See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/328404105

Hot Utilitarianism and Cold Deontology

Article  in  Social Neuroscience · April 2018

CITATIONS READS

0 346

4 authors, including:

Alejandro Rosas Hugo Viciana

National University of Colombia Universidad de Sevilla
55 PUBLICATIONS   253 CITATIONS    28 PUBLICATIONS   172 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

The evolution of human sociality and the cognitive neuroscience of moral judgment View project

All content following this page was uploaded by Hugo Viciana on 20 October 2018.

The user has requested enhancement of the downloaded file.

 Social Neuroscience

ISSN: 1747-0919 (Print) 1747-0927 (Online) Journal homepage: http://www.tandfonline.com/loi/psns20

Hot utilitarianism and cold deontology: Insights

from a response patterns approach to sacrificial
and real world dilemmas

Alejandro Rosas, Hugo Viciana, Esteban Caviedes & Alejandra Arciniegas

To cite this article: Alejandro Rosas, Hugo Viciana, Esteban Caviedes & Alejandra Arciniegas
(2018): Hot utilitarianism and cold deontology: Insights from a response patterns approach to
sacrificial and real world dilemmas, Social Neuroscience

To link to this article: https://doi.org/10.1080/17470919.2018.1464945

Published online: 18 Apr 2018.

Submit your article to this journal

View related articles

View Crossmark data

Full Terms & Conditions of access and use can be found at

http://www.tandfonline.com/action/journalInformation?journalCode=psns20
SOCIAL NEUROSCIENCE, 2018
https://doi.org/10.1080/17470919.2018.1464945

ARTICLE

Hot utilitarianism and cold deontology: Insights from a response patterns

approach to sacrificial and real world dilemmas
Alejandro Rosasa, Hugo Vicianab, Esteban Caviedesa and Alejandra Arciniegasa,c
a
Department of Philosophy, Universidad Nacional de Colombia, Bogotá, Colombia; bInstituto de Estudios Sociales Avanzados, CSIC, Córdoba,
Spain; cDepartment of Philosophy, Carnegie Mellon University, Pittsburgh, USA

ABSTRACT ARTICLE HISTORY

Research on moral judgment with moral dilemmas suggests that “utilitarian” responses (UR) to Received 26 September 2016
sacrificial high-conflict dilemmas are due to decreased harm aversion, not only in individuals with Revised 31 March 2018
clinical conditions, but also in healthy participants with high scores in antisocial personality traits. Published online 18 April
2018
We investigated the patterns of responses to different dilemma types in healthy participants and
present evidence that some URs to sacrificial dilemmas are morally motivated, as indicated by their KEYWORDS
empathic concern (EC) or primary psychopathy (PP) scores. In study 1 (N = 230) we tested students Moral judgment; moral
with four categories of sacrificial dilemmas featuring innocent victims. In study 2 (N = 590) we dilemmas; empathic
tested students with two categories of sacrificial dilemmas and two “real-world” moral dilemmas, concern; primary
where the agent can improve the lot of strangers by making a personal sacrifice. Results in both psychopathy; utilitarianism
studies showed no decreased harm aversion in a pattern where the only UR is to the sacrificial
dilemma where the number of saved people is very high, and significantly lower harm aversion
only in the pattern of all-deontological respondents in Study 2. The analysis by response patterns
allowed a better discrimination of the moral motivations of participants and showed that at least
some of them express moral concerns in their URs.

Introduction reported primary psychopathy (Bartels & Pizarro, 2011).

Their responses may not express moral motivations.
Research on moral judgment with sacrificial dilemmas
Researchers who believe that decreased harm-aversion
(Greene, Sommerville, Nystrom, Darley, & Cohen, 2001,
is the only relevant proximate cause of URs to sacrificial
2004) asks participants to judge the action of sacrificing
dilemmas (Duke & Bègue 2014; Kahane et al., 2015; Wiech
one stranger in order to save several others. A wealth of
et al., 2013) pose a serious challenge. They cast doubt on
papers reported increased utilitarian responses in partici-
whether research with sacrificial dilemmas can discern
pants with either clinical or subclinical deficits in pro-social
something useful about moral judgment.
emotions – low levels of empathic concern (EC) and high
Studies that report correlations between UR to sacrifi-
levels of primary psychopathy (PP) – (Bartels & Pizarro,
cial dilemmas and low EC/high PP scores deny that those
2011; Carmona-Perera, Verdejo-García, Young, Molina-
responses count as expression of moral dispositions. To
Fernandez, & Pérez-García, 2012; Ciaramelli, Muccioli,
mention a recent example, Kahane and colleagues gath-
Làdavas, & Di Pellegrino, 2007; Duke & Bègue 2014;
ered evidence to support this view (Kahane et al., 2015).
Gleichgerrcht & Young, 2013; Kahane, Everett, Earp,
They reported statistical correlations between pooled URs
Farias, & Savulescu, 2015; Koenigs et al., 2007; Koenigs,
to a variegated set of sacrificial dilemmas, scores of pri-
Kruepke, Zeier, & Newman, 2012; Mendez, Anderson, &
mary psychopathy (PP) obtained with self-report ques-
Shapira, 2005; Patil & Silani, 2014). Researchers disagree
tionnaires, and URs to “real-world” utilitarian dilemmas,
on how to interpret these results. Some argue that low
i.e., those dilemmas where a greater good requires altruis-
empathy levels facilitate cognitive control of prepotent
tic sacrifices on the part of the agent (in “sacrificial” dilem-
harm-aversion, enabling rational and utilitarian decisions
mas, the agent sacrifices a stranger instead). They found
(Greene, 2007; Greene, Morelli, Lowenberg, & Nystrom,
that PP scores correlate positively with UR to sacrificial
2008). Others argue that there is not much to control if
dilemmas in several studies. They also found that PP
participants who give URs display low empathic concern
correlated negatively with real-world measures of
and decreased harm-aversion, or get high scores in self-

CONTACT Alejandro Rosas arosasl@unal.edu.co Department of Philosophy, Universidad Nacional de Colombia, Sede Bogotá, Ciudad Universitaria,
Kra 30 # 45-03, Bogotá, Colombia
Supplemental data for this article can be accessed here.
© 2018 National University of Colombia
2 A. ROSAS ET AL.
utilitarian judgment in two of their four studies. They
found no correlation between sacrificial and real-world
URs: participants who responded positively to sacrificial
dilemmas were neither more nor less likely to be utilitar-
ian regarding real-world utilitarian challenges. From this
set of analyses they infer that UR to sacrificial dilemmas
cannot be interpreted as expressing concern for the
greater good, but rather a diminished aversion to harm
and a maximizing “cold” rationality (see also Bartels &
Pizarro, 2011, p.156).
However, this and other studies leave room for the
possibility that at least some participants delivering UR
to sacrificial dilemmas have moral utilitarian concerns.
This is suggested both by the size of the effect reported
in the correlations between UR and PP (small to moderate,
r between .17 and .32, Cohen, 1988) and by the fact that
no correlation was found between UR to sacrificial and UR
to “real-world” dilemmas. This latter result suggests that
participants giving URs to sacrificial dilemmas are a mixed
bunch. Some of them seem to be utilitarians when the
requirement is to sacrifice their freedom or property.
In order to discriminate different types of “utilitarian”
participants, we first need to realize that the traditional set
of sacrificial dilemmas contains different kinds of incen-
tives to victimize a stranger (Rosas & Koenigs, 2014). Some
of these incentives – like an extremely high number of
people saved – may speak directly to a utilitarian sensitiv-
ity; while others may be attractive mainly to antisocial
personalities – like when the agent is among the five
people saved. This gives us the opportunity to tap into
participants’ specific motivations through their response
patterns. In real life it makes sense, when inferring moti-
vations, to observe a person’s pattern of behavior across
different situations. Those with utilitarian moral concerns
would give URs to sacrificial dilemmas preferably only in
those cases with high utilitarian benefits, while those with
low moral concern would be less discriminating. Their
different motivations should show up in significantly dif-
ferent EC and PP scores. In this debate, it is assumed that
EC and PP scores are correlated with moral concern
(Kahane et al., 2015). Although correlation does not in
itself prove that empathic concern is part of the moral
mechanism, evidence from lesion and neuroimaging stu-
dies suggests that it is (Decety & Cowell, 2014). We rely on
these other studies when we say, throughout this paper
that EC, or EC and PP levels (studies 1 and 2 respectively)
indicate whether the URs of participants express or lack
moral concern.
Study 1
In study 1 we used eight dilemmas in four categories:
Strict, Selfish, Doomed and 100K (see “Materials” below,
and the Appendix). The contrast between the category
Strict and the other three categories is that the latter
offer additional incentives in favor of the sacrifice, over
and above the strict utilitarian incentive of saving five
lives by sacrificing one. The additional incentives – self-
ish stakes, doomed victim and the high number of lives
saved – are each isolated to a separate category.
Hypothesis
We hypothesized that only the incentive given by the
high number of people saved is univocally utilitarian.
The selfish incentive – killing an innocent person to
save a small number that includes you – is not particu-
larly utilitarian, and it would be particularly (though not
uniquely) attractive to participants with antisocial pro-
pensities. Doomed provides a mixed incentive: Pareto
considerations (Huebner, Hauser, & Pettit, 2011) make it
fairly utilitarian – the victim is not made worse off while
her sacrifice saves five lives. But plausibly, some callous-
ness is needed to pick any one person to be sacrificed.
Accordingly, moral concerns should be mainly observa-
ble in participants who give an UR to the 100K category
only. These should be equivalent to the concerns
expressed by all-deontological respondents, but higher
than those expressed in all-utilitarian respondents to
this particular set of dilemmas. The EC scores of the
former two should not differ statistically; but all-utilitar-
ian respondents should exhibit significantly lower EC
scores as expression of decreased harm aversion.
Participants
Data were collected anonymously with pen and paper in
five sessions among undergrad students from 4 different
disciplines at the National University of Colombia. A total
of 234 students participated. Written informed consent
was obtained from all participants. Four did not complete
the Interpersonal Reactivity Index (IRI) and were excluded
from the analyses, giving a total N = 230, of which 155
were males, 69 females and 6 missing data. Mean age
was 21 years and mean college education 1.25 years.
Materials and procedures
Interpersonal reactivity index
The IRI (Davis, 1980) is a self-report scale consisting of four
subcomponents, 7 items each, assessing emotional and
cognitive aspects of empathy (empathic concern, personal
distress, perspective taking and fantasy, respectively). We
used the empathic concern (EC) subcomponent in the
analysis, generally considered to be an important constitu-
ent of moral judgment (Decety & Cowell, 2014) and widely
 SOCIAL NEUROSCIENCE 3

used to assess motivations in research with moral dilem- Table 1. Descriptive statistics – study 1.
mas. Participants responded on a Likert scale running from N Range Mean % Std. Dev.
0 (does not describe me well) to 4 (describes me very well). StrictPk 230 1 .28 .451
StrictFootb 230 1 .30 .461
SelfishFlames 230 1 .43 .497
Sacrificial dilemmas SelfishExplos 230 1 .63 .483
DoomedBoat 229 1 .45 .499
Our eight sacrificial dilemmas are personal, that is, the DoomedGuide 229 1 .57 .496
sacrifice is up-close, using personal force (or some resem- Dam100k 230 1 .54 .500
Assistant100k 229 1 .66 .476
blance of it), and the victim is used as a means (see Gender 224 1 .69 .463
Christensen, Flexas, Calabrese, Gut, & Gomila, 2014). In EC 230 27 16.64 5.386
Valid N (listwise) 218
the category Strict, dilemmas have the characteristics of
Footbridge: the victim is 1) an innocent person, i.e., is not
a source of the imminent threat pending over others, 2) negatively correlated with EC, consistent with data from
not doomed to die and 3) the agent has no selfish stakes other studies. We found no significant correlation in our
in the sacrifice. 100K dilemmas have these same charac- sample between EC and gender. Results are shown in
teristics, but the number of people saved increases from Table 2.
5 to 100K. Selfish is like Strict, except that the agent also Although we aimed at making the same-category
saves herself. Doomed is like Strict, except that the victim items as formally identical as possible, we observed
would die anyway, along with others if not sacrificed. The some divergence in their proportion of URs (see
two intra-category scenarios are constructed as similar as Table 1). A McNemar’s test revealed that the differences
possible to increase the data-points per category. We in UR between same-category items in Selfish, Doomed
asked participants to judge their hypothetically per- and 100K are statistically significant (all p ≤ 0.005, see
formed utilitarian action: “Given that there was no other Supplement table S1). We conjecture that the greater
way to save those people, was your action all right?” The proportion of UR in Selfish-Explosives (.63) compared to
answer was in dichotomous (YES/NO) format. NO was Selfish-Flames (.43) and in Doomed-Guide (.57) com-
coded as 0, YES as 1. The battery was presented in pared to Doomed-Boat (.45) is due to the fact that the
Spanish. An English translation is appended. “personal force” component is rather indirect in the
Participants were asked to complete the EC compo- items with higher proportions (see the items in the
nent of the Spanish version of the IRI (Pérez-Albeniz et al. Appendix). This approximates these items to the cate-
2003) after responding to the eight dilemmas in a within- gory of impersonal dilemmas. Similarly, both in Selfish-
subjects design. We implemented a within-subjects Explosives and 100K-Assistant (.66) the victim causally
design because comparison between mean EC scores contributes (with no intention) to the lethal threat
for the different patterns requires that all participants impending on others. Some participants might tend to
confront and respond to the same dilemmas. Only if the view this causal contribution as distinguishing the vic-
possible patterns are the same for all participants can the tims in these items from the entirely innocent victims of
different patterns adopted indicate individual differences. other items. Both features increase the URs.
The order of presentation of the dilemmas was counter- However, our aim was not to test the sensitivity of
balanced. We ran a Two-Step Cluster analysis to observe participants to intra-category differences, but to check
how participants cluster into response patterns and then whether the variables used to construct our categories
ran subsequent multiple comparisons searching for sig- significantly influenced the participant’s responses. We
nificant differences in mean EC scores across clusters. The therefore computed an average proportion of UR for
p values of the multiple comparisons were adjusted using each category and checked with McNemar’s Test whether
the Bonferroni correction.

Table 2. Point Biserial correlation. EC to gender and to UR in

Results the four dilemma-categories.
Strict Selfish Doomed 100K Gender EC
The percentages of UR by item (non-UR = 0, UR = 1), of EC Pearson −.254** −.212** −.355** −.140* −.043 1
male participants (male = “1”, female = “0”) and the Correlation
Sig. (2-tailed) .000 .001 .000 .034 .521
mean score for empathic concern are shown in Table 1. N 230 230 228 230 224 230
A Point Biserial correlation1 between EC and the UR to *Correlation is significant at the 0.05 level (2-tailed).
the four categories and gender showed all four categories **Correlation is significant at the 0.01 level (2-tailed).

1
SPSS runs the Point Biserial Correlation as a special case of the Pearson Correlation; thus “Pearson Correlation” appears in the
generated table.
4 A. ROSAS ET AL.
participants sensed a difference across categories. The ana-
lysis revealed that the additional incentives were in fact
treated by participants as incentives over and above the
strict utilitarian incentive of saving five instead of one. The
proportions of UR were significantly different for all paired
categories (all p < 0.05, see Supplement table S2) except for
the pair Selfish vs. Doomed (p = 1.000). This does not tell us
which incentives speak to utilitarian sensitivity in particular,
and which ones speak to antisocial propensities, but it is
plausible that we deal here with incentives of both kinds.
Therefore, we ran a cluster analysis and a subsequent
multiple comparison of mean EC scores, hypothesizing
that participants giving URs only to the dilemma where
the sacrifice saves 100K lives would express moral con-
cerns, contrasted to participants indiscriminately giving
URs to all categories. Participant responses to the different
categories were created from our data in this way: a parti-
cipant giving a UR to one or both items in each category
counts as giving a UR to that category, otherwise she
counts as giving a deontological response.
We ran the Two-Step Cluster analysis with four dichot-
omous inputs (the URs to the 4 categories of sacrificial
dilemmas), letting the auto-clustering statistic determine
the optimal number of clusters automatically and using
Schwarz’s Bayesian clustering criterion. According to the
largest Ratio of Distances, the optimal number of clusters
is 2. However, since the cluster quality (silhouette measure
of cohesion and separation) for 2 clusters = .5 (with 1.0 as
highest value) we decided not to use the largest Ratio of
Distances but the lowest Bayesian Information Criterion
coefficient to determine the optimal number of clusters: 9
by this method. The cluster quality (silhouette measure of
cohesion and separation) improved significantly to 0.9
and the predictor importance of each input ranged
between 0.95 and 1.00 (1.00 is the highest score). In
Table 3 we only show the three clusters of relevance for
our hypothesis: No-UR (members of the cluster give deon-
tological responses in every category), All-UR (members
give UR in every category) and 100K (members give UR
only to 100K). (The data for all 9 clusters are shown in the
Supplement as Fig. S1.) Each row in Table 3 corresponds
to one of these three clusters, i.e., patterns of responses.
The name of the pattern appears in the left column. The
row cells show whether participants responded NO (0) or
YES (1) to the 4 different categories in the columns
Table 3. TwoStep cluster analysis – study 1.
INPUTS
Cluster Std.
Labels N DOOMED STRICT 100K SELFISH EC Dev.
No-UR 27 0 (100%) 0 (100%) 0 (100%) 0 (100%) 19.44 4.972
100K-UR 21 0 (100%) 0 (100%) 1 (100%) 0 (100%) 19.48 4.033
All-UR 75 1 (100%) 1 (100%) 1 (100%) 1 (100%) 14.80 5.631
(inputs). All relevant patterns for our study show perfect
consistency: all members of a given cluster follow without
exception the cluster’s pattern. We also show the EC mean
score and standard deviation for each cluster.
Clusters are independent groups: no individual
belongs to more than one cluster. A Kruskal-Wallis H
test showed a statistically significant difference in mean
EC score between the different clusters, H(8) = 28.789,
p = .000, with a mean rank EC of 148.81 for No-UR,
148.88 for 100K-UR and 91.96 for All-UR. Post hoc tests
showed significant differences for No-UR vs. All-UR
(p = .000, adj. p = .004) and for 100K-UR vs. All-UR
(p = .000, adj. p = .016). 100K-UR was not significantly
different from No-UR (p = .997, adj. p = 1.000). Adjusted
p-values were obtained with the Bonferroni correction.
Discussion
We hypothesized that the participants in the 100K-UR
pattern would display a mean EC score significantly higher
than the non-discriminating All-UR pattern and would not
display a mean EC score significantly lower than partici-
pants in the No-UR pattern. Both predictions were con-
firmed. This suggests that members of the cluster 100K-UR
are no less morally concerned than members of No-UR (all-
deontological respondents). These results support our
hypothesis that participants approving of UR to sacrificial
dilemmas are a mixed bunch, and that some of them
approve the sacrifice out of moral concern. We conducted
a second study to validate and expand these findings.
Study 2
We proceeded to test participants with six items: four sacri-
ficial dilemmas, where the agent sacrifices the life of an
innocent stranger to save more lives, and two non-sacrificial
“real-world” dilemmas borrowed from Kahane et al. (2015),
namely Firefighter (renamed to Let Mother Die) and
Veronica’s Comfortable Lifestyle (renamed to Sacrifice
Comfort). In the latter two items the agent has to make a
personal sacrifice for the sake of improving the lot of stran-
gers. We examined the moral concern underlying different
response patterns by measuring both EC and PP scores.
Hypothesis
In a similar fashion to Study 1, we hypothesized that
participants in the pattern combining UR to 100K and to
the two real-world dilemmas would show no less moral
concern than participants following a consistently
deontological pattern, but significantly higher moral
concern than participants with a consistent utilitarian
pattern. Additionally, we hypothesized that participants

in the pattern combining URs to 100K and to real-world
dilemmas would show no less moral concern than par-
ticipants in the pattern UR to real-world dilemmas only.
We established whether participants cluster in these
patterns running a Two-Step Cluster analysis. The p
values of the subsequent multiple comparisons of
mean EC and PP scores between clusters were adjusted
using the Bonferroni correction.
Participants and procedures
Data were collected anonymously with pen and paper in
one session among undergrad students in a massive
humanities course open to students from any discipline
at the National University of Colombia. A total of 595
students participated. Written informed consent was
obtained from all participants. Five did not complete the
Interpersonal Reactivity Index or the Levenson Self-Report
Psychopathy Scale and were excluded, giving a total
N = 590, of which 396 were males, 181 females, and 13
did not report their gender. Mean age was 22.14 years and
mean college education 1.68 years. Participants were
asked to complete the IRI and the Levenson Self-Report
Psychopathy Scale (Levenson, Kiehl, & Fitzpatrick, 1995),
after responding to the six dilemmas in a within-subjects
design. The order of presentation of the different items
was counterbalanced.
Measures
Interpersonal reactivity index
We used the same Spanish version of the IRI as in the
previous study.
Levenson Self-report Psychopathy Scale. The Scale con-
sists of 26 items purporting to describe the participant,
16 of them measuring Primary Psychopathy and 10
measuring Secondary Psychopathy. We used only the
16 items for Primary Psychopathy in a Spanish transla-
tion. Participants responded on a Likert scale running
from 0 (totally disagree) to 4 (totally agree).
Dilemmas
The battery consisted of six items: two Strict dilemmas
(Aquarium and Footbridge); two 100K dilemmas (Vaccine
and a version of Assistant eliminating the victim’s caus-
ality in the anticipated harm); and two “real world”
dilemmas: Sacrifice Comfort and Let Mother Die, both
borrowed from study 4 in Kahane et al. (2015). In Let
Mother Die Albert – a firefighter in action – has to choose
between saving a renowned, successful and still active
peace negotiator or an uneducated housekeeper who
happens to be his mother. In Sacrifice Comfort Veronica
could sacrifice some of her well-earned comfort by
SOCIAL NEUROSCIENCE 5
donating money to help save people from starving to
death in other parts of the world. In the sacrificial items,
we asked participants: “Is it all right to cause the death of
the innocent person in order to save five/100K persons?”
The answer was entered in a dichotomous (YES/NO)
format. In the “real world” items we asked participants:
“How right or wrong would it be for Veronica/Albert to
decide not to donate to charity/not to save the peace
negotiator?” The answer was entered in a five-point
Likert scale running from 0 (Totally Right) to 4 (Totally
Wrong). The dilemmas were presented in Spanish. The
English versions are included in the Appendix.
Results
We computed an average proportion of UR for the two
categories of sacrificial dilemmas (Strict and 100K) in the
following way: a participant giving a UR to one or both
items in a category counts as giving a UR to that cate-
gory; otherwise she counts as giving a deontological
response. The proportions (means in the “real world”
items) of UR to the four items (two categories of sacrifi-
cial and two “real world” dilemmas) are shown in Table 4
and the Point Biserial Correlations between EC and PP on
one hand, and the UR and gender on the other, are
shown in Table 5. The negative correlation between PP
and EC scores was moderate to strong (r = .490, p = .000,
not shown in Table 5). We found no correlation between
the “real world” dilemmas and any of the two categories
of sacrificial dilemmas (not shown in Table 5). In the
correlations between the URs and the EC and PP scores
we obtained a neat asymmetry: negative and positive
correlations with EC and PP respectively in the sacrificial
dilemmas, positive and negative correlations with EC and
PP respectively in the real-world dilemmas. We could not
replicate Kahane et al.’s (2015, pp. 204–205) result of a
positive correlation between PP and UR in Let Mother Die
(Firefighter). Our data show a negative correlation: higher
PP scores are less likely to endorse the UR. Higher PP
scores are therefore consistently less utilitarian in both
types of real-world dilemmas. For all items, except for
Sacrifice Comfort, the significant correlations of UR with
EC scores disappeared when controlling for PP
Table 4. Descriptive statistics – study 2.
N Range Mean % Std. Dev.
Strict 583 1 .16 .365
100k 583 1 .58 .494
Sacrifice Comfort 590 4 1.94 1.172
Let Mother Die 588 4 .82 .937
EC 590 28 17.82 4.661
PP 590 58 18.10 9.050
Gender 577 1 .69 .464
Valid N (listwise) 564
6 A. ROSAS ET AL.

Table 5. Point Biserial correlations. EC and PP to UR and
gender.
Sacrifice
Strict 100k Comfort
EC Pearson −.090* −.138** .384**
Correlation
Sig. (2-tailed) .030 .001 .000
N 583 583 590
PP Pearson .147** .192** −.378**
Correlation
Sig. (2-tailed) .000 .000 .000
N 583 583 590
*Correlation is significant at the 0.05 level (2-tailed).
**Correlation is significant at the 0.01 level (2-tailed).
(Supplement, Table S3). When controlling for EC, all sig-
nificant correlations with PP were preserved, although
with smaller effect sizes (Supplement, Table S4). We
continue showing both scores in our remaining analyses.
As in Study 1, we proceeded to explore the different
response patterns present in the sample. The inputs to
the Two-Step Cluster algorithm (the inputs used to
build the clusters), correspond to four dilemma types:
Strict, 100K, Sacrifice Comfort and Let Mother Die.
Intuitively, we considered the latter two dilemmas as
belonging to different categories. This was confirmed
by a Wilcoxon Signed Rank test (Z = 7,579, p = .000).
We ran the Two-Step Cluster analysis letting the auto-
clustering statistic automatically determine the opti-
mal number of clusters. According to the largest Ratio
of Distances, the optimal number of clusters is three.
The cluster quality at three clusters (silhouette mea-
sure of cohesion and separation) = 0.5 and the pre-
dictor importance for the four inputs ranged from 0
(no importance at all) for both Sacrifice Comfort and
Let Mother Die to 1.0 (full importance) for Strict. This
tells us that the two former input variables were not
used to form clusters. We therefore used the lowest
Bayesian Criterion coefficient, present at 8 clusters, to
determine the optimal cluster number. The cluster
quality with 8 clusters remained at 0.5, but the pre-
dictor importance for the four inputs improved and
ranged from 0.75 for Sacrifice Comfort to 1.0 for Let
mother die.
We give the data for the eight clusters in Table 6. We
label the clusters citing the dilemmas to which their
Let members give an UR (0 = deontological response;
Mother
Die Gender 1 = UR). The percentages are of the members of the
.117** −.205** group following the cluster-formation pattern. The
numbers under the “real world” dilemma inputs show
.004 .000
588 577 the mean response in the Likert scale from 0 to 4, with
−.161** .266** 2 as the neutral midpoint and 4 as full agreement with
.000 .000 the UR. Means below the midpoint count as deontolo-
588 577 gical (see “Discussion” below); means above the mid-
point count as utilitarian. We also show the EC and PP
mean score and standard deviation for each cluster.
Inspection of the table reveals that all clusters show
good internal consistency: all cluster- members follow the
cluster pattern at 100% in the sacrificial categories, except
in cluster 2, where consistency in relation to condition
100K lies high at 81.5%. In the case of the “real world”
utilitarian inputs, the clusters are separated by low mean
(below the neutral midpoint 2 of the 0 to 4 scale) vs. high
mean (above the neutral midpoint of the scale).
To support our hypotheses, we focus on four clusters
or patterns: pattern 5 is utilitarian in 100K and in the two
“real world” dilemmas (100K&RealW-UR); pattern 2 is uti-
litarian in all dilemmas, (All-UR); pattern 6 is utilitarian
only in the “real world” dilemmas (RealW-UR) and pattern
7 is deontological in response to all dilemmas (No-UR).
A Kruskal-Wallis test showed a statistically significant
difference in mean rank EC score between the different
patterns of responses, H(7) = 64.024, p = .000, with a mean
rank EC of 353.80 for RealW-UR, 337.71 for 100K&RealW-
UR, 260.00 for All-URs and 226.27 for No-UR. Post hoc tests
showed significantly higher EC scores for both
100K&RealW-UR and RealW-UR over No-UR (p = .001, adj.
p = .021 and p = .000, adj. p = .000, respectively), but not
over All-UR. Mean EC scores were not significantly differ-
ent between 100K&RealW-UR and RealW-UR (p = .628, adj.
p = 1.0). Adjusted p values were obtained with the
Bonferroni correction. Regarding PP scores, a Kruskal-
Wallis test showed significant differences in mean rank
PP score between the different patterns, H(7) = 70.875,
p = .000, with a mean rank PP of 209.85 for RealW-UR,
230.10 for 100K&RealW-UR, 328.71 for No-UR and 359.56

Table 6. TwoStep cluster analysis – study 2.

INPUTS
Std. Std.
Cluster Labels N LET MOTHER DIE STRICT 100K SACRIF. COMF. EC Dev. PP Dev.
1. Strict&100K-UR 62 0.29 1 (100%) 1 (100%) 1.79 16.82 5.902 20.95 11.224
2. All-UR 27 2.33 1 (100%) 1 (81.5%) 2.04 17.15 4.696 21.44 7.963
3. 100K&Comf-UR 120 0.49 0 (100%) 1 (100%) 2.59 18.48 4.264 17.13 7.750
4. 100K-UR 87 0.54 0 (100%) 1 (100%) 0.62 15.24 4.863 23.00 9.368
5. 100K&RealW-UR 40 2.30 0 (100%) 1 (100%) 2.70 19.15 4.130 15.12 7.304
6. RealW-UR 67 2.07 0 (100%) 0 (100%) 2.51 19.72 3.741 13.63 7.910
7. No-UR 68 0.32 0 (100%) 0 (100%) 0.56 16.28 4.263 19.90 8.248
8. Comf-UR 105 0.33 0 (100%) 0 (100%) 2.67 19.40 3.733 14.73 7.590

for All-UR. Post hoc tests showed significantly lower PP
score for RealW-UR than for No-UR and All-UR (p = .000,
adj. p = .001 and p = .000, adj. p = .002 respectively). They
showed significantly lower PP scores for 100K&RealW-UR
than for All-UR, but only marginally significant lower PP
scores than for No-UR (p = .002, adj. p = .05 and p = .003,
adj. p = .082 respectively). Mean PP scores were not sig-
nificantly different between 100K&RealW-UR and RealW-
UR (p = .542, adj. p = 1.0). Adjusted p values were obtained
with the Bonferroni correction.
We should note that the cluster 100K&Comf-UR
(N = 120) provides further evidence that some partici-
pants endorse URs to both sacrificial and self-sacrificial
(for the sake of strangers) items. Though it includes only
one of the two self-sacrificial items – Sacrifice Comfort –
this item is arguably a better measure of self-sacrificial
utilitarianism than Let Mother Die. The latter is emotionally
demanding and probably the reason why the cluster
100K&RealW-UR contains only 40 participants. The cluster
100K&Comf-UR does not differ statistically from RealW-UR
neither in mean EC nor in mean PP scores (p = .074, adj.
p = 1.0 and p = .012, adj. p = .329 respectively, Bonferroni
corrected). Compared to the all-deontological cluster No-
UR, the mean EC scores of 100K&Comf-UR are significantly
higher, (p = .001, adj. p = .03) while their mean PP scores
are lower, but not significantly (p = .029, adj. p = .823).
Adding the clusters 100K&Comf-UR and 100K&RealW-UR
together, we obtain N = 160 (see Table 6, cluster 3 and 5),
which makes 30% of our sample.
Discussion
Study 2 showed that participants who consistently gave
deontological responses scored significantly lower in EC
compared to a pattern containing UR to 100K, and
higher in PP, though only marginally significant, than
that same pattern. This result assumes that a score
below the midpoint to Sacrifice Comfort and Let Mother
Die counts as deontological, because it expresses that
not-helping strangers is morally permitted. It is typical of
deontology to claim that harming is wrong and deserves
punishment, while nobody can be punished for refusing
to help; and also to claim that the state is allowed to
apply force to punish harms, but not to punish refusals
to help, except in the context of some family relation-
ships (Smith, 1759 TMS, 2.2.1; Kant, 1996 MS 6: 231). This
means that nobody has a deontological obligation to
sacrifice something dear to her for the sake of strangers,
except perhaps when the cost is very low, which is not
the case of the “real world” dilemmas in Study 2 (Kahane
et al., 2017, p.18, col.1 adopt this same view of deontol-
ogy). If we accept this interpretation of deontology,
Study 2 reveals that participants with utilitarian concerns
SOCIAL NEUROSCIENCE 7
expressed in relation to 100K sacrificial dilemmas have
stronger pro-social attitudes than consistently deontolo-
gical participants, as indicated by their EC and PP mean
rank scores.
General discussion
Kahane et al. (2015) deserve credit for pointing out that
dilemmas where the utilitarian option is to sacrifice a
stranger are not representative of many real world dilem-
mas, which demand instead some degree of self-sacrifice
for the sake of strangers. Utilitarianism, in fact, makes high
moral demands on the self. This does not necessarily
mean, however, that dilemmas targeting the sacrifice of
innocent third persons never speak to these high moral
demands. As we have attempted to show here, an
approach relying only on correlations of pooled UR with
EC/PP and using the variegated items introduced by
Greene et al. (2001) is too coarse to discriminate utilitarian
moral motivations from antisocial propensities. Instead,
the response patterns approach offers a better chance of
discrimination. Study 1 shows that participants who deli-
ver URs to sacrificial dilemmas are a mixed bunch with at
least two types: one type behaves conditionally towards
sacrificial dilemmas and is no less morally concerned than
deontological respondents, while the other type is low on
moral concern, lacking deontological sensitivity. Study 2
confirms that some participants who give utilitarian
responses to sacrificial, particularly to 100K dilemmas,
also show a willingness to make sacrifices for strangers.
Furthermore, study 2 reveals that if we judge overall moral
concern by EC and PP scores, we should conclude that
morally unconcerned individuals exist also among consis-
tent deontological respondents. Participants who consis-
tently give deontological responses to a battery including
both sacrificial and real-world dilemmas show signifi-
cantly lower empathic concern (higher EC scores) and
marginally significantly lower harm aversion (higher PP
scores) than participants who give URs to 100K. Response
patterns allow us to more accurately discriminate
between presence or absence of moral concerns in
participants.
The response-patterns approach suggests that overall
moral concern is subdivided in two types of pre-theore-
tical moral sensitivity: deontological and utilitarian (see
also Bialek & de Neys 2017). Although sacrificial dilem-
mas evoke “utilitarian” responses from participants with
decreased harm aversion and apparently no deontologi-
cal sensitivity, not all URs to sacrificial items come from
these participants. Participants that give URs only to
100K and to at least one of the real-world items express
the same moral concerns as those who give URs only to
the real-world items, as indicated by the statistical
8 A. ROSAS ET AL.

equivalence of their EC and PP scores. Importantly, the
former type are not an insignificant proportion of the
population, adding up to 30% of our sample of 530
participants.2
Acknowledgments
This research was supported by scholarships to Alejandra
Arciniegas and Esteban Caviedes, under the Program “Young
Researchers”, sponsored by Colciencias and the National
University of Colombia.
Disclosure statement
No potential conflict of interest was reported by the authors.
Funding
This work was supported by the Departamento Administrativo
de Ciencia, Tecnología e Innovación Colciencias [617, 2013];
References
Bartels, D. M., & Pizarro, D. A. (2011). The mismeasure of
morals: Antisocial personality traits predict utilitarian
responses to moral dilemmas. Cognition, 121(1), 154–161.
Bialek, M., & De Neys, W. (2017). Dual processes and moral con-
flict: Evidence for deontological reasoners’ intuitive utilitarian
sensitivity. Judgment and Decision Making, 12(2), 148–167.
Carmona-Perera, M., Verdejo-García, A., Young, L., Molina-
Fernandez, A., & Pérez-García, M. (2012). Moral decision-
making in polysubstance dependent individuals. Drug and
Alcohol Dependence, 126, 389–392.
Christensen, J. F., Flexas, A., Calabrese, M., Gut, N. K., & Gomila,
A. (2014). Moral judgment reloaded: A moral dilemma vali-
dation study. Frontiers in Psychology, 5, 607.
Ciaramelli, E., Muccioli, M., Làdavas, E., & Di Pellegrino, G.
(2007). Selective deficit in personal moral judgment follow-
ing damage to ventromedial prefrontal cortex. Social
Cognitive and Affective Neuroscience, 2(2), 84–92.
Cohen, J. (1988). Statistical power analysis for the behavioral
sciences (2nd ed.). Hillsdale, NJ: Erlbaum.
Davis, M. H. (1980). A multidimensional approach to individual
differences in empathy. JSAS Catalog of Selected Documents
in Psychology, 10, 85.
Decety, J., & Cowell, J. (2014). The complex relation between
morality and empathy. Trends on Cognitive Science, 18(7),
337–339.
Duke, A. A., & Bègue, L. (2014). The drunk utilitarian: Blood
alcohol concentration predicts utilitarian responses in
moral dilemmas. Cognition, 134, 121–127.
Gleichgerrcht, E., & Young, L. (2013). Low levels of empathic
concern predict utilitarian moral judgment. PLoS ONE, 8(4),
e60418.
Greene, J. D. (2007). Why are VMPFC patients more utilitarian?
A dual-process theory of moral judgment explains. Trends
in Cognitive Science, 11(8), 322–323.
Greene, J. D., Morelli, S. A., Lowenberg, K., & Nystrom, L. E.
(2008). Cognitive load selectively interferes with utilitarian
moral judgment. Cognition, 107, 1144–1154.
Greene, J. D., Nystrom, L. E., Engell, A. D., Darley, J. M., &
Cohen, J. D. (2004). The neural bases of cognitive conflict
and control in moral judgment. Neuron, 44, 389–400.
Greene, J. D., Sommerville, R. B., Nystrom, L. E., Darley, J. M., &
Cohen, J. D. (2001). An fMRI investigation of emotional
engagement in moral judgment. Science, 293, 2105–2108.
Huebner, B., Hauser, M. D., & Pettit, P. (2011). How the source,
inevitability and means of bringing about harm interact in
folk-moral judgments. Mind & Language, 26, 210–233.
Kahane, G., Everett, J. A. C., Earp, B. D., Caviola, L., Faber, N. S.,
Crockett, M. J., & Savulescu, J. (2017). Beyond sacrificial harm: A
two-dimensional model of utilitarian psychology. Psychological
Review, Advance online publication. doi:10.1037/rev0000093
Kahane, G., Everett, J. A. C., Earp, B. D., Farias, M., & Savulescu,
J. (2015). ‘Utilitarian’ judgments in sacrificial moral dilem-
mas do not reflect impartial concern for the greater good.
Cognition, 134(3), 193–209.

2
As we revised this paper, Kahane and collaborators proposed and published a thorough analysis of utilitarian morality as
consisting in two dimensions: instrumental harm for the greater good (this is measured with sacrificial dilemmas), and self-
sacrificing impartial beneficence (this is measured with the “real-world” dilemmas from Kahane et al., 2015). With these two
dimensions they constructed and validated the Oxford Utilitarianism Scale (OUS) (Kahane et al., 2017). They use the OUS, among
other things, to estimate a proportion of the lay population that endorses both dimensions of utilitarian morality. This is
important, because it tells us whether the two dimensions cohere only in virtue of a philosophical construct or alternatively point
towards some pre-theoretical psychological reality, as a “natural kind” so to say. Kahane et al. (2017) come to an estimate of
around 4% to 5%, and doubt whether both dimensions compose a unitary, stable pre-theoretical phenomenon. Our study 2
provides preliminary evidence for the rival view: of the 160 participants that give URs to at least one of two sacrificial and one of
two self-sacrificial items, all answered “yes” to 100K; 79 scored a noncommittal 2 in Sacrifice Comfort (the neutral midpoint in a 5
point scale from 0 to 4) while the remaining 81 participants – 15% of our sample – scored 3 or 4, suggesting that 15% of the lay
population endorses both dimensions of utilitarianism, when the utilitarian “rate of return” is large. Moreover, the mean EC and
PP scores of this subgroup are not statistically different from the mean EC and PP scores of participants endorsing only self-
sacrificial dilemmas for the good of strangers, supporting the view that both dimensions of utilitarianism have some probability
of being a natural kind in this subgroup. These diverging results might arise because we measured utilitarian inclinations with
vignettes representing concrete situations rather than with abstract propositions as in the OUS. Another factor possibly
explaining divergence is that 3 of the 4 items of the Instrumental Harm OUS subscale suggest practices of militarism and
political conservatism (items 2, 3 and 4: (“political oppression”, “torture” and “collateral damage”) (Kahane et al., 2017, p. 16).
These items could prove specially unappealing to participants who would endorse instrumental harm for the greater good
preferably in non-ideological contexts.

Kant, I. (1996). The metaphysics of morals. Cambridge:
Cambridge University Press.
Koenigs, M., Kruepke, M., Zeier, J., & Newman, J. P. (2012).
Utilitarian moral judgment in psychopathy. Social Cognitive
and Affective Neuroscience, 7(6), 708–714.
Koenigs, M., Young, L., Adolphs, R., Tranel, D., Cushman, F.,
Hauser, M., & Damasio, A. (2007). Damage to the prefrontal
cortex increases utilitarian moral judgments. Nature, 446
(19), 908–911.
Levenson, M. R., Kiehl, K. A., & Fitzpatrick, C. M. (1995).
Assessing psychopathic attributes in a noninstitutionalized
population. Journal of Personality and Social Psychology, 68
(1), 151–158.
Mendez, M. F., Anderson, E., & Shapira, J. S. (2005). An inves-
tigation of moral judgment in frontotemporal dementia.
Cognitive and Behavioral Neurology, 18(4), 193–197.
Patil, I., & Silani, G. (2014). “Reduced empathic concern leads
to utilitarian moral judgments in trait alexithymia. Frontiers
in Psychology, 5, 501.
Pérez-Albéniz, A., De Paúl, J., Etxeberría, J., Montes, M. P., &
Torres, E. (2003). Adaptación de interpersonal reactivity
index (IRI) al español. Psicothema, 15(2), 267–272.
Rosas, A., & Koenigs, M. (2014). Beyond ‘utilitarianism’:
Maximizing the clinical impact of moral judgment research.
Social Neuroscience, 9, 661–667.
Smith, A. (1759). The theory of moral sentiments. Oxford:
Oxford University Press. 1976.
Wiech, K., Kahane, G., Shackel, N., Farias, M., Savulescu, J., &
Tracey, I. (2013). Cold or calculating? Reduced activity in the
subgenual cingulate cortex reflects decreased emotional
aversion to harming in counterintuitive utilitarian judg-
ment. Cognition, 126(3), 364–372.
Appendix. English translation of the original
Spanish scenarios
The following hypothetical scenarios describe unusual but
possible situations. They pose decision problems where dis-
agreement has been shown to exist between people. In
giving your answer, please take the standpoint of the
agent in the story who makes the decision. Even though
the decisions made could involve legal consequences, we
ask you not to take them into account and assume that the
law would agree with the decision you reached. Please
answer honestly and in line with your own conscience.
Study 1
1. Park
You are in an amusement park, waiting in line for a
ride. Suddenly, you notice that the mechanism accelerates
and spins out of control. Five people are about to go
flying through the air, and will surely die. The turn-off
switch does not work. The gears could be stopped if you
could jam them with heavy object, but only a very mas-
sive person is nearby. You decide to push this person
without warning. The mechanism stops, the person dies,
but the other five people are saved.
SOCIAL NEUROSCIENCE 9
Given that there wasn’t any other way to save the five people,
was your action all right?
YES NO
2. Footbridge
A runaway and silent trolley is heading down the tracks toward a
group of workmen who will be killed by the impact. You are
seeing this from a footbridge over the tracks, between the
approaching trolley and the five workmen. Next to you there’s
a large and heavy person. Given the impossibility to alert the five
workmen, you decide to push this person off the bridge without
warning, onto the tracks. The stranger dies, but the trolley is
derailed and the five workmen are saved.
Given that there was no other way to save the five people,
was your action all right?
YES NO
3. Selfish Flames
You and other five people are trapped in a burning building.
There is only one emergency exit, but it is blocked by burning
debris. A sixth person tries to crawl through a hole at the base
of that exit. You and the other five lack the time to do the
same thing. If you pushed that person through the debris to
unlock the exit, you and the other five people could get out,
but that person will die. You decide to push that person
through the debris, and you and the other five are saved.
Given that there wasn’t any other way to save yourself and
the other five people, was your action all right? YES NO
4. Selfish Explosives
You are an explosives expert that works for a demolition com-
pany. You are examining an explosive charge on the eighth
floor, and you see that someone inadvertently is going to acti-
vate the detonator for the seventh floor’s charge. The explosion
will uncontrollably collapse the eighth floor, and this will kill you
and other five workmen who are with you. The person doesn’t
hear your warning shouts. If you drop a heavy block of concrete
on him, he will die, but you will prevent the explosion and you
and the other five workmen will be saved. You decide to drop
the block, the person dies, but you and the other five are saved.
Given that there wasn’t any other way to save yourself and
the other five people, was your action all right? YES NO
5. Doomed Guide
You are the guide of a group of survivors from an airplane
crashed in the desert. Among the survivors there is one with a
broken leg. The group faces more than a week’s journey on
foot, without food or water. You are trained to survive in
those harsh conditions, so you decide to carry the injured
person on your back. However, you advance slowly and you
realize that, at this pace, the injured and the remaining peo-
ple will die. You decide to leave the injured person behind,
and you successfully guide the other five to safety, out of the
desert.
Given that there was no other way to save the five people,
was your action all right?
YES NO
 10 A. ROSAS ET AL.
6. Doomed Boat
You are traveling on a ship that catches fire, but you manage
to get on a life-boat. This boat can carry up to five people, but
it’s carrying seven, and sinking due to the weight. You are a
good swimmer so you jump into the water, but the boat, now
carrying six, keeps sinking. None of those six wear life jackets,
and none of them can swim. All of them are going to die,
unless someone jumps into the sea; as no one does this, you
decide to throw into the water the first person you reach. This
person will drown in the stormy waters, but you will have
saved the other five.
Given that there was no other way to save the five people,
was your action all right?
YES NO
7. Assistant 100K
You are the assistant of a scientist that has just discovered a new
energy source that is cheap, clean and safe, and that will end
pollution and poverty. However, the scientist wants to sell this
discovery to some people, who seem decent, but are in fact
ruthless criminals, who will use it to commit genocide: eliminate
a city inhabited by one hundred thousand people. You have no
evidence and the scientist simply does not believe your story.
You decide to kill him, in order to stop the sale and the genocide,
and you save one hundred thousand lives.
Given that there was no other way to save the one hundred
thousand lives, was your action all right?
YES NO
8. Dam
You work at a large dam. Failures in the dam’s construction
material have created a hole in it. The water pressure on that
point will break the dam. The river will flood in no time a near city
inhabited by one hundred thousand people. There isn’t enough
time to warn and evacuate. Standing on the dam’s edge, you look
around but only find a workman whose body-volume would
cover the hole. You decide to push him into the water, the
water current drags him and clogs the hole, which gives enough
time to implement a permanent repair. The workman evidently
drowns, but you save one hundred thousand lives.
Given that there wasn’t any other way to save one hundred
thousand lives, was your action all right?
YES NO
Study 2.
1. Aquarium
You are attending a show of sharks being fed in an aqua-
rium pool. Near you a group of five people falls into the
pool as the metal fence inexplicably collapses. The hungry
sharks are attracted to the splashes. Another man close to
you is bleeding from his nose. If you push him into the
water, his blood will attract the sharks away from the
other five persons. The bleeding man will die, but the
other five will swim to safety.
Is it all right to push the innocent person to his death in
order to save five persons? YES NO
2. Footbridge
A runaway trolley is speeding silently toward five people
working on the tracks, who will be killed in the impact. You
are standing on a footbridge over the tracks midway
between the trolley and the five workmen. A large stranger
is standing beside you. There is no time to alert the work-
men. If you push this large stranger off the footbridge into
the path of the trolley, this stranger will stop the trolley die,
but you will save the other five people.
Is it all right to cause the death of the innocent person
in order to save five persons? YES NO
3. Vaccine
A viral epidemic has spread across the globe. You are a
medical researcher who has developed two substances in
your laboratory, a vaccine and a lethal poison. But due to a
mistake both have been labeled as vaccine. Two people are
working in a room next door, and you can identify the
vaccine by injecting each with one substance, without
warning them. One person will die, but this is the only
way to immediately begin saving lives. Very accurate epi-
demiological projections predict that 100 thousand per-
sons will die if the vaccine does not immediately go into
production.
Is it all right to cause the death of the innocent person
in order to save one hundred thousand people?
YES NO
4. Assistant
You are an assistant of a scientist who has recently dis-
covered a procedure for producing clean energy. Your
boss has received a generous offer from interested buyers.
However, you accidentally overheard a secret conversa-
tion, where they mentioned a plan to commit genocide
with this technology, wiping out a city of 100 thousand
inhabitants. You have no proof besides this, and your boss
is unwilling to believe you. But you devise a plot to
provoke the buyers into killing another assistant, with
whom you are well acquainted, and videotape the
crime. This is the only way for you to stop the genocide
and no one will ever know you were behind this.
Is it all right to cause the death of the innocent person
in order to save 100 thousand people?
YES NO

5. Let Mother Die (FromKahane et al., 2015:
Firefighter)
Albert is a firefighter who is trying to rescue people
from a burning building. The building is about to
collapse and in the time left, Albert will only be
able to rescue one more person. In the room he
has entered, Albert finds two trapped people that
he immediately recognizes. One is a famous peace
negotiator well known for his work resolving armed
conflicts around the world and who is likely to
continue doing this important work if he survives.
The other is a poor, uneducated housekeeper. The
housekeeper is Albert’s mother. Albert has to decide
which of these to save, and the one he does not
save will die.
How right or wrong would it be for Albert not to save
the peace negotiator?
0 1 2 3 4 0
TOTALLY RIGHT TOTALLY WRONG
View publication stats
SOCIAL NEUROSCIENCE 11
6. Sacrifice Comfort (From Kahane et al.,2015:
Veronica’s Comfortable Lifestyle)
Veronica is a writer who has written several successful
popular books, and therefore has a rather ample income.
She has worked hard for this income. Veronica has
enough money to live a comfortable life, with things like
vacations and even the latest gizmos and gadgets – the
fruit of many years of scholarly labor. Veronica realizes
that she could survive at a decent level of happiness with
much less money, and she has good evidence that if she
were to give away large sums of money to effective
charities, she could actually save dozens of people from
squalor and even death. But Veronica would have to give
up the cozy life she’s worked so hard for and live more
modestly.
How right or wrong would it be for Veronica to decide
not to donate to charity?
1 2 3 4
TOTALLY RIGHT TOTALLY WRONG