a letter from a charity company, asking them to make a $50 donation. The letter explains that their donation will go toward helping a poor family of hurricane victims in the local community. The company offers each person the following four donation choices:

(1) They can donate anonymously, such that the family will never know who the donor is, and the donor will never know who the family is.

(2) The donor receives an envelope containing a photo of the family they are donating to, and will thus know who the family is; but, the family will never know who the donor is.

(3) The family receives an envelope containing a photo of the donor, and will thus know who the donor is; but, the donor will never know who the family is.

(4) Both the donor and the family receive an envelope containing both a photo of the donor and a photo of the family. In other words, the donor will know who the family is and the family will know who the donor is, and both the donor and the family will know that they each know who the other is.

We designed these options to reflect four ways to give: double- blind; donor receives photo, that is, anonymous donor, known beneficiary; donor sends photo, that is, known donor and unknown beneficiary; and donor exchanges photos, that is, donor and ben- eficiary have common knowledge of the gift. These options cor- respond with Maimonides’ Rungs 2–5, respectively.

1 As in the opening scene of The Godfather, in which Vito Corleone says to a supplicant “Someday, and that day may never come, I’ll call upon you to do a service for me. But until that day, accept this justice as a gift on my daughter’s wedding day.” (Ruddy & Coppola, 1972).

2 Due to a technical error, we did not collect demographics for Experi- ments 1-4.

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

161ATTRIBUTIONS OF CHARITABILITY

Participants then read about four donors (Joe, Brian, Mike, and Charles) who each chose a different option; the donors were presented on separate pages in a random order. For each donor, participants rated: how charitable they thought the donor was, the donor’s likelihood of donating again, the beneficiary’s gratitude, and the beneficiary’s embarrassment about receiving the charity (the questions were always presented in the same order). For example: How genuinely charitable do you think Joe is? (0 � very selfish, 100 � very charitable). Finally, participants answered four multiple-choice comprehension questions and provided basic de- mographic information. We focus the analysis on the two items about the donor to test the main hypotheses about judgments of charity.

Thus, Experiment 1 has a 2 (Donor Receives Photo: yes, no) � 2 (Donor Sends Photo: yes, no) within-subject design, based on which photos were received and/or sent by the donor.3 We first use a 2 � 2 ANOVA to analyze whether these forms of knowledge affect participants’ perceptions of charity. Then, we conduct planned comparisons between adjacent levels of the charitability ladder, which postulates four levels of increasingly diagnostic cues of dispositional generosity. Specifically, we compare double-blind versus receive photo, receive photo versus send photo, and send photo versus exchange photos. (For similar analyses of knowledge levels, see Thomas et al., 2016; Thomas et al., 2014.)

Results and Discussion

Table 1 reports 2 � 2 ANOVAs showing a significant interac- tion and a main effect of send photo for both measures of chari- tability. We follow these up with planned comparisons for each measure (see Table 2).

Charitability. Participants rated the donor who gave double- blind as more charitable than the donor who received the benefi- ciary’s photo. They rated the donor who received the beneficiary’s photo as more charitable than the donor who sent his photo. However, they rated the donor who sent his photo as less charitable than the donor who exchanged photos (Figure 1 and Table 2)—an exception to the ordering of the charitability ladder, which the participants otherwise followed.

Likelihood of donating again. Participants rated the donor who gave double-blind as more likely to donate again than the donor who received the beneficiary’s photo. They rated the donor who received the beneficiary’s photo as more likely to donate again than the donor who sent his photo. However, paralleling the charitability ratings, they rated the donor who sent his photo as less likely to donate again than the donor who exchanged photos (Figure 1 and Table 2). Once again, these results follow the ordering of the charitability ladder with the exception that ex- changing photos and thereby generating common knowledge was perceived as more charitable than the donor sending his photo and thereby revealing his identity unilaterally.

For this experiment and Experiments 2 and 4, we report data on ratings of the beneficiary’s gratitude and embarrassment in the online supplemental materials.

These results confirm that evaluations of a donor’s charitability depend on the state of knowledge that the donor opts for when making a donation. Donors who gave double-blind were rated as more charitable than donors who received a photo, who in turn were rated as much more charitable than donors who sent their photo. In general, a donor forfeited the greatest impression of charitability by sending a photo. This is consistent with the hy- pothesis that a donor who reveals his or her identity comes across as motivated, at least in part, by reputation.

However, donors who opted to send their photos but not receive one from the beneficiary were seen as slightly less charitable than those who exchanged photos, violating the order of the charitabil- ity hierarchy. The unexpected finding that exchanging photos did not reduce the donor’s perceived charitability, as it might have if common knowledge obligated the beneficiary to reciprocate, sug- gests that some psychological factor in addition to direct and indirect reciprocity affects attributions of charitability. We will explore this factor in subsequent experiments.

Experiment 2: Personal Context

Experiment 1 presented scenarios involving donations to strang- ers. Yet such scenarios, which depend on a formal institution mediating the donation, is a far cry from the face-to-face commu- nities of our ancestral ecology. Such a discrepancy may explain the divergence between our results and the predictions laid out by the partner choice hierarchy (Hagen & Hammerstein, 2006; Krasnow et al., 2013). Experiment 2 investigated charitable donations in a more personal context, in which the donor and beneficiary live in the same community and know each other. An observer might be more sensitive to a donor’s potential motives for direct and indirect reciprocity when the donor gives to people with whom he or she regularly interacts in a local community, because the donor will have more opportunities to receive: (a) an exchange of favors from the people he helps (direct reciprocity); and (b) favors from third parties in the community who have heard of the donor’s good reputation (indirect reciprocity). If so, exchanging photos should especially be seen as obligating the beneficiary to repay the favor,

3 With this 2 � 2 design, we do not imply that exchanging photos is simply a combination of sending and receiving photos. Conceptually, the donor who exchanges photos creates common knowledge of the gift, which is a distinct category from the two kinds of private knowledge created when only one photo is shared.

Table 1 2 � 2 ANOVA for Ratings of Donor Charitability and Likelihood of Donating Again in Experiments 1–3

Experimental condition

Charitability Likelihood of donating again

F �p 2 F �p

E1 Impersonal Charity Receive photo 0.2 .00 0.2 .00 Send photo 141.0�� .09 70.8�� .05 Receive � Send (Exchange) 43.2�� .03 15.7�� .01

E2 Personal Charity Receive photo 4.1� .00 4.5�� .00 Send photo 200.0�� .14 196.5�� .14 Receive � Send (Exchange) 18.0�� .01 14.8�� .01

E3 Relationship and Status Receive photo 3.5 .01 4.4� .02 Send photo 67.3�� .19 31.4�� .10 Receive � Send (Exchange) 3.3 .01 0.0 .00

� p � .05. �� p � .001.

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

162 DE FREITAS, DESCIOLI, THOMAS, AND PINKER

and observers may now judge that the donor’s motives are mostly driven by reciprocity, bringing their judgments into complete, rather than partial, alignment with the charitability ladder and the partner choice hypothesis.

Method

After excluding 80 participants who made errors on the com- prehension questions, we end up with a sample of 307 participants. The procedure was the same as in Experiment 1 except for the following changes. The vignette described a personal setting in which everyone in the community already knew everyone else:

Ashborne is a small town where everyone knows each other. During one year, a dozen of the houses were seriously flooded. A townsper- son started collecting donations to help the families who were af- fected, and because all the families know each other, he offered donors several ways to give: . . .

We further clarified that “In all cases, the family is also in- formed about which option their donor has chosen,” (but not the donor’s identity, unless the donor chose to reveal it). In addition to the questions asked in Experiment 1 (donor charitability, donor’s likelihood of donating again, beneficiary gratitude, and beneficiary embarrassment), we added two questions at the end about the donor’s motivations by direct and indirect reciprocity, respec- tively: “How much do you think [name] was motivated to give by a desire to improve his reputation with the family? How much do you think [name] was motivated to give by a desire to improve his reputation with the broader community?” Finally, we added a comprehension item about the personal context, specifically

whether the community was a big city, medium-sized town, or small town.

Results and Discussion

Table 1 reports the 2 � 2 ANOVAs showing a significant interaction and main effects of send photo and receive photo for both donor measures. We follow these up with planned compari- sons for each measure.

The patterns of ratings of charitability and likelihood of donat- ing again were similar to those in Experiment 1, in which the context was impersonal (Figure 2, Table 2). We again found that judgments of charitability followed the ordering of the charitability ladder except that participants did not judge the donor who ex- changed photos as less charitable than the donor who only sent their own photo, even in this more personal context.

In this experiment, however, ratings of whether the donor was motivated by a desire to improve his reputation with the family and the broader community did correspond to all four levels of the charitability ladder (Figure 3, supplemental materials Table S4). Possible explanations for this discrepancy will be tested in Exper- iment 3.

Follow-Up Experiments 2a and 2b

In a pair of supplemental Experiments 2a and 2b (see online supplemental materials for full details), we use the same scenarios while adding that the donor had a potential motive for direct reciprocity (E2a) or indirect reciprocity (E2b). In the direct reci- procity case, the donor needed a favor from the beneficiary; in the

Table 2 Ratings of Donor Charitability and Likelihood of Donating Again, Experiments 1–5

Experimental condition

Charitability Likelihood of donating again

M SD t d M SD t d

E1 Impersonal Charity Double-blind 86.5 15.4 8.0�� .60 80.0 19.0 4.2�� .32 Receive photo 79.3 17.7 9.6�� .72 76.3 17.8 8.1�� .61 Send photo 67.7 22.8 6.1a�� .46 66.9 22.2 4.6a�� .35 Exchange photos 73.9 20.1 71.6 20.2

E2 Personal Charity Double-blind 89.5 13.5 8.9�� .72 86.0 16.0 8.1�� .66 Receive photo 82.7 17.1 11.1�� .90 79.1 17.5 11.1�� .90 Send photo 69.6 22.6 2.5a�� .21 65.6 23.4 2.0a� .16 Exchange photos 72.0 21.1 67.5 21.9

E3 Relationship and Status Double-blind 90.1 14.6 3.2�� .55 82.0 19.7 1.6 .27 Receive photo 81.8 15.3 4.7�� .80 77.1 17.7 2.6�� .44 Send photo 67.9 19.5 0.0 .01 68.5 21.0 1.5 .24 Exchange photos 67.6 22.6 62.9 25.1

E4 Shared Knowledge Versus Common Knowledge

Double-blind 88.0 14.1 2.7�� .50 83.1 15.7 1.9� .37 Optional photos 80.3 16.6 2.6�� .47 77.0 17.7 2.4�� .44 Exchange photos 71.3 21.6 68.2 21.9

Note. The t-statistics and effect sizes are from the tests comparing each condition with the one below it (hence, the fourth row is empty for each experiment). The tests were paired comparisons for Experiments 1 and 2, and independent-sample t-tests for Experiments 3– 4. Sample sizes for the tests were n � 358 and 307 for Experiments 1 and 2, respectively; Experiment 3: 64, 76, 66, 86; Experiment 4: 54, 66, 61. a Indicates that the difference is in the opposite direction from the charitability ladder. � p � .05. �� p � .01. �� p � .001.

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

163ATTRIBUTIONS OF CHARITABILITY

indirect reciprocity case, he needed a favor from the beneficiary’s friends. In addition, these studies had a between-subjects design. We tested whether, in light of explicitly stated self-serving mo- tives, participants judged exchanging photos as less charitable than sending a photo, as predicted by the charitability ladder.

Figure 4 shows the main results. Judgments of charitability followed the same pattern as before except the drop-off in chari- tability ratings for the receive-photo and exchange photos condi- tions was more pronounced than in Experiment 2. Most relevant here, participants continued to judge the donor who sent a photo as similar in charitability to the donor who exchanged photos, rather than as less charitable.

Experiment 3: What Revealed Identities Say About Relationship and Status

In Experiments 1 (impersonal context) and 2 (personal context), contrary to the ordering of the charitability ladder and the corre- sponding considerations from partner-choice theory, participants judged a donor who gave money with common knowledge as no less charitable than a donor who revealed only his own identity. Experiments 2a and 2b, which made the potential motives for direct or indirect reciprocation explicit, also found no reduction in charitability for giving with common knowledge. This experiment explores this divergence from what would seem like a natural ranking based on states of mutual knowledge. We suggest that when people assess charitability, they do not just estimate the gross probability of acts of reciprocation, whether they be direct or indirect; they also assess the nature of the social relationship that provides the context of the reciprocation.

Presumably, a donor who remains hidden (i.e., in the double- blind and receive photo conditions) is modest about his charita- bility and thus maximally charitable, whereas a donor who di- vulges his identity might have a self-serving reputational motive. When that donor also seeks to know the beneficiary (by exchang- ing photos), his motive may be even more self-serving (by know- ing who he can tap for a favor later), but it could also have a more benign interpretation: that he is interested in entering into an

100

Do ub

le- Bl

ind

Re ce

ive Re

cip ien

t P ho

Se nd

Do no

r P ho

Ex ch

an ge

Ph oto

M ea

n R

at in

gs Charitable Donate again

Figure 1. Mean ratings of the donor’s charitability and likelihood of donating again in Experiment 1. Error bars are standard errors. See the online article for the color version of this figure.

100

Do ub

le- Bl

ind

Re ce

ive Re

cip ien

t P ho

Se nd

Do no

r P ho

Ex ch

an ge

Ph oto

M ea

n R

at in

Charitable Donate again

Figure 2. Mean ratings of the donor’s charitability and likelihood of donating again in Experiment 2. Error bars are standard errors. See the online article for the color version of this figure.

100

Do ub

le- Bl

ind

Re ce

ive Re

cip ien

t P ho

Se nd

Do no

r P ho

Ex ch

an ge

Ph oto

M ea

n R

at in

Direct Reciprocity Indirect Reciprocity

Figure 3. Mean ratings of the extent to which the donor was motivated by direct and indirect reciprocity in Experiment 2. Error bars are standard errors. See the online article for the color version of this figure.

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

164 DE FREITAS, DESCIOLI, THOMAS, AND PINKER

ongoing and symmetrical social relationship. This relationship could take the form of a communal bond that serves as a context for spontaneous and unstinting exchange (as when Vito Corleone chided his beseecher “We’ve known own each other for many years, but this is the first time you came to me for help. I can’t remember the last time that you invited me to your house for a cup of coffee”). The idea that people can perceive favors not just as opening moves in a series of exchanges but as invitations to consummate an ongoing relationship is consistent with the com- mon anthropological findings that people in many cultures use gifts to form and cement relationships (van Baal, 1975), and that they distinguish communal relationships with spontaneous sharing from equality-matching relationships with calculated quid pro quo reciprocity (Fiske, 1992). Perhaps this possibility can offset the seemingly self-serving nature of a common-knowledge donation, with its implicit demand for reciprocation. Thus, the first hypoth- esis we test is whether observers infer that a donor who generates common knowledge by exchanging photos is seeking to initiate an ongoing cooperative relationship with the beneficiary.

A related hypothesis is that evaluators judge that donors who reveal only their own identities see themselves as higher in status than their beneficiaries, making their gift a mere costly signal of status with all its perquisites, like any other conspicuous flaunting of wealth, with no concern (indeed, perhaps even contempt) for the well-being of the beneficiaries. In contrast, a donor who exchanges photos may be seen as signaling that he sees himself as equal in status to the beneficiary, and that the gift was offered out of sheer beneficence, with no ulterior motive for dominance.

Either or both of these factors could explain why a donor whose gift is common knowledge is not judged worse than one who only reveals his identity. Here we test whether participants do in fact make these inferences, with a focus on the contrast between a

donor exchanging photos (common knowledge) and unilaterally revealing his identity.

Method

After excluding 82 participants who made comprehension er- rors, we arrived at a final sample of 292. The procedure was the same as in Experiment 2, with its personal context, except that the design was between-subjects, and there were two additional ques- tions about the protagonists’ relationship and status, both rated on a 100-point scale: Do you think that Brian would be interested in meeting the family and getting to know them in the future? (defi- nitely not interested to definitely interested); Do you think that Brian sees himself as equal in status to the family, or as higher in status? (definitely higher in status– definitely equal in status).

Results and Discussion

Charitability and likelihood of donating again. Table 1 reports the 2 � 2 ANOVAs for the charitability and likelihood of donating again measures. For charitability, we find a main effect of send photo and no interaction. For likelihood of donating again, we find no interaction and significant main effects of send photo and receive photo. We follow these up with planned comparisons for each measure. Perceptions of whether the donor was charitable and likely to donate again showed similar patterns of ratings to the previous studies (Figure 5 and Table 2): no lower with common knowledge than with a revealed donor, failing to conform to the ladder ordering, though in this case not reversing it outright.

Initiating a relationship. Table 3 reports the 2 � 2 ANOVAs for initiating a relationship, which shows significant main effects of receive photo and send photo and no interaction. We next look

100

Do ub

le- Bl

ind

Re ce

ive Re

cip ien

t P ho

Se nd

Do no

r P ho

Ex ch

an ge

Ph oto

M ea

n R

at in

100

Do ub

le- Bl

ind

Re ce

ive Re

cip ien

t P ho

Se nd

Do no

r P ho

Ex ch

an ge

Ph oto

Charitable Donate again Friends

Figure 4. Mean ratings of the donor’s charitability, likelihood of donating again, and desirability as a friend in (A) Experiment 2a (direct reciprocity) and (B) Experiment 2b (indirect reciprocity). Error bars are standard errors. See the online article for the color version of this figure.

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

165ATTRIBUTIONS OF CHARITABILITY

at the pattern of means. Perceptions of whether the donor was interested in initiating a relationship zigzagged across the levels: Donors who received the beneficiary’s photo or exchanged photos were viewed as most interested in initiating a relationship with the beneficiary (Figure 6, Table 4). This pattern, qualitatively different from what we have seen so far, suggests that a different psycho- logical mechanism is in play than the reciprocity calculations that have driven the other ratings. And, it may help explain the main surprise in those ratings: that participants did not judge exchanging photos as less charitable than sending a photo. Namely, if partic- ipants inferred that a photo-exchanging donor was interested in a long-term social relationship with the beneficiary, it could have

offset the negative inference that the donor was selfishly motivated by reciprocity, adding up (either across raters or across conflicting interpretations by each rater) to a more neutral evaluation of the donor’s charitability. A post hoc comparison reveals that the donor who exchanged photos was also viewed as more interested in getting to know the beneficiary than the donor who only received the beneficiary’s photo, presumably because establishing common knowledge shows greater willingness to strike up a relationship, t(160) � 2.41, p � .017, d � 0.38.

If this interpretation is correct, these considerations should be specific to the conditions in which the donor reveals himself by sending or exchanging photos, because only in these conditions can the donation be interpreted in either of the two ways, reciprocity-seeking or relationship-initiating. Indeed, this was the case: Judgments of relationship intentions and charitability were

100

Do ub

le- Bl

ind

Re ce

ive Re

cip ien

t P ho

Se nd

Do no

r P ho

Ex ch

an ge

Ph oto

M ea

n R

at in

gs Charitable Direct Reciprocity (Reversed) Equal Status

Get to Know Indirect Reciprocity (Reversed)

Figure 5. Mean ratings of the donor’s charitability, interest in a relation- ship, and perceived relative status in Experiment 3. We have also juxta- posed the perceived motives for direct and indirect reciprocity from Ex- periment 2 (personal context)—reverse-coded so that higher values indicate less motives for reciprocity—in order to illustrate how the various factors unveiled in Experiments 1–3 add up to produce favorable charita- bility judgments. Because likelihood of donating again mirrored charita- bility, we have omitted it from this already complicated plot. Error bars are standard errors. See the online article for the color version of this figure.

Table 3 2 � 2 ANOVAs for Ratings of Donor Interest in Initiating a Relationship and Perceived Status Relative to the Beneficiary in Experiment 3

Experimental condition

Initiating relationship

Status relative to beneficiary

F �p 2 F �p

Receive photo 46.5�� .14 .01 .00 Send photo 7.0�� .02 12.61�� .04 Exchange photos .5 .00 11.69�� .03

�� p � .01. �� p � .001.

100

Do ub

le- Bl

ind

Op tio

na l P

ho tos

Ex ch

an ge

Ph oto

M ea

n R

at in

g s

Charitable Donate again

Figure 6. Mean ratings of the donor’s charitability and likelihood of donating again in Experiment 4. Error bars are standard errors. See the online article for the color version of this figure.

Table 4 Ratings of Donor Interest in Initiating a Relationship and Perceived Status Relative to the Beneficiary, Experiment 3

Experimental condition

Initiating relationship Status relative to

beneficiary

M SD t d M SD t d

Double-blind 38.9 23.7 4.2�� .72 64.0 26.8 2.4� .41 Receive photo 56.8 25.6 2.9�� .49 53.2 25.5 2.8�� .48 Send photo 43.8 27.6 5.3�� .88 41.1 24.5 2.4� .39 Exchange photos 66.2 23.8 51.7 28.5

Note. The independent-sample t-statistics and effect sizes are from the tests comparing each condition with the one below it (hence, the fourth row is empty for each experiment). Sample sizes for each of the conditions were: 64, 76, 66, 86. � p � .05. �� p � .01. �� p � .001.

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

166 DE FREITAS, DESCIOLI, THOMAS, AND PINKER

significantly positively correlated when the donor sent his photo or exchanged photos, but not when the donor was double-blind or received the beneficiary’s photo: double-blind, r(62) � �.02, p � .896; receive beneficiary photo, r(74) � .06, p � .617; send donor photo, r(64) � .41, p � .001; exchange photos, r(84) � .65, p � .001.

Status relative to beneficiary. For status relative to benefi- ciary, the 2 � 2 ANOVA shows a significant interaction and main effect of send photo. Turning to the pattern of means, perceptions of whether the donor felt equal in status to the beneficiary followed the order of the charitability ladder with a familiar exception: Donors who exchanged photos were rated as acting more equal in status to their beneficiaries than donors who sent a photo (Figure 5, Table 4). These results may further explain why giving with common knowledge is viewed as no less charitable than giving with one’s identity revealed: Although the photo-exchanging do- nor may seek future favors, he at least seems more egalitarian than one who would seem to be lording his advantage over the bene- ficiary by only sending his photo.

We also looked at correlations across participants between judg- ments of charitability and judgments of status for each of the conditions: double-blind, r(62) � 0.24, p � .060; receive photo, r(74) � 0.11, p � .354; send photo, r(64) � 0.27, p � .028; exchange photos, r(84) � 0.33, p � .002. In the key conditions when the donor’s identity was revealed, participants who thought the donor felt more equal to the beneficiary also judged the donor as more charitable.

Taken together, the results suggest that when a donor reveals only his own identity, observers perceive that the donor projects his superiority over the beneficiary. In contrast, a donor who exchanges photos conveys a more egalitarian disposition toward beneficiaries and possibly a willingness to initiate an ongoing communal relationship with them. This may explain the main exception in charitability ratings from the charitability ladder, in which donations with common knowledge are seen as equally or more charitable than donations with only the donor’s identity revealed.

One way to understand this finding is to think back to Mai- monides’ own example of a donation in which the donor is known to the beneficiary but not vice versa: a well-off man carrying coins in a scarf behind his back for the poor to pick out. Though the donor cannot be accused of trying to secure a debt with the beneficiary, we still perceive such an action as haughty, high- handed, or demeaning, compared with the more socially (if not financially) egalitarian gesture of placing the money into the beneficiary’s hand.

Thus, judgments of charitability are driven not by the size of the gift alone, nor by the state of mutual knowledge (and hence opportunity for reciprocity) alone, but also by the status imbalance and the donor’s interest in initiating a communal relationship. This suggests that charitability is attributed relative to the context of distinct relationship types involving status and personal intimacy.

Experiment 4: Common Knowledge Versus Shared Knowledge

Theories of common knowledge emphasize its logical differ- ence from shared knowledge, which is any state of mutual knowl- edge that falls short of common knowledge (Chwe, 2001; Clark,

1996; Clark & Marshall, 1981; Rubinstein, 1989; Schelling, 1960; Smith, 1982). Pinker (2007), Lee and Pinker (2010), and Thomas, DeScioli, Haque, and Pinker (2014) propose that people are sen- sitive to this distinction whenever they must coordinate their behavior with each other. Common knowledge ratifies an under- standing that each will behave in a way that works to their joint benefit, including conforming to the expectations in stable rela- tionships such as intimates, friends, neighbors, superordinate and subordinate, or trading partners. Knowledge states short of com- mon knowledge allow for uncertainty or deniability, and thus enable breaches of these expectations.

Here we test whether common knowledge is also potent in attributions of charitability, as suggested by the fact that four out of the eight rungs in Maimonides’s Ladder are differentiated by states of mutual knowledge. Theoretically, a donor who gives with shared knowledge has imposed less of an obligation on the bene- ficiary to reciprocate, compared with a donor who gives with common knowledge. This is because with shared knowledge the donor and beneficiary are uncertain about whether they know each other, which creates ambiguity about the beneficiary’s indebted- ness to the donor. If observers are sensitive to this distinction, they should judge donors who give with only shared knowledge as more charitable than donors who give with common knowledge.

Method

After excluding 55 of 236 MTurk participants who made com- prehension errors, we ended up with a sample of 181. The proce- dure and scenario were adapted from Experiment 2 in a between- subjects design in which the donor was described as having chosen to give money in one of three ways:

1. ANONYMOUS: They can donate anonymously, such that the family will never know who the donor is, and the donor will never know who the family is.

2. KNOW EACH OTHER: Both the donor and the family receive an envelope containing both a photo of the donor and a photo of the family. In other words, the donor knows who the family is and the family knows who the donor is, and both the family and the donor know this.

3. OPTIONAL KNOW EACH OTHER: The donor and the family give their photos to the townsperson. The donor chooses whether or not to receive the family’s photo from the townsperson. Likewise, the family chooses whether or not to receive the donor’s photo. The townsperson will keep these choices confidential so neither will know if the other side received their photo.

The third alternative is a new one that represents shared knowl- edge, as opposed to the common knowledge created by exchanging photos directly. After being informed of the donor’s options, participants read which alternative he in fact chose. For the donor who chose Number 3, we further clarified how this created shared knowledge as follows:

Mike donated $50 to the family, and chose Option 3. That is, both he and the family gave their photos to the townsperson, and then they got to choose whether to see the other’s photo.

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

167ATTRIBUTIONS OF CHARITABILITY

As it happened, both Mike and the family ended up asking the townsperson to show them the other’s photo, and the townsperson kept their choices confidential. Therefore, both sides know who the other is, but they don’t know that the other side knows this.

Results

One-way ANOVAs showed a main effect of knowledge state on both judgments of charitability, F(1, 179) � 25.3, p � .001, �p2 � .12; and likelihood of donating again, F(1, 179) � 18.3, p � .001, �p

2 � .09. We conducted two planned comparisons: double-blind versus optional photos, and optional photos versus exchange pho- tos. For judgments of charitability we found that donors who gave double-blind were rated as more charitable than donors who gave with shared knowledge, who in turn were rated as more charitable than donors who gave with common knowledge (Figure 6, Table 2). We found a similar pattern for ratings of likelihood of donating again. This pattern supports the hypothesis that common knowledge, more than mere shared knowledge, ratifies the bene- ficiary’s reciprocal obligation, and as a consequence reduces at- tributions of generosity to the donor.

This finding indicates that people’s ascriptions of charitability are cognitively and strategically sophisticated. With both shared knowledge and common knowledge, the donor and beneficiary knew each other’s identities. The critical difference was that in the exchange photos condition they also knew that they both knew this, whereas in the optional photos condition they didn’t. Partic- ipants’ charitability judgments picked up on this subtle distinction.

More generally, this result supports a nuanced prediction of the partner choice hypothesis. Recall that a donor who gives with common knowledge (exchange photos) cements the beneficiary’s obligation to repay their favor, similar to the way that lenders use legal contracts to create common knowledge of the obligations between a creditor and debtor. In contrast, a donor who gives with shared knowledge (optional photos) imposes less of an obligation on the beneficiary because it is unclear whether each side knows the other. In this study, participants were sensitive to this crucial distinction, ascribing greater charitability to the donor who made the photos optional, even though both sides privately chose to retrieve the other’s photo. Moreover, this finding confirms a dis- tinctive prediction of partner choice that does not follow from major alternative accounts, such as traditional reciprocity or classic attribution theories that do not make reference to cooperative strategies.

Experiment 5: The Price of Reputation

In Experiment 5, we quantify how much participants weigh a donor’s anonymity, which does not materially affect a beneficia- ry’s well-being, relative to the amount of the donation, which does. Specifically, we asked them to compare the charitability of a donor who gave a double-blind donation of $10,000 with another who revealed his identity while donating a specific amount, which varied between participants. This titration method allowed us to zero in on the switch point at which a revealed donor’s gift is large enough to equal or exceed the charitability of a double-blind donor. We varied the amount of the revealed donation first in broad increments (e.g., 1x, 2x, 3x), and, after finding a switch point between 1x and 2x, we then varied the revealed donation in

finer increments that afforded higher resolution near the switch point (e.g., $12k, $14k, etc.).

We conduct separate titrations for three kinds of revealed donor: ones who send a photo, exchange photos, or deliver the gift in person. Technically, the latter two do not differ in level of knowl- edge: Both are common knowledge. But previous research argues that people infer common knowledge based on particular cues (Pinker, 2007; Thomas et al., 2016), and face-to-face contact is likely to be a particularly powerful cue that generates common- knowledge.

Method

In the send photo condition, we excluded 172 MTurk partici- pants who made comprehension errors (the questions are included in the online supplemental materials), yielding a final sample of 555 (Mage � 38; 46% female). In the exchange photos condition, we excluded 184 participants, yielding a final sample of 530 (Mage � 20; 54% female). In the in-person condition, we excluded 167, yielding a final sample of 475 (Mage � 38; 61% female).

Participants read about two donors, one who made a double- blind donation ($10,000), and another who revealed his identity while donating a specific amount, which varied in between- subjects conditions (specified below). In three separate titrations, we compared the double-blind donor with a donor who sent a photo, exchanged photos, or gave in person.

In the send photo condition, participants read:

Ashborne is small town where everyone knows each other. During one year, a few dozen families in a poor neighborhood are affected by a severe flood and become homeless. It costs $1000 per family to provide temporary food and shelter for a month while their homes are being repaired.

The local newspaper has asked the citizens of Ashborne to please help by mailing $1000 checks to the local municipality, who will then distribute these donations to the homeless families without revealing who these beneficiaries are.

One donor decides to help anonymously by sending 10 envelopes with no return address containing $1000 each.

A second donor sends [X] envelopes containing $1000 each to [X] families, and the envelopes also include the donor’s business card with his name and a photo of himself.

Participants then answered a forced-choice question:

Who do you think is a more charitable person? The donor who gives 10 donations of $1000 anonymously, or the donor who gives [X] donations of $1000 sent with the donor’s business card.

The exchange photos condition was introduced as follows: “Donors can drop off the checks at the shelter either by giving them to the shelter director to distribute to a family or by request- ing to share photos with each beneficiary, which means that both the donor and beneficiary receive an envelope with each other’s photos.” Then, it said that the second donor in fact chose this second option.

The in-person condition was introduced in the same way except that the second option was described as follows: “. . . or by handing them directly to the family in person.” Then, it said that the second

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

168 DE FREITAS, DESCIOLI, THOMAS, AND PINKER

donor had chosen this option. See the online supplemental mate- rials for the full vignettes and comprehension questions.

Results and Discussion

Table 5 and Figure 7 show the results. For the donor who sent his photo, we found, surprisingly, that there was no amount he could give to make himself appear as charitable as a double-blind donor—not even 100 times as much money. Participants’ evalua- tions thus showed a striking insensitivity to magnitude (Fether- stonhaugh, Slovic, Johnson, & Friedrich, 1997). To ensure that this extreme finding was not due to the materials being biased to suggest a crass commercial motive, we ran three follow-up studies in which we specified that he had the same income and assets as the double- blind donor and that he revealed his identity without using a business card (which could imply a direct solicitation). The pattern of ratings was the same (see online supplemental materials).

With the donor who exchanged photos, participants started to consistently favor him when he gave roughly four times as much money as the double-blind donor. With the donor who gave in person, participants began to consistently favor him when he gave twice as much as the double-blind donor. The fact that participants preferred the donor who gave in person over the donor who exchanged photos might be because an in-person donation indi- cates the donor’s interest in a forming an ongoing relationship with the beneficiary (similar to the findings from Experiment 3).

Finally, we compared perceptions of altruistic traits across the conditions. Participants generally viewed the donor who sent a photo as less charitable than the donor who exchanged photos or who gave in person (see Figure 7). For instance, if we focus on revealed donors who gave twice as much as the anonymous donor, we see that participants judged the one who sent his photo unfa- vorably compared to the one who exchanged photos (p � .008, Fisher’s exact test) or who gave in person (p � .001, Fisher’s exact test). This violation of the ordering of the charitability ladder, with common knowledge not being judged as least charitable, is con- sistent with the previous experiments.

Experiment 6: The Beneficiary’s Perspective

The previous experiments examined participants’ judgments from the perspective of an impartial third-person observer. We

now look at how participants view charity from the perspective of the beneficiary. Our previous results show that an observer’s judgments of charitability diverge from what is materially best for the beneficiary. Beneficiaries, whose first priority is not choosing among partners but improving their lives, may see things different, and prefer to receive more money than less money, while caring less about how the gift was given.

In particular, partner choice theory predicts that observers’ judgments of donors will not necessarily match the beneficiaries’ preferences, because they face different problems: The observer evaluates the donor’s generous disposition, while the beneficiary aims to evaluate disposition and maximize the immediate benefits he or she receives from the donor. This theory predicts a discrep- ancy in which observers praise more righteous donors while ben- eficiaries prefer greater benefits, caring less about righteousness. This mismatch could be quite consequential: It means that observ- ers could criticize and deter donations that would most help beneficiaries.

We examined this possibility by asking participants to put themselves in the shoes of the beneficiary and to indicate whether they preferred to receive a small gift from a double-blind donor or a larger gift from a donor who includes his photo with the check. Participants also judged which donor was more charitable. We compared these decisions with a control condition in which par- ticipants took the role of an observer who judged which donor was more charitable.

The key comparison is between what beneficiaries say they would prefer (larger gift or anonymous donor) versus who observ- ers judge as more charitable. This approach mirrors classic work on moral hypocrisy (Batson & Thompson, 2001; Batson, Thomp- son, Seuferling, Whitney, & Strongman, 1999), which examines discrepancies between what observers say that others should do versus what they choose to do themselves. For instance, this work found that participants say that others should flip a coin to fairly allocate a boring task to themselves or a partner, but when choos- ing themselves, participants do not actually flip the coin or abide by the outcome. We test for analogous discrepancies by comparing participants’ judgments in different roles (observer or beneficiary) for the same set of gifts with measures that are specific to each role (judging or receiving).

Table 5 Proportion Who Chose the Revealed Donor as More Charitable Than the Double-Blind Donor Who Gave $10,000

Send photo Exchange photos In person

Donation n % p Donation n % p Donation n % p

10,000 68 4 �.001 10,000 68 12 �.001 10,000 60 13 �.001 12,000 63 21 �.001 20,000 63 57 .314 12,000 62 55 .526 14,000 60 22 �.001 30,000 60 55 .519 14,000 51 61 .161 16,000 62 29 .001 32,000 62 63 .056 16,000 57 67 .016 18,000 62 34 .015 34,000 53 58 .272 18,000 61 59 .200 20,000 66 33 .009 36,000 59 69 .004 20,000 57 65 .033 50,000 53 45 .583 38,000 60 60 .155 30,000 58 74 �.001

100,000 59 56 .435 40,000 53 66 .027 40,000 69 71 �.001 1000,000 62 47 .704 50,000 52 71 .003 50,000 56 73 �.001

Note. Donation refers to the dollar amount given by the revealed donor. The double-blind donor always gave $10,000. p values are derived from a two-tailed binomial test of whether the observed percentage differed from 50%.

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

169ATTRIBUTIONS OF CHARITABILITY

Method

After excluding 42 participants who made comprehension errors (using the same items as in Experiment 5), we ended up with a sample of 114 (Mage � 37; 51% female). Participants were as- signed to the observer or beneficiary condition, and they read a vignette in which their town was hit by a flood (similar to Exper- iment 5). In the observer condition, they chose which donor was more charitable: one who gives $10,000 double-blind or one who gives $20,000 with their photo. In the beneficiary condition, par- ticipants read:

Imagine that you are a member of one of the families affected by the flood and are currently staying at the emergency shelter. The director tells you that there have been a number of donations and you can choose between two of them. One donation is for $10,000 and is anonymous, and the other donation is for $20,000 and includes the donor’s photo. Which donation would you choose to receive?

After making this choice, participants also answered which donor they thought was more charitable.

Results and Discussion

As in Experiment 5, participants who took the observer’s per- spective were divided about who was more charitable, 48% pick- ing the revealed donor who gave $20,000 with the photo, a proportion that did not differ from 50% (p � .90, binomial test). In contrast, among participants who took the beneficiary’s perspec- tive, 85% (p � .001) chose the larger donation from the revealed

donor, and a slight majority (65%, p � .036) said that he was the more charitable.

Imagined beneficiaries were significantly more likely than ob- servers to choose the larger gift from the revealed donor over the smaller gift from the anonymous one (p � .001, Fisher’s exact test), and marginally (p � .089) more likely to judge him as more charitable. The fact that beneficiaries showed only a marginal difference in judgments of charitability suggests that they are sensitive to how the donor gives, even if they choose to take a bigger gift because they want the money. Indeed, it’s possible that the only reason that their ratings were even marginal is because they experienced a sort of “spillover” effect of positivity after hypothetically receiving the bigger gift.

In a supplemental Experiment 6a (see online supplemental ma- terials for full details), we test an analogous scenario where par- ticipants take the perspective of the donor. We find that as donors, participants are more likely to say they would send their gift with a photo rather than stay anonymous, compared with observers judging charitability, who favored the anonymous gift. This con- trast depending on the kind of judgment being made is consistent with the idea that, like beneficiaries, donors have different interests than observers; specifically, donors gain by promoting their repu- tation.

Overall, these results confirm that observers are more concerned about whether a donor has a generous disposition, whereas bene- ficiaries are more concerned about the amount of the donation. The results also show a kind of moral hypocrisy that is related to findings from previous research on altruism (Batson & Thompson,

0 20

4 0 60

80 10

Multiple Given by Revealed Donor Compared to Double-Blind Donor

P er

ce nt

ag e

W ho

P r e

fe r

R e v

ea le

d D

on or

(% )

1X 2X 3X 4X 5X 10X 100X

Send Photo Exchange Photos In Person

Figure 7. Titration results for Experiment 5. The x-axis shows the multiple by which the revealed donor’s gift exceeded the double-blind donor’s gift of $10,000. The shaded region indicates an approximate range for values that do not significantly differ from 50%, such that values below it indicate significant preference for the double-blind donor, and values above it indicate significant preference for the revealed donor. See Table 5 for exact significance levels.

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

170 DE FREITAS, DESCIOLI, THOMAS, AND PINKER

2001; Batson et al., 1999;). Batson and Thompson (2001) argue that such effects occur because people surrender their goal to be moral when they have to endure the costs of being moral, and they characterize the difference as a form of hypocrisy. Our results point to a similar inconsistency: As beneficiaries, participants said they preferred the larger gift; but as observers, who did not receive the gift themselves, they judged the larger gift as less charitable than the anonymous one.

General Discussion

We began by asking why people are so impressed by anony- mous charitable giving when anonymity makes no difference to the welfare of the beneficiaries, noting that the preference is so pervasive that it manifests itself both in popular culture and in philosophical analyses, from Maimonides in the 12th Century to those in the present. Drawing on partner choice theory, we hy- pothesized that people’s judgments of charity are attuned to cues of a donor’s generous disposition, including the way the donor reveals his or her identity. These cues affect people’s attributions of a generous or selfish disposition to the donor, a disposition which is thought to generalize across beneficiaries and situations (Gilbert & Malone, 1995; Kelley, 1967) and thus give the observ- ers a reason to seek out generous donors. This allowed us to explain the ordering of four of the rungs in Maimonides’ ladder of giving from most to least charitable (the donor gives double-blind, reveals the beneficiary but not himself, reveals himself but not the beneficiary, and generates common knowledge of each other’s identities). We proposed that these rungs decrease in perceived charitability because the donor creates more opportunities and obligations for receiving future favors in return, making him appear less dispositionally charitable and thus less desirable as a cooperation partner.

Experiments 1–2 found that participants’ judgments largely followed this hierarchy, supporting the partner choice account, with a notable exception: They did not judge a donor who gave with common knowledge as less charitable than a donor who only revealed himself. We hypothesized that while common knowledge may add an obligation for the beneficiary to repay the gift, it could also signal a desire for a symmetrical relationship of communal sharing without quid pro quo obligations or differences in status or authority, like a friendship. Experiment 3 tested this conjecture and confirmed that a donor who exchanged photos rather than sent one unilaterally was perceived as more interested in an ongoing rela- tionship and as more likely to view the beneficiary as a social equal. These positive inferences may have partially offset the negative inference that the donor wished to call in a favor in the future.

Experiment 4 tested a distinctive prediction from the partner choice hypothesis, by leveraging the difference between common knowledge and shared knowledge (Thomas et al., 2014): While common knowledge cements an obligation to repay, lesser states of shared knowledge leave more leeway. This predicts that partic- ipants will see a donor as more generous when he gives with shared knowledge (the donor and beneficiary know who each other are, but they do not know that other one knows this) than when the donor gives with common knowledge. That is what we found: When the donor gave the beneficiary the option of finding out who gave the gift, leaving the donor uncertain whether the beneficiary

did so, participants judged him as more charitable than when the two simply exchanged photos. In both cases, the donor knew who the beneficiary was and vice versa, but only in the common knowledge condition did both parties also know that they each knew this.

Experiment 5 aimed to quantify the difference in charitability between an anonymous donation and several forms of revealed donations, using a titration method to find the switch point at which a revealed donor’s gift is judged as more charitable than a double-blind gift. Surprisingly, we found that when donors send a photo (a blatant signal of the beneficiary’s indebtedness), there is no greater amount of money (even 100 times as much) that they can give to make them appear more charitable than a double-blind donor. In contrast, when donors give while exchanging photos or in person, they can potentially appear more charitable than the double-blind donor, provided they give roughly four times and twice as much money, respectively.

Finally, Experiment 6 tested an additional prediction of the partner choice account based on the perspective of the one doing the judging (assessed by asking participants to imagine themselves in the shoes of one of the parties in a charitable act). The Ladder of Charity, and its interpretation within the theory of partner choice, applies to an observer evaluating a donor’s dispositional charitability. The theory implies that beneficiaries and donors should see things differently: A beneficiary should be more con- cerned with the amount of the gift; a donor with promoting his reputation. Experiment 6 confirmed that imagined beneficiaries preferred a large revealed gift to a small anonymous one, but observers judged the small anonymous one to be more charitable. The results underscore how the incongruent goals of observers and beneficiaries affect their judgments of charitability: observers may deprecate the very sorts of gifts that beneficiaries prefer. The divergence can also lead to a certain kind of hypocrisy, where how people judge others as donors diverges from how people prefer to receive gifts when they are in the beneficiary’s shoes.

Partner Choice and Mental State Reasoning

These experiments could provide insight into the coevolution of partner choice and mental state reasoning in the human lineage. We humans have sophisticated abilities for mental state reasoning that allow us to keep track of different people’s beliefs and desires, including recursive beliefs about others’ beliefs (e.g., reviewed in Frith & Frith, 2003). As humans evolved these abilities, they could put them to use in choosing the best partners in the implicit cooperation marketplace. A skilled mentalizer who observes an act of giving can go beyond a donor’s mere behavior and decipher the donor’s underlying motives and thus his or her disposition for unstinting generosity. Indeed, we saw that people in our experi- ments were sensitive not only to clear cues to disposition such as a willingness to give double-blind, but also took into account subtle details such as whether the gift was shared knowledge or common knowledge.

Relevance to Charitable Organizations

Many donors to charity want to be acknowledged for their gifts. Objectively, this does not diminish the good they can do, espe- cially if this motive can be leveraged to do even more good.

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

171ATTRIBUTIONS OF CHARITABILITY

Donors who anticipate praise may be likelier to donate now and in the future, and may donate more generously. Hence, charitable organizations face the challenge of satisfying this desire while ensuring that donors do not, ironically, end up being viewed as uncharitable, which could ultimately reduce donations.

The current experiments suggest that at least one approach already employed by charitable organizations may achieve this goal. Many charitable organizations ask big donors to go on tours in which they become personally involved with the beneficiaries. These tours may satisfy at least some of the evolved psychological criteria for being involved directly with beneficiaries and the community. The tours may signal that the donors are not just motivated by a concern with their reputations, that they are not asserting a higher status than the beneficiaries, and that they are genuinely interested in establishing relationships with those in need. (We note that even the world’s largest charity driven by rigorous evaluation of efficacy, the Bill and Melinda Gates Foun- dation, often features photographs of the eponymous donors min- gling with beneficiaries in the developing world.) More generally, we hope that empirical studies and theoretical analyses of the evolved psychology of altruism can inform, and be informed by, the best practices in institutional charitable giving.

References

Barclay, P. (2016). Biological markets and the effects of partner choice on cooperation and friendship. Current Opinion in Psychology, 7, 33–38. http://dx.doi.org/10.1016/j.copsyc.2015.07.012

Batson, C. D., & Thompson, E. R. (2001). Why don’t moral people act morally? Motivational considerations. Current Directions in Psycholog- ical Science, 10, 54 –57. http://dx.doi.org/10.1111/1467-8721.00114

Batson, C. D., Thompson, E. R., Seuferling, G., Whitney, H., & Strong- man, J. A. (1999). Moral hypocrisy: Appearing moral to oneself without being so. Journal of Personality and Social Psychology, 77, 525–537. http://dx.doi.org/10.1037/0022-3514.77.3.525

Baumard, N., André, J. B., & Sperber, D. (2013). A mutualistic approach to morality: The evolution of fairness by partner choice. Behavioral and Brain Sciences, 36, 59 –78. http://dx.doi.org/10.1017/S0140525X11002202

Chwe, M. S. Y. (2001). Rational ritual: Culture, coordination, and com- mon knowledge. NJ: Princeton University Press.

Clark, H. H. (1996). Using language. UK: Cambridge University Press. http://dx.doi.org/10.1017/CBO9780511620539

Clark, H. H., & Marshall, C. R. (1981). Definite reference and mutual knowledge. In A. K. Joshi, B. L. Webber, & I. A. Sag (Eds.), Elements of discourse understanding (pp. 10 – 63). UK: Cambridge University Press.

Cosmides, L., & Tooby, J. (1994). Better than rational: Evolutionary psychology and the invisible hand. The American Economic Review, 84, 327–332.

Critcher, C. R., & Dunning, D. (2011). No good deed goes unquestioned: Cynical reconstruals maintain belief in the power of self-interest. Jour- nal of Experimental Social Psychology, 47, 1207–1213. http://dx.doi .org/10.1016/j.jesp.2011.05.001

Crump, M. J., McDonnell, J. V., & Gureckis, T. M. (2013). Evaluating Amazon’s Mechanical Turk as a tool for experimental behavioral re- search. PLoS ONE, 8, e57410. http://dx.doi.org/10.1371/journal.pone .0057410

Curry, O., & Chesters, M. J. (2012). “Putting ourselves in the other fellow’s shoes:” The role of “theory of mind” in solving coordination problems. Journal of Cognition and Culture, 12, 147–159. http://dx.doi .org/10.1163/156853712X633974

Dawkins, R. (1976). The selfish gene. New York, NY: Oxford University Press.

Dunlap, D. W. (1997). $3 million zoo gift revoked because plaque is too small. Retrieved from http://www.nytimes.com/1997/05/15/nyregion/3- million-zoo-gift-revoked-because-plaque-is-too-small.html

Fetherstonhaugh, D., Slovic, P., Johnson, S., & Friedrich, J. (1997). In- sensitivity to the value of human life: A study of psychophysical numb- ing. Journal of Risk and Uncertainty, 14, 283–300. http://dx.doi.org/10 .1023/A:1007744326393

Fiske, A. P. (1992). The four elementary forms of sociality: Framework for a unified theory of social relations. Psychological Review, 99, 689 –723. http://dx.doi.org/10.1037/0033-295X.99.4.689

Frith, U., & Frith, C. D. (2003). Development and neurophysiology of mentalizing. Philosophical Transactions of the Royal Society of London Series B, Biological Sciences, 358, 459 – 473. http://dx.doi.org/10.1098/ rstb.2002.1218

Gilbert, D. T., & Malone, P. S. (1995). The correspondence bias. Psycho- logical Bulletin, 117, 21–38. http://dx.doi.org/10.1037/0033-2909.117 .1.21

Hagen, E. H., & Hammerstein, P. (2006). Game theory and human evo- lution: A critique of some recent interpretations of experimental games. Theoretical Population Biology, 69, 339 –348. http://dx.doi.org/10.1016/ j.tpb.2005.09.005

Hamilton, W. D. (1996). Narrow roads of gene land (Vol. 1). New York, NY: Freeman.

Hardy, C. L., & Van Vugt, M. (2006). Nice guys finish first: The com- petitive altruism hypothesis. Personality and Social Psychology Bulletin, 32, 1402–1413. http://dx.doi.org/10.1177/0146167206291006

Kelley, H. H. (1967). Attribution theory in social psychology. Nebraska Symposium on Motivation, 15, 192–238.

Krasnow, M. M., Delton, A. W., Tooby, J., & Cosmides, L. (2013). Meeting now suggests we will meet again: Implications for debates on the evolution of cooperation. Scientific Reports, 3, 1747. http://dx.doi .org/10.1038/srep01747

Lee, J. J., & Pinker, S. (2010). Rationales for indirect speech: The theory of the strategic speaker. Psychological Review, 117, 785– 807. http://dx .doi.org/10.1037/a0019688

Lin-Healy, F., & Small, D. A. (2013). Nice guys finish last and guys in last are nice: The clash between doing well and doing good. Social Psycho- logical & Personality Science, 4, 692– 698. http://dx.doi.org/10.1177/ 1948550613476308

Maimonides, M. (1170/1180). Laws of gifts to the poor. Mishneh torah, 10, 7–14.

Miller, C. B. (2018a). Generosity: A preliminary account of a surprisingly neglected virtue. Metaphilosophy, 49, 216 –245. http://dx.doi.org/10 .1111/meta.12298

Miller, C. B. (2018b). True generosity involves more than just giving. Retrieved from https://aeon.co/ideas/true-generosity-involves-more- than-just-giving

Newman, G. E., & Cain, D. M. (2014). Tainted altruism: When doing some good is evaluated as worse than doing no good at all. Psychological Science, 25, 648 – 655. http://dx.doi.org/10.1177/0956797613504785

Noë, R., & Hammerstein, P. (1995). Biological markets. Trends in Ecology & Evolution, 10, 336 –339. http://dx.doi.org/10.1016/S0169- 5347(00)89123-5

Nowak, M. A., & Sigmund, K. (2005). Evolution of indirect reciprocity. Nature Reviews, 437, 1291–1298.

Pallotta, D. (2009). Uncharitable: How restraints on nonprofits undermine their potential. Lebanon, NH: University Press of New England.

Pinker, S. (2007). The stuff of thought: Language as a window into human nature. New York, NY: Viking.

Rubinstein, A. (1989). The electronic mail game: Strategic behavior under “almost common knowledge.” The American Economic Review, 79, 385–391.

Ruddy, A. S. (Producer), & Coppola, F. F. (Director). (1972). The godfa- ther [Motion Picture]. United States: Paramount Pictures.

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

172 DE FREITAS, DESCIOLI, THOMAS, AND PINKER

Schelling, T. C. (1960). The strategy of conflict. Cambridge, MA: Harvard University Press.

Smith, N. (1982). Mutual knowledge. Orlando, FL: Academic Press. Thomas, K. A., & Clifford, S. (2017). Validity and Mechanical Turk: An

assessment of exclusion methods and interactive experiments. Comput- ers in Human Behavior, 77, 184 –197. http://dx.doi.org/10.1016/j.chb .2017.08.038

Thomas, K. A., De Freitas, J., DeScioli, P., & Pinker, S. (2016). Recursive mentalizing and common knowledge in the bystander effect. Journal of Experimental Psychology: General, 145, 621– 629. http://dx.doi.org/10 .1037/xge0000153

Thomas, K. A., DeScioli, P., Haque, O. S., & Pinker, S. (2014). The psychology of coordination and common knowledge. Journal of Per- sonality and Social Psychology, 107, 657– 676. http://dx.doi.org/10 .1037/a0037037

Tooby, J., & Cosmides, L. (1996). Relationship and the banker’s paradox: Other pathways to the evolution of adaptations for altruism. Proceedings of the British Academy, 88, 119 –144.

Trivers, R. (1971). The evolution of reciprocal altruism. The Quarterly Review of Biology, 46, 35–57. http://dx.doi.org/10.1086/406755

Trivers, R. (1985). Social evolution. Menlo Park, CA: Benjamin Cum- mings.

van Baal, J. (1975). Reciprocity and the position of women. Assen, the Netherlands: Van Gorcum.

Williams, G. C. (1966). Adaptation and natural selection. NJ: Princeton University Press.

Wilson, E. O. (1975). Sociobiology. Cambridge, MA: Harvard University Press.

Wu, J., Balliet, D., & Van Lange, P. A. (2016). Gossip versus punishment: The efficiency of reputation to promote and maintain cooperation. Sci- entific Reports, 6, 23919. http://dx.doi.org/10.1038/srep23919

Received October 5, 2017 Revision received August 2, 2018

Accepted August 8, 2018 �

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

173ATTRIBUTIONS OF CHARITABILITY