How Foreign Language Shapes Moral Judgment

Abstract

We investigated whether and how processing information in a foreign language as opposed to the native language affects moral judgments. Participants judged the moral wrongness of several private actions, such as consensual incest, that were depicted as harmless and presented in either the native or a foreign language. The use of a foreign language promoted less severe moral judgments and less confidence in them. Harmful and harmless social norm violations, such as saying a white lie to get a reduced fare, were also judged more leniently. The results do not support explanations based on facilitated deliberation, misunderstanding, or the adoption of a universalistic stance. We propose that the influence of foreign language is best explained by a reduced activation of social and moral norms when making moral judgments.

Full text

How foreign language shapes moral judgment☆
Janet Geipel
a,
⁎, Constantinos Hadjichristidis
b,c,
⁎⁎, Luca Surian
a
a
Department of Psychology and Cognitive Sciences, University of Trento, Italy
b
Department of Economics and Management, University of Trento, Italy
c
Research Centre for Decision Making, Leeds Business School, Leeds University, United Kingdom
HIGHLIGHTS
•We investigated whether and how foreign language influences moral judgment.
•Foreign language prompted more lenient judgments for moral transgressions.
•Foreign language reduced confidence in people's moral evaluations.
•Violations of everyday norms were judged less harshly in a foreign language.
•Foreign language might act through a reduced activation of social and moral norms.
abstractarticle info
Article history:
Received 20 June 2014
Revised 29 January 2015
Available online 11 February 2015
Keywords:
Moral judgment
Foreign language
Bilingualism
Emotion
Social and cultural norms
We investigated whether and how processing information in a foreign language as opposed to the native
language affects moral judgments. Participants judged the moral wrongness of several private actions, such as
consensual incest, that were depicted as harmless and presented in either the native or a foreign language. The
use of a foreign language promoted less severe moral judgments and less confidence in them. Harmful and
harmless social norm violations, such as saying a white lie to get a reduced fare, were also judged more leniently.
The results do not support explanations based on facilitated deliberation, misunderstanding,or the adoption of a
universalistic stance. We propose that the influence of foreign language is best explained by a reduced activation
of social and moral norms when making moral judgments.
© 2015 Elsevier Inc. All rights reserved.
Introduction
Imagine reading about the case of a brother and sister who have an in-
cestuous relationship. What would your moral reaction be? Most people
judge incest as wrong, even in circumstances where potential harm is
minimized (Haidt, 2001). Now imagine reading the same story in a foreign
language that you comprehend well. Would your moral reaction change?
It shouldn't —the story is the same (principle of description invariance
(Tversky & Kahneman, 1981)orextensionality(Arrow, 1982)).
1
But
psychological research on moral violations suggests that it might: A
higher proportion of participants judge that it is acceptable to shove a
man into the path of a trolley to save five lives, when the scenario and
questions are printed in a foreign language rather than in their native
language (Cipolletti, McFarlane, & Weissglass, 2015; Costa, Foucart,
Hayakawa, et al., 2014; Geipel, Hadjichristidis, & Surian, 2014).
Here we aimed to extend the foreign language effect to actions that
are relatively harmless, but nevertheless typically condemned. We ex-
pected that foreign language would distanceparticipants from intuition
and gut-feelings, and through that promoteless harsh moral judgments.
We considered two competing hypotheses. Costa, Foucart, Hayakawa,
et al. (2014) proposed that a “muted”intuition could make the moral
machinery switch from the default automatic, intuitive mode, to a con-
trolled mode, thus focusing the attention to the harmless consequences
(see also Cipolletti et al., 2015). We call this thecontrolled-processing hy-
pothesis. Alternatively, the moral machinery might remain on the auto-
matic, intuitive mode, but the muted intuition would nevertheless
promote less harsh moral judgments. We call this the automatic-pro-
cessing hypothesis. This could happen either through an attenuation of
the typical aversive reaction (see the affect heuristic;Kahneman &
Journal of Experimental Social Psychology 59 (2015) 8–17
☆We wish to thank Marta Degani, Sabine Stricker, Nina Knoll, Maria Micaela Coppola, Helena
Steiger, Cristina Algranati, Silvia Toniolo, Vera Schladitz del Campo, Jane Elizabeth Price, and
Catherine Elizabeth Riley for allowing us to visit their lectures and all their students who volunteered
to participate. We also wish to thank Catherine Caldwell-Harris for helpful discussions and support.
⁎Correspondence to: J. Geipel, Department of Psychology and Cognitive Sciences,
University of Trento, Corso Bettini 31, 38068 Rovereto, Italy.
⁎⁎ Correspondence to: C. Hadjichristidis, Department of Economics and Management,
University of Trento, Via Inama 5, 38122 Trento, Italy.
E-mail addresses: janet.geipel@unitn.it (J. Geipel), k.hadjichristidis@unitn.it
(C. Hadjichristidis).
1
The principle of description invariance or extensionality holds that the way options
are described should not influence a person's preferences about them.
http://dx.doi.org/10.1016/j.jesp.2015.02.001
0022-1031/© 2015 Elsevier Inc. All rights reserved.
Contents lists available at ScienceDirect
Journal of Experimental Social Psychology
journal homepage: www.elsevier.com/locate/jesp

Frederick, 2002; Slovic, Finucane, Peters, & MacGregor, 2002), or a re-
duction of the mental accessibility of moral and social rules (e.g., Bond
& Lai, 1986; Dewaele, 2010). Moral and social rules are learned and ex-
perienced through interactions involving the native language, and so a
foreign language might evoke them to a lesser extent (see Marian &
Neisser, 2000).
Prior research
Foreign language has been shown to attenuate emotional response to
words and phrases (for reviews see Caldwell-Harris, 2014; Pavlenko,
2012). For example, Harris and colleagues found that childhood repri-
mands, such as “Don't do that!”, evoked reduced skin conductance re-
sponses when they were read aloud in a foreign language (Harris,
Ayçiçeĝi, & Gleason, 2003; Harris, Gleason, & Ayçiçeği, 2006). Moreover,
a large questionnaire-based study has shown that late bilinguals rated
taboo words and swearwords as less emotional in a foreign language
than in a native language (Dewaele, 2004; Pavlenko, 2004). Studies
also suggest that a foreign language facilitates people to discuss topics
that are considered off-limits or taboo in their native language. For ex-
ample, Bond and Lai (1986) found that Chinese–English bilinguals
spoke longer about embarrassing topics, such as sexual attitudes, in a
foreign language. In the same vein, Dewaele (2010) found that several
UK-based multilinguals preferred using swearwords in a foreign
language, stating that a foreign language allows them to escape from
social andcultural restrictions. However, some studies have failedto de-
tect an attenuation of emotions (e.g., Ayçiçegi-Dinn & Caldwell-Harris,
2009; Eilola, Havelka, & Sharma, 2007; Sutton, Altarriba, Gianico, &
Basnight-Brown, 2007). To reconcile these findings, Harris and colleagues
proposed that the relative emotionality of a foreign versus a native lan-
guage depends on a complex interplay between age of acquisition, level
of proficiency, and the emotional context in which the foreign language
is learned and used (Caldwell-Harris, 2014; Harris et al., 2006).
Foreign language has also been shown toreduce decision biases that
are believed to have an emotional basis (Keysar, Hayakawa, & An,
2012). Moreover, recent studies demonstrated that it also influences
moral judgment (Costa, Foucart, Hayakawa, et al., 2014; Geipel et al.,
2014). This research was confined to trolley dilemmas (Foot, 1978;
Thomson, 1985) that create tension between a characteristically utili-
tarian perspective, which aims at maximizing net benefit, and a charac-
teristically deontological perspective, which forbids actions that harm
innocent others. You are informed that a runaway trolley will kill five
people unless an action is performed, either pulling a lever (standard
trolley dilemma) that would make the trolley switch to alternative
tracks where one workman is standing, or by pushing a person off a
bridge (footbridge dilemma). Is it morally acceptable to perform such
actions? Adults and children by the age of four typically respond that
it is acceptable to pull the lever, but not to push the person (Cushman,
Young, & Hauser, 2006; Pellizzoni, Siegal, & Surian, 2010).
The dual-process theory of moral judgment (e.g., Greene, Morelli,
Lowenberg, Nystrom, & Cohen, 2008; Greene, Sommerville, Nystrom,
Darley, & Cohen, 2001) explains these findings by suggesting that re-
sponses result from a competition between an automatic, emotional
system that prompts a deontological response, and a slow, controlled
system that favors a utilitarian response.
2
When the proposed action
is emotionally salient(pushing a person off a bridge), the emotional sys-
tem predominates; when it is not (pulling a lever), the controlled sys-
tem overrides the emotional system and produces a characteristically
utilitarian response (see also Koenigs et al., 2007).
When these trolley dilemmas were presented in a foreign language,
utilitarian responses increased but just for the footbridge dilemma
(Costa, Foucart, Hayakawa, et al., 2014; Geipel et al., 2014). Further-
more, as proficiency in the foreign language increased, language differ-
ences decreased. These results were robust across a variety of
foreign—native language combinations and cultures (for a replication,
see Cipolletti et al., 2015). The proposed explanation is that foreign lan-
guage triggers emotional distance, which privileges controlled process-
ing (controlled-processing hypothesis). Its effects are observed in the
footbridge dilemma, as this is typically processed by the emotional sys-
tem, but not in the trolley dilemma, which is commonly processed by
the controlled system (Greene et al., 2001).
Notice that these findings are also compatible with the automatic-
processing hypothesis. The footbridge dilemma involves a prohibited
action (pushing a person; see Cushman, 2013), whereas the trolley di-
lemma does not. It could be that foreign language promoted utilitarian
responses for the footbridge dilemma, because it allowed people to
see past the taboo action (either by reducing the aversive response
linked to the prohibition, and/or by deactivating social and moral
norms). This interpretation is consistent with recent research that
shows that characteristically utilitarian responses do not necessarily
imply controlled processes, but may also arise from impaired social cog-
nition, such as reduced empathy (see Duke & Bègue, 2015; Kahane,
Everett, Earp, Farias, & Savulescu, 2015).
Previous studies examining the role of foreign language on moral
judgmenthave four limitations. First, they have examined only the trol-
ley dilemmas, which involve severe personal harm and concern con-
trived cases distant from the participants' experience (Hare, 1981;
Sunstein, 2005). Second, these dilemmas involve a numerical tradeoff
(killing one in order to save five). As processing information in a foreign
language is difficult, people might have treated the dilemmas as simple
math problems (Bloom, 2011). Third, these studies offer no empirical
support for the central claim that language has a cooling effect on
emotions, or that this cooling effect prompts controlled (utilitari-
an) reasoning. Fourth, the results are open to an in-group out-
group interpretation (Caldwell-Harris, 2014). Participants reading
the materials in a foreign language might have inferred that the
scenarios concerned foreign people (out-group), whereas those
reading them in the native language might have inferred that
they concerned co-nationals (in-group). Research suggests that
feeling socially connected to the characters portrayed in a scenario
influences moral judgment (e.g., Bloom, 2011; Greene, 2013; Lucas
& Livingston, 2014). Thus, the observed foreign language effect
might reflect added assumptions, rather than the use of foreign lan-
guage per se. In the present study we address all these issues.
Present research
Our first aim was to broaden the scope of the foreign language effect
on moral judgment. We examined different types of violations that, ac-
cording to the categorization proposed by Shweder, Much, Mahapatra,
and Park (1997;seealsoGuerra & Giner-Sorolla, 2010), concern the ethics
of Community (e.g., violations of loyalty), Autonomy (e.g., violations of
fairness) and Divinity (e.g., violations of purity) (CAD for short; for an ex-
tension of this model see Graham, Haidt, & Nosek, 2009; Haidt & Joseph,
2008). We selected violations that did not involve physical harm, such
as siblings having consensual and safe sex (see e.g., Björklund, Haidt, &
Murphy, 2000; Eyal, Liberman, & Trope, 2008; adapted from Haidt,
2001). People typically judge such behaviors as ethically wrong, but
struggle to supply moral justifications (moral dumbfounding;Haidt,
Koller, & Dias, 1993). To test the generalizability of this effect, we also
asked participants to evaluate relatively harmful and harmless social
norm violations in community and autonomy ethics. We predicted
that foreign language would promote less harsh moral judgments.
Our second aim was to test whether the effect of foreign language on
moral judgment is underpinned by an attenuation of emotions, as
2
We follow Greene (2014) in using deontological and utilitarian to meanrespectively
“characteristically deontological”and “characteristically utilitarian”asa function of the re-
sponse content, not the underlying motivation.
9J. Geipel et al. / Journal of Experimental Social Psychology 59 (2015) 8–17

previous studies have suggested. A third aim was to differentiate be-
tween the automatic-andcontrolled-processing hypotheses. To this pur-
pose, we used two tasks. We asked participants to state their confidence
in their moral evaluations, and to answer a tricky question (see Study
3), which tests the ability to overridean intuitive wrong answer. Finally,
we assessed a number of deflationary explanations of the foreign lan-
guage effect, such as that it is simply due to misunderstanding, or in-
group out-group considerations.
Study 1
In Study 1, we examined whether reading moral transgressions in a
foreign versus a native language influences moral wrongness judg-
ments (see Table 1 for descriptions of the scenarios). We tested native
German speakers and native Italian speakers. For both, the foreign lan-
guage was English. We predicted that the use of a foreign language
would promote less harsh moral judgments.
Methods
Participants
Study 1a. Forty-eight students (34 females, 14 males; mean age =
27.27 years, range: 18–70 years) from the Free University of Berlin par-
ticipated at the beginning of a lecture.
3
Twenty-seven students were
randomly assigned to the foreign language condition and received a
questionnaire in English and 21 to the native language condition and re-
ceived its German equivalent. Ten participants were excluded from the
analyses, as they were not native German speakers. For all studies, de-
tails of the participants assigned to the foreign language conditions are
presented in Appendix A.
Study 1b. Sixty-four students (56 females, 6 males, 2 unknown; mean
age = 20.56 years, range: 19–24 years) from the University of Trento
participated at the beginning of an English class for credits.
4
Thirty-six
were randomly assigned to the foreign language condition (English),
and 28 to the native language condition (Italian). Four participants
were excluded from the analyses, as they were not native Italian
speakers.
Materials and procedure. We used four scenarios (adapted from Haidt
et al., 1993; Eyal et al., 2008) each describing one or more persons com-
mitting a moral violation (see Table 1, dog, incest, exam and flag items).
Participants were asked to judge the wrongness of each action on a scale
ranging from 0 (perfectly ok)to9(extremely wrong). Two presentation
orders were created and counterbalanced across participants. In all
our studies the original materials were in English, they were translated
by individuals who are highly proficient in the foreign and native lan-
guages used, and checked by bilingual speakers for comparability. The
two versions of each scenario were also closely matched for word
counts.
Results
Study 1a
The results are illustrated in the left panel of Fig. 1. We submitted
the mean wrongness ratings to a 2 (Language: foreign vs. native) × 4
(Scenario: 1–4) mixed factor analysis of variance (ANOVA), with
repeated measures on scenario. As predicted, there was a significant
main effect of language, F(1, 34) = 6.44, p=.016,f= .43. The scenarios
were judged less harshly in the foreign language (M= 3.57, CI [2.83,
4.31])
5
than in the native language (M= 4.91, CI [4.13, 5.70]). There
was also a significant main effect of scenario, F(3, 102) = 24.52,
pb.001, f= .85. Mean wrongness ratings for the dog scenario (M=
6.14, CI [5.21, 7.07]) and the incest scenario (M= 5.53, CI [4.57, 6.49])
were significantly higher than those for the exam scenario (M=3.42,
CI [2.40, 4.44]) and the flag scenario (M= 1.72, CI [0.95, 2.50]). There
was no language × scenario interaction, F(3, 102) = 1.34, p= .265,
f=.20.
Study 1b
The results are illustrated in the right panel of Fig. 1.Wesubmitted
the mean wrongness ratings to a 2 (Language: foreign vs. native) × 4
(Scenario: 1–4) mixed-factor ANOVA, with repeated measures on sce-
nario. As in Study 1a, we found a significant main effect of language,
F(1, 58) = 4.93, p=.030,f= .29. The scenarios were judged less harsh-
ly when presented in the foreign language (M= 5.40, CI [4.73, 6.07])
than in the native language (M= 6.55, CI [5.76, 7.34]). There was a sig-
nificant main effect of scenario, F(3, 174) = 15.57, pb.001, f=.52.The
mean wrongness ratings of the incest scenario (M= 6.93, CI [6.23,
7.64]), the dog scenario (M= 6.20, CI [5.38, 7.02]), and the exam
Table 1
Scenarios used in Studies 1, 2 and 3 (English versions).
Scenario Description
Moral
Dog
1,2,3
Frank's dog was killed by a car in front of his house. Frank had heard
that in China people occasionally eat dog meat, and he was curious
what it tasted like. So he cut up the body and cooked it and ate it for
dinner. [Study 1a: Franz; Study 1b: Frank; Study 2: Franco; Study 3:
Lorenzo]
Incest
1,2,3
A brother and sister are alone in the house and decide to make love
just once. The sister is already taking birth control pills and the
brother uses a condom. They both enjoy the act but decide not to do
it again. They promise each other to keep it a secret.
Exam
1,2,3
A student who doesn't know the answers to some of the questions in
an exam copies them from a student sitting in front of him. He
doesn't get caught and he and the other student both get good
grades. [Study 3: Silvia]
Flag
1,2
A woman is cleaning out her closet, and she finds a national flag. She
decides to cut it up into small pieces and uses the pieces to clean the
toilet.
Bonus
3
Two employees have worked equally toward a project. The project
went well so they are entitled to a collective bonus of 1000 Euros.
The manager, Giulia, is a friend of one of the employees and wants to
allocate the entire bonus to him. Giulia sends an email to the finance
office, but the email never arrives due to a server failure. As a result,
each employee gets 500 Euros.
Non-moral
Train
3
Francesca lives in Florence and would like to visit one of her friends
in Imola. If she takes the Eurostar she has to pay 21 Euro but the
service is very comfortable. If she instead takes the regional train she
pays 6 Euro but the wagons are a bit cold and dirty. Francesca takes
the regional train.
Brand
3
Marco has a strong headache. He goes to the pharmacy with the
intention of buying a headache medicine from Bayer. The pharmacy
is out of the medicine from Bayer Marco was looking for, but has a
generic product which is, in his words, “exactly the same”as the
product he intended to buy. Marco buys the generic brand medicine.
Note. Superscripts indicate the studies in which the scenarios were used.
3
The sample size was determined based on an a-priori power analysis using G*power
(Faul, Erdfelder, Lang, & Buchner, 2007) with the following settings: statistical
power = .80, effect size f= .35 (medium to large effect size, based on Geipel et al.,
2014), p= .05, number of groups = 2 (language conditions), number of repeated
measures = 4 (scenarios), correlation between repeated measures ρ= 0.3 (estimated).
The analysis indicated a minimum sample size of 34. We tested more participants than
the power analysis suggested because the present studies were conducted during classes
in which a greaternumber of participants wasavailable (this applies to all reported stud-
ies). In the present studies, no interim analyses or stopping rules were applied.
4
To determine the appropriate sample size we conducted an a-priori power analysis
utilizing the estimates from Study 1a: effect size f= .43, alpha level = .05, power = .8,
and ρ= 0.4. The minimum sample size suitable to detect a maineffect of language con-
dition was 26.
5
We report 95% CIs unless otherwise stated.
10 J. Geipel et al. / Journal of Experimental Social Psychology 59 (2015) 8–17

scenario(M= 6.40, CI [5.77, 7.03]) were all significantly higher than the
mean wrongness rating of the flag scenario (M= 4.36, CI [3.67, 5.05]).
No other differences were observed. There was no language × scenario
interaction, F(3, 174) = 0.54, p= .656, f=.10.
In sum, Studies 1a and 1b show that the foreign language effect on
moral judgment generalizes to private violations that are offensive but
involve relatively harmless consequences. As anticipated, the use of a
foreign language promoted less harsh moral judgments.
Study 2
In Study 2, we presented a new sample of late Italian–English bilin-
guals with the same four scenarios. In addition to moral judgments, we
also asked participants to rate their emotional reactions. Here, our main
aim was to examine whether foreign language influences moral judg-
ments by attenuating emotions.
Methods
Participants
Seventy-eight English majors (61 females, 15 males, 2 unknown;
mean age = 23.11 years, age range: 20–38 years) from the University
of Verona participated at the beginning of an English lesson; 42 were
randomly assigned to the foreign language condition (English), and 36
to the native language condition (Italian).
6
Materials and procedure
Following each scenario, participants judged the wrongness of the
action that was depicted in it. They were also asked to rate how upset,
worried,disgusted,sad, and angry they felt while reading the scenario
using a 5-point scale (1 = very slightly or not at all,2=alittle,3=
moderately,4=quite a bit,5=extremely;fromWatson, Clark, &
Tellegen, 1988). The presentation order of the moral judgment and
emotion rating tasks was counterbalanced. Preliminary analyses re-
vealed no order effects and so we dropped this factor from the analyses.
Results and discussion
Emotion ratings
The five emotion scales were highly associated (Cronbach's alpha = .85
in the native language condition, Cronbach's alpha = .85 in the foreign
language condition). Thus, we reduced the five scales into an emotion
index score by taking their average (see Fig. 2). The resulting scores
were submitted to a 2 (Language: foreign vs. native) × 4 (Scenario: 1–4)
mixed-factor ANOVA, with repeated measures on scenario. Although
there was no main effect of language, F(1, 76) = 1.24, p= .268,
f= .13, there was a significant language × scenario interaction, F(3,
228) = 5.56, p= .001, f= .26. Simple one-way ANOVAs revealed
that foreign language attenuated emotions in the dog scenario (M
FL
=
3.33, M
NL
=3.88),F(1, 76) = 6.30, p=.014,f= .29, and in the incest
scenario (M
FL
=2.84,M
NL
=3.48),F(1, 76) = 7.22, p= .009, f=.31,
but not in the exam scenario (M
FL
= 2.26, M
NL
= 2.05), F(1, 76) b1,
p= .341, f= .11, and flag scenario (M
FL
= 2.45, M
NL
= 2.12), F(1,
76) = 1.62, p= .207, f= .15. There was also a main effect of scenario,
F(3, 228) = 42.89, pb.001, f= .75. Post hoc comparisons indicated
that the mean emotion ratings of the dog scenario (M=3.61,CI[3.39,
3.83]), the incest scenario (M= 3.16, CI [2.92, 3.40]) and the flag scenar-
io (M= 2.28, CI [2.03, 2.54]) were significantly higher than the mean
emotion rating of the exam scenario (M= 2.16, CI [1.93, 2.38]).
Moral judgments
Previous studies have suggested that foreign language influences
moral judgment by attenuating emotions. We thus grouped the scenar-
ios into those that showed an attenuation of emotions (dog, incest)and
those that did not (exam, flag). We predicted that foreign language
would promote less harsh moral judgments but for only the dog and in-
cest scenarios. Furthermore, these items were relatively more emotion-
al than the exam and flag scenarios, and previous research suggests that
the effect of foreign language is confined to high emotion items, such as
the footbridge dilemma (Costa, Foucart, Hayakawa, et al., 2014).
The findings are illustrated in Fig. 3 and are consistent with this pre-
diction. We analyzed the mean wrongness ratings with simple one-way
ANOVAs. For the dog and incest scenarios foreign language promoted
Fig. 1. Mean wrongness of action ratings (0 = perfectly ok;9=extremely wrong) by scenario and language condition. In Study 1a (a) the native language was German and the foreign
language was English. In Study 1b (b) the native language was Italian and the foreign language was English. Error bars represent 95% CIs.
6
The sample size was determined via an a-priori power analysis using the estimates
from Study1b: effect size f= .29, alpha level = .05,power = .8, and ρ= 0.6. The indicat-
ed minimum sample size was 68.
*
Fig. 2. Mean emotion rat ings by scenario and language condition (Study 2). Native
language: Italian; Foreign language: English. Error bars indicate 95% CIs. *pb.05, **pb.01.
11J. Geipel et al. / Journal of Experimental Social Psychology 59 (2015) 8–17

less harsh moral judgments (M= 7.75, CI [7.22, 8.27]) than the na-
tive language (M= 8.47, CI [8.15, 8.79]), F(1, 76) = 5.17, p=.026,
f= .23. For the exam and flag scenarios, foreign language (M= 5.59,
CI [4.88, 6.15]) and native language (M= 5.68, CI [5.08, 6.33]) induced
similar moral judgments, F(1, 76) = 0.04, p=.851,fb.01.
Correlations between emotion and moral judgment ratings
If foreign language promotes dispassionate controlled thinking, then
moral judgments should rely less on emotions and gut feelings. There-
fore the correlation between moral judgments and emotion ratings
should be weaker in the foreign language condition than in the native
language condition (see Shiv & Fedorikhin, 1999). Contrary to this pre-
diction, both correlations were statistically significant (native language:
r(36) = .70, pb.001; foreign language: r(41) = .56, pb.001), and did
not differ from one another (z=1.01,p=.311).
Mediation analyses
We performed a mediation analysis using the SOBEL macro by
Preacher and Hayes (2004). We used only the dog and incest scenarios,
for which we detected an attenuation of emotions. We used the
bootstrapping procedure (5000 bootstrapped re-samples). Fig. 4 illus-
trates the results. There was a significant direct effect of language on
moral judgment (B=−.72, SE = .32, p= .026). The indirect effect con-
trolling for emotions lies between −1.020 and −0.085 with 99% confi-
dence (B=−.46, SE = .19). Because this intervaldoes not include 0, we
can conclude that emotion mediates the association between language
and moral judgment.
As a check of the proposed mediation hypothesis, we also conducted
an analysis based on a reverse mediation model. We assessed whether
the association between language and emotions is mediated by moral
judgments. There was a direct effect of language on emotions (B=
−.55, SE =.18,p= .003). The indirect effect of language on emotions
controlled for moral judgment lies between −0.460 and 0.009 with
99% confidence (B=−19, SE = .09). Since this interval includes 0,
we can conclude that there is no reverse mediation. Taken together,
the analyses suggest that, for the dog and incest scenarios, emotions
mediated the effect of language on moral judgment.
In sum, in Study 2 we found that the use of a foreign language
promotes less harsh moral judgments but only for the dog and incest
scenarios. For these scenarios, foreign language influenced moral judg-
ment through an attenuation of emotions. Moreover, in both language
conditions moral judgments and emotion ratings were significantly cor-
related and to a similar extent, which is consistent with the automatic-
processing hypothesis.
Study 3
In Study 3 we further examined whether foreign language influ-
ences moral judgment through an attenuation of emotions. Study 2
found support for this hypothesis but only in two out of four scenarios.
Interestingly, these scenarios concerned violations of purity, whereas
the other scenarios concerned a violation of fairness (exam) and loyalty
(flag). Could it be that the effect of foreign language on moral judgment
is mostly confined to purity violations? Studies 1a and 1b, as well as pre-
vious studies on trolley dilemmas (Costa, Foucart, Hayakawa, et al.,
2014; Geipel et al., 2014), suggest that this is not the case, but Study 2
leaves open this possibility. In Study 3 we addressed this question by
testing two violations of purity (dog, incest) and two violations of fair-
ness (exam, bonus; see Table 1). As a further test of the generalizability
of the foreign language effect, we also asked participants to rate the
moral wrongness of 15 items containing relatively harmful (e.g., Sell
someone a defective car) and harmless (e.g., Fail to vote in minor elec-
tions) social norm violations in community and autonomy ethics (see
Appendix B).
A second aim of Study 3 was to provide evidence to distinguish
between the two competing hypotheses. To this end, we used two
new tasks: the Moses illusion task and the confidence-rating task. In
the Moses illusion task (Song & Schwarz, 2008;seeAppendix B for full
instructions) participants are asked: “How many animals of each kind
did Moses take on to the ark?”The correct answer is “can't say”(since
the biblical character was Noah), but most people are unable to override
the automatic response “two”(Alter, 2013). If foreign language
promotes deliberation, then it should improve performance. The
automatic-processing hypothesis predicts no such improvement. If any-
thing, the increased burden on cognitive resources might deteriorate
performance.
In the confidence-rating task, participants rated how sure they were
in their moral evaluations. The automatic-processing hypothesis pre-
dicts that when usinga foreign language people would be less confident
in their judgments because they might lack the “it feels wrong!”signal
that accompanies a strong aversive reaction,
7
which could be grounded
on emotions and/or sociocultural norms. In contrast, the controlled-
processing hypothesis predicts that people would be more confident,
as their judgments will be a product of careful analysis (for further
evidence that deliberative thinking leads to higher confidence, see
Mata, Ferreira, & Sherman, 2013).
In Study 3 we also controlled for the possibility that the foreign lan-
guage effectis due to misunderstanding (we askedparticipants to trans-
late the materials). Furthermore we assessed whether it is constrained
to the use of English, the modern lingua franca, as a foreign language
**
Fig. 3. Moral wrongness ratings for the groups of items by language condition (Study 2).
Higher scores indicate higher moral wrongness ratings. Native language: Italian; Foreign
language: English. Error bars represent 95% CIs. **pb.01.
-.26
ns.
.84***
-.55**
Language
Condition
Emotion
Moral
Judgment
-.72*
Language
Condition
Moral
Judgment
a
b
Fig. 4. Illustration of the direct effect (a) and indirect effect (b) of language on moral
judgment (Study 2). Numbers refer to unstandardized beta weights. *pb.05, **pb.01,
***pb.005.
7
We thank Catherine Caldwell-Harrisfor suggesting this possibility to us.
12 J. Geipel et al. / Journal of Experimental Social Psychology 59 (2015) 8–17

(here we used German). Participants were told explicitly that the sce-
narios took place in their native country and involved co-nationals, in
order to rule out possible in-groupout-group factors. To assess whether
foreign language promotes a universalistic stance, we asked participants
to rate how close they feel to People in my community,Italians,and
People around the world.
8
Methods
Participants
Seventy-four German major students (67 females, 7 males; mean
age = 21.03 years, age range: 18–30 years) from the University of Tren-
to participated at the beginning of a German lesson.
9
Participants were
randomly assigned to either the foreign language condition (German;
n= 37) or the native language condition (Italian; n= 37). Analyses
of the translations revealed that two participants assigned in the foreign
language condition mistranslated one scenario each. The scores for
these two scenarios were excluded from the analyses.
Materials and procedure
We used two purity violations (dog, incest), two fairness violations
(exam, bonus), and two non-moral scenarios (brand, train; see
Table 1), both of which should be judged as “perfectly ok.”Following a
scenario, participants received either the moral judgment task or the
emotion-rating task (in counterbalanced order). We used the same
scales as in Study 2. Preliminary analyses revealed no effect of order,
so we dropped this factor from the analyses. Following the moral judg-
ment task, participants were asked “How sure are you in your evalua-
tion?”and were given a scale ranging from 1 (not at all sure)to7
(very sure). Next, participants received the Moses illusion task (see
Appendix B), and then a subscale of the Identification with All Humanity
Scale (McFarland, Webb, & Brown, 2012). Participants were asked:
“How close do you feel to each of the following groups?”:People in my
community,Italians,People around the world and were given a scale
ranging from 1 (not at all close)to5(very close). Finally, participants
evaluated 15 violations of everyday moral and social norms on a scale
ranging from 1 (not wrong)to4(severely wrong)(seeAppendix B).
Results and discussion
Emotion ratings
As in Study 2, the five emotion scales were highly associated
(Cronbach's alpha = .92 in the native language condition, Cronbach's
alpha = .85 in the foreign language condition). We thus computed an
emotion index by taking the mean score over the five scales. These
mean emotion scores were submitted to a 2 (Language: foreign vs. na-
tive) × 4 (Scenario: 1–4) mixed-factor ANOVA, with repeated measures
on scenario. There was no main effect of language, F(1, 69) = 0.46,
p= .501, f= .08, but a marginally significant language × scenario in-
teraction, F(3, 207) = 2.32, p=.077,f= .18. Simple one-way ANOVAs
revealed that foreign language attenuated emotions for only the dog
scenario (M
FL
= 3.70, CI [3.37, 4.03], M
NL
= 3.13, CI [2.73, 3.53], F(1,
72) = 4.98, p=.029,f= .26. As in the previous studies, there was a sig-
nificant main effect of scenario, F(3, 207) = 29.79, pb.001, f= .66. Post
hoc comparisons indicated that the mean emotion rating of the dog sce-
nario (M= 3.42, CI [3.16, 3.68]) was significantly higher than the mean
emotion ratings of the bonus (M= 2.91, CI [2.64, 3.18]) and exam sce-
narios (M= 2.06, CI [1.83, 2.29]). The mean emotionrating of the incest
scenario (M= 3.14, CI [2.88, 3.39]) was significantly higher than the
mean emotion rating of the exam scenario.
We next tested whether proficiency in a foreign language is associat-
ed with emotion ratings. We created a proficiency score by aggregating
a participant's self-ratings in reading and understanding (both scales
ranged from 1 = almost none,to5=very good). The highest possible
score is 10,which we also assigned to the participants in the native lan-
guage condition. We found no significant association between profi-
ciency and mean emotion ratings, r(72) = .12, p= .322.
Moral judgments
If the influence of foreign language on moral judgments is mediated
by an attenuation of emotion, then we should observe no language ef-
fect on moral judgment, or perhaps an effect for only the dog scenario.
The results of a 2 (Language: foreign vs. native) × 4 (Scenario: 1–4)
mixed-factor ANOVA do not support this prediction (see Fig. 5). There
was a significant main effect of language condition, F(1, 68) = 8.28,
p=.005,f= .35, which was not qualified by a language × scenario in-
teraction, F(3, 204) = 0.67, p= .573, f= .10. The scenarios were judged
less harshly in the foreign language (M= 6.77, CI [6.33, 7.21]) than in
the native language (M= 7.65, CI [7.23, 8.07]). There was also a signif-
icant main effect of scenario, F(3, 204) = 30.51, pb.001, f=.66. The in-
cest scenario received the highest mean moral wrongness rating (M=
8.15, CI [7.75, 8.56]), followed by the bonus scenario (M= 7.92, CI
[7.45, 8.39]), the dog scenario (M= 7.47, CI [6.97, 7.97]), and the
exam scenario (M= 5.30, CI [4.67, 5.93]). Post hoc comparisons indicat-
ed that the mean wrongness ratings of the dog scenario, the incest sce-
nario and thebonus scenario were all significantly higher than the mean
wrongness rating of the exam scenario.
We also computed a correlation between proficiency in a foreign
language and moral judgment ratings. We found that the higher the
language proficiency, the harsher the moral judgment: r(72) = .25,
p= .034.
Correlations between emotion and moral judgment ratings
As in Study 2,within each language condition we computed a corre-
lation between moral judgments and (negative) emotion ratings.
Both correlations were statistically significant (native language:
r(35) = .49, p= .002; foreign language: r(35) = .41, p= .012), and
not different from oneanother (z=−0.39, p= .699). In both languages,
higher negative emotion ratings were associated with more severe
moral wrongness judgments.
Confidence ratings
The automatic-processing hypothesis predicts that foreign language
would decrease confidence in one's moral evaluations, whereas the
controlled-processing hypothesis suggests the opposite. The results
8
We thank the action editor, RogerGiner-Sorolla,and an anonymous reviewerfor sug-
gesting several of these alternative explanations.
9
The appropriate sample size was calculated based on an a-priori sample size calcula-
tion using the estimates from Study 1b: effect size f= .29 (medium effect), alpha
level = .05,power = .8, and ρ= 0.6.The analysis indicated a minimumsample size of 68.
0
1
2
3
4
5
6
7
8
9
Dog Incest Bonus Exam
Wrongness of action
Native language
Foreign language
Fig. 5. Moral wrongness ratings by scenario and language condition (Study 3). Native
language: Italian; Foreign language: German. Higher scores indicate higher mo ral
wrongness ratings. Error bars represent 95% CIs.
13J. Geipel et al. / Journal of Experimental Social Psychology 59 (2015) 8–17

from a 2 (Language: foreign vs. native) × 4 (Scenario: 1–4) mixed-factor
ANOVA support the automatic-processing hypothesis (see Fig. 6). There
was a significant main effect of language, F(1, 68) = 9.61, p= .003,
f= .38, which was not qualified by a language × scenario interaction,
F(3, 204) = 2.35, p= .074, f= .18. Participants in the foreign lan-
guage condition were less confident in their moral judgments
(M= 6.07, CI [5.81, 6.33]) than participants in the native language
condition (M= 6.63, CI [6.38, 6.88]). There was also a significant
main effect of scenario, F(3, 204) = 4.41, p= .005, f=.25.Theincest
scenario received the highest mean confidence rating (M=6.65,CI
[6.42, 6.88]), followed by the bonus scenario (M= 6.46, CI [6.21,
6.71]), the dog scenario (M= 6.18, CI [5.90, 6.47]), and the exam sce-
nario (M= 6.10, CI [5.78, 6.43]). Post hoc comparisons indicated that
the mean confidence rating of the incest scenario was significantly
higher than the mean confidence ratings of thedog and exam scenarios.
We then computed a correlation between proficiency in a foreign
language and confidence ratings. It was significant and positive:
r(72) = .42, pb.001; the higher the language proficiency, the higher
the participants' confidence in their moral judgments.
Moses illusion task
If foreign language promotes analytic reasoning, as the controlled-
processing hypothesis claims, then it should increase the frequency of
correct responses inthis task. It did not. Inthe native language condition
35.1% of participants responded correctly (“can't say”), compared
to 16.2% in the foreign language condition, χ
2
(1, N= 74) = 3.47,
p=.062,φ=−.22. For the control item, no differences were observed
between the two language conditions, χ
2
(1, N=74)=2.38,p=.123,
φ=−.24.
Identification with All Humanity Scale
If foreign language promotes a universalistic stance, then we should
observe differences in terms of how much participants identify with
close and distant social groups. We analyzed the data using a 2
(Language: foreign vs. native) × 3 (Social group: people in my commu-
nity vs. Italians vs. people around the world) ANOVA, with repeated
measures on the last factor. There was no main effect of language: Par-
ticipants in the foreign language condition gave similar closeness rat-
ings (M= 3.56, CI [3.33, 3.78]) as participants in the native language
condition (M= 3.69, CI [3.47, 3.92]), F(1, 71) = 0.75, p= .389,
f= .10. Also, there was no main effect of social group, F(2, 142) =
1.05, p= .353, f= .12, and no language × social group interaction,
F(2, 142) = 0.90, p= .409, f=.11.
Everyday moral and social norms
Here we addressed whether the foreign language effect generalizes
to the evaluation of norm violations that concern relatively harmless
and harmful actions in community and autonomy ethics. To the extent
that it does, foreign language should promote less harsh moral judg-
ments. We tested this hypothesis by conducting two analyses of vari-
ance, one treating subjects as a random factor (F
1
), the other items
(F
2
). In accord with the automatic-processing hypothesis, participants
in the foreign language condition gave less harsh moral judgments
(M= 2.67, CI[2.54, 2.80]) than participants in the native language con-
dition (M= 2.94, CI [2.83, 3.05]), F
1
(1, 73) = 10.17, p= .002, f=.45.
This result was robust in the analysis by items, F
2
(1, 14) = 11.76,
p= .004. In 13 (out of 15) items, the means were in the expected direc-
tion (the exact binomial probability of getting 13 ormore hits out of 15
trials is .007, two-tailed). We also computed a correlation between
moral judgments and proficiency. It was significant and positive:
r(72) = .36, p= .001; the higher the language proficiency, the harsher
the moral judgment.
General discussion
The use of a foreign language, as opposed to a native language, elic-
ited less harsh moral judgments for actions that violate purity, fairness,
and loyalty norms, but have relatively harmless consequences. This was
true across three native-foreign language combinations: German–
English, Italian–English, and Italian–German. The use of a foreign lan-
guage also elicited less harsh moral judgments for fifteen violations of
everyday social and moral norms in community and autonomy ethics.
Thus, the present findings consolidate and extend previous ones re-
garding the trolley dilemmas (Costa, Foucart, Hayakawa, et al., 2014;
Geipel et al., 2014). Critically, in contrast to previousstudies, the present
findings are not open to explanations based on misunderstanding,
added assumptions concerning who is involved in the scenarios
(e.g., in-group or out-group members), a generic bias that distorts the
use of the rating scale (the effect was present in scenarios that induce
both low and high levels of acceptance, and was absent from non-
moral scenarios), or people reducing a moral judgment to a simple
math problem.
The present studies provide limited support for the claim that the
effect of foreign language is mediated by an attenuation of emotions.
Such an effect was found only for two out of four violations in Study
2. In Study 3 we found a main effect of foreign language on moral
judgments, but no attenuation of emotions. The failure to detect a
widespread attenuation of emotions could be related to how we
measured them. Research suggests that emotional scales with verbal
anchors (e.g., 1 = not at all disgusted to 5 = extremely disgusted)elic-
it higher ratings when the anchors are in a foreign language than in a
native language (the anchor contraction effect;seede Langhe,
Puntoni, Fernandes, & van Osselaer, 2011). Presumably emotional
anchors are felt less strongly in a foreign language, and thus partici-
pants compensate by selecting more extreme ratings. Another po-
tential issue is that some emotion words might lack direct
translation equivalents in a foreign language (Pavlenko, 2008). Fu-
ture research could overcome these issues by eliciting emotions
through emotional scales labeled in the native language, scales that
are supplemented by nonverbal cues such as emoticons or colors
(see de Langhe et al., 2011,Studies8&9),orbyusingmoredirect
measures of emotions such as facial affect.
The present findings are not consistent with the idea that foreign
language promotes a switch from intuitive to controlled processes
(see Costa,Foucart, Hayakawa,et al., 2014; Keysar et al., 2012), but rath-
er suggest that intuitive processes remain active (see also
Hadjichristidis, Geipel, & Savadori, in press). First, foreign language pro-
moted less confidence in one's moral evaluations. This finding suggests
that foreign language makes people judge in accord to weakened or
confused intuitions rather than enlightened utilitarian reasoning. An
enlightened utilitarian should consider carefully all outcomes and so
be confident in his or her decision. Standard economic theory cannot
explain certain doubts in a rational agent (see Shafir, Simonson, &
1
2
3
4
5
6
7
Dog Incest Bonus Exam
Confidence ratings
Native language
Foreign language
Fig. 6. Confidence ratings inone's moral evaluationsby scenario and language conditions
(Study 3). Native language: Italian; Foreign language: German. Higher scores indicate
higher confidence ratings. Bars represent 95% CIs.
14 J. Geipel et al. / Journal of Experimental Social Psychology 59 (2015) 8–17

Tversky, 1993). Rational agents compute the (expected) utility associ-
ated with each option, and choose the one with the highest value
(see also Mata et al., 2013). However, an individual who relies on
weak intuitions should be much less confident. The reduction in
gut feelings that makes this person less radical in his or her moral po-
sitions, also explains the reduced confidence in those positions.
Second, foreign language did not improve performance on the
Moses illusion task. Costa, Foucart, Arnon, et al., 2014, Study 4) have
similarly failed to find improved performance on the Cognitive Reflec-
tion Test (CRT; Frederick, 2005), a logical task where correct responding
necessitates the inhibition of an intuitive answer that is incorrect. These
authors suggest that the foreign language effect might be confined to
problems that have an emotional component. An independent reason
to doubt the controlled-processing hypothesis is that thinking in a for-
eign language should increase cognitive load, and thus hinder rather
than facilitate analytical and deliberative reasoning (see Keysar et al.,
2012).
One reason why foreign language might promote less harsh moral
judgments can be traced to memory and socio-cultural learning pro-
cesses. The language in which an event is encoded facilitates its recall
(e.g., Marian & Neisser, 2000; Schrauf & Rubin, 2000, 2004). All the
moral violations we studied concerned norms that have been learned
directly or indirectly through social interactions involving the native
language. Therefore, a native language is more likely to activate these
social and moral norms than a foreign language. In support of this,
Gawinkowska, Paradowski, and Bilewic (2013) demonstrated that
bilinguals use stronger words to translate swearwords from a native-
to-a-foreign language than vice versa, especially for politically incorrect
swearwords, such as ones directed at social groups (ethnophaulisms).
These authors argue that a foreign language exempts bilinguals from
self- or socially-imposed norms, thus making them more prone to
offending others (see also Bond & Lai, 1986; Dewaele, 2010).
This account helps explaining the present findings: all five moral
transgressions as well as thefifteen violations of everyday moral andso-
cial norms involved behaviors which have been learned and experi-
enced predominantly in contexts in which the native language was
used. It can also explain the findings with the trolley dilemmas. The for-
eign language effect was present in the footbridge dilemma but absent
in the trolley dilemma, because only the footbridge dilemma involves
a prohibited action (pushing a person; see also Cushman, 2013). Simi-
larly, the effect was absent from the non-moral dilemmas, because
these dilemmas did not involve social norms.
A further possibility is that the use of a foreign language might
prompt a generic feeling of uncertainty, which in turn promotes less ex-
treme moral judgments. Here we cannot address this hypothesis as the
confidence ratings were tied to the moral evaluations. But future studies
could examine whether foreign language also reduces confidence in
one's responses concerning emotion-neutral items, such as general
knowledge questions.
Conclusion
The present research extends the foreign language effect to
harmless-but-offensive actions, but also to relatively harmful and harm-
less violations of everyday social norms. Foreign language promoted
less harsh moral judgments and less confidence in one's moral evalua-
tions. The present findings do not support the view that the use of a for-
eign language turns people into enlightened rationalists, reasoning
coldly in terms of utilitarian principles. Rather, the picture that emerges
is of people who are guided by a muted intuition, perhaps due to re-
duced activation of relevant moral and cultural norms. Whatever the
final verdict might be in thetheoretical arena, studying how foreign lan-
guage influences moral judgment is of applied interest, as international
public policy involves communicating and processing materials in a for-
eign language before taking decisions that impact on the populations of
many countries.
Appendix A. Details of participants in the foreign
language conditions
Table A.1
Details of participants in the foreign language conditions.
Means, 95% CI
Study 1a (n= 19)
Start age of English education 8.82, [8.13, 9.54]
Self-ratings of language skills in English
a
3.82, [3.46, 4.14]
Study 1b (n= 35)
b
Start age of English education 9.29, [8.64, 9.91]
Self-ratings of language skills in English
a
3.64, [3.45, 3.78]
Self-ratings of comprehension of the materials
c
97%, [96%, 99%]
Study 2 (n= 42)
b
Start age of English education 8.40, [7.76, 8.98]
Self-ratings of language skills in English
a
3.99, [3.87, 4.17]
Study 3 (n= 37)
b
Start age of German education 11.78, [10.42, 13.22]
Self-ratings of language skills in German
d
3.72, [3.40, 4.00]
Note.
a
Participants evaluated their language skills in terms of conversational fluency,
reading, writin g, and understanding, each on a 5-point scale (1 = almost none,2=
poor,3=fair,4=good,5=very good; scale adapted from Caldwell-Harris & Ayçiçeği-
Dinn, 2009). Here we report the mean rating across these scales.
b
All participants had at
least an intermediate level certificate (B1 or B2) in the foreign language as specified by
the Common European Framework of Reference for Languages (CEFR; see page 24 in.
http://www.coe.int/t/dg4/linguistic/Source/Framework_EN.pdf for descriptors).
c
We
asked participants to rate how well they understood each scenario on a 6-point scale
ranging from 50% (some understanding) to 100% (excellent understanding).
d
Participants
evaluated their language skills in terms of reading and understanding, each on a 5-point
scale (1 = almost no ne,2=poor,3=fair,4=good,5=very good). We repor t the
mean rating across these scales.
Appendix B. English versions of the Moses illusion and everyday
social and moral norms tasks (Study 3)
B.1. Moses illusion task
This task was developed by Erickson and Mattson (1981) (see also
Reder & Kusbit, 1991,andSong & Schwarz, 2008). Following Song and
Schwarz (2008), participants were instructed:
“You will read a couple of trivia questions and answer them. You can
write the answer in the blank. In case you do not know the answer,
please write ‘don't know.’You may ormay not encounter ill-formed
questions which do not have correct answers if taken literally. For
instance, you might see the question ‘Garfield is the dog of which
cartoon?’In fact, Garfield is not a dog it is a cat. Please, write ‘can't
say’for this type of question.”
Following these instructions, participants had to respond to two
questions:
(A) “Which country is famous for cuckoo clocks, chocolate, banks, and
pocket knives?”(control question; correct answer: Switzerland)
(B) “How many animals of each kind did Moses take on the Ark?”
(Moses illusion; correct answer: can't say)
B.2. Everyday social and moral norms
These materials were taken from Khemiri, Guterstam, Franck, and
Jayaram-Lindström (2012), who selected them from Mendez,
Anderson, and Shapira (2005). Participants were given to evaluate 15
items, each on a scale ranging from 1 (not wrong)to4(severely wrong).
15J. Geipel et al. / Journal of Experimental Social Psychology 59 (2015) 8–17

In your opinion, how wrong is it to…
1) Fail to keep minor promises
2) Take the last seat on a crowded bus
3) Sell someone a defective car
4) Drive after having one drink
5) Cutinlinewheninahurry
6) Don't give blood during blood drives
7) Are mean to someone you don't like
8) Say a white lie to get a reduced fare
9) Drive out the homeless from your neighborhood
10) Not help someone pick up their dropped papers
11) Keep excess-change at a store
12) Not offer to help after an accident
13) Ignore a hungry stranger
14) Fail to vote in minor elections
15) Keep money found on the ground
References
Alter, A. L. (2013). The benefits of cognitive disfluency. Current Directions in Psychological
Science,22,437–442. http://dx.doi.org/10.1177/0963721413498894.
Arrow, K. J. (1982). Risk perception in psychology and economics. Economic Inquiry,20,
1–9.
Ayçiçegi-Dinn, A., & Caldwell-Harris, C. L. (2009). Emotion-memory effects in bilingual
speakers: A levels-of-processing approach. Bilingualism: Language and Cognition,12,
291–303. http://dx.doi.org/10.1017/S1366728909990125.
Björklund, F., Haidt, J., & Murphy, S. (2000). Moral dumbfounding: When intuition finds
no reason. Lund Psychological Reports,1(2), 1–15.
Bloom, P. (2011). Family, community, trolley problems, and the crisis in moral psycholo-
gy. TheYaleReview,99,26–43. http://dx.doi.org/10.1111/j.1467-9736.2011.00701.x.
Bond, M. H., & Lai, T. -M. (1986). Embarrassment and code-switching into a second lan-
guage. Journal of Social Psychology,126,179–186.
Caldwell-Harris, C. L (2014, May 20). In a foreign language, “killing 1 to save 5”may be
more permissible. ScientificAmerican(Retrieved from http://www.scientificamerican.
com/article/in-a-foreign-language-killing-1-to-save-5-may-be-more-permissible/).
Caldwell-Harris, C. L., & Ayçiçeği-Dinn, A. (2009). Emotion and lying in a non-native lan-
guage. International Journal of Psychophysiology,71, 193–204. http://dx.doi.org/10.
1016/j.ijp sycho.2008.09.006.
Cipolletti, H., McFarlane, S., & Weissglass, C. (2015). The moral foreign-language effect.
Philosophical Psychology.http://dx.doi.org/10.1080/09515089.2014.993063 (Advance
online publication).
Costa, A., Foucart, A., Arnon, I., Aparici, M., & Apesteguia, J. (2014). ‘Piensa’twice: On the
foreignlanguage effect in decision making.Cognition,130,236–254. http://dx.doi.org/
10.1016/j.cognition.2013.11.010.
Costa, A., Foucart, A., Hayakawa, S., Aparici, M., Apesteguia, J., Heafner, J., et al. (2014).
Moral judgment depends on language. PLoS ONE,9, e94842. http://dx.doi.org/10.
1371/journal.pone.0094842.
Cushman, F. (2013). Action, outcome and value: A dual-system framework for morality.
Personality and Social Psychology Review,17,273–292. http://dx.doi.org/10.1177/
1088868313495594.
Cushman, F., Young, L., & Hauser, M. (2006). The role of conscious reasoning and intui-
tions in moral judgment: Testing three principles of harm. Psychological Science,17,
1082–1089. http://dx.doi.org/10.1111/j.1467-9280.2006.01834.x.
de Langhe, B., Puntoni, S., Fernandes, D., & van Osselaer, S. M. J. (2011). The anchor con-
traction effect in international marketing research. Journal of Marketing Research,
48,366–380. http://dx.doi.org/10.1509/jmkr.48.2.366.
Dewaele,J. -M. (2004). The emotional forceof swearwords and taboo words in thespeech
of multilinguals. Journal of Multilingual and Multicultural Development,25,204–222.
http://dx.doi.org/10.1080/01434630408666529.
Dewaele, J. -M. (2010). ‘Christ fucking shit merde!’Language preferences for swearing
among maximally proficient multilinguals. Sociolinguistic Studies,4,595–614. http://
dx.doi.org/10.1558/sols.v4i3.595.
Duke, A. A., & Bègue, L. (2015). Thedrunk utilitarian: Blood alcoholconcentration predicts
utilitarian responses in moral dilemmas. Cognition,134,121–127. http://dx.doi.org/
10.1016/j.cognition.2014.09.006.
Eilola, T. M., Havelka, J., & Sharma, D. (2007). Emotional activation in the first and second
language. Cognition and Emotion,21, 1064–1076. http://dx.doi.org/10.1080/
02699930601054109.
Erickson, T. D., & Mattson, M. E. (1981). From words to meaning: A semantic illusion.
Journal of Verbal Learning and Verbal Behavior,20,540–551. http://dx.doi.org/10.
1016/S0022-5371(81)90165-1.
Eyal, T., Liberman, N., & Trope, Y. (2008). Judging near and distant virtue and vice. Journal
of Experimental Social Psychology,44, 1204–1209. http://dx.doi.org/10.1016/j.jesp.
2008.03.012.
Faul, F., Erdfelder, E., Lang, A. -G., & Buchner, A. (2007). G*Power 3: A flexible statistical
power analysis program for the social, behavioral, and biomedical sciences.
Behavior Research Methods,39,175–191. http://dx.doi.org/10.3758/BF03193146.
Foot, P. (1978). Virtues and vices and otheressays in moral philosophy. Berkeley: University
of California Press.
Frederick, S. (2005). Cognitive reflection and decision making. Journal of Economic
Perspectives,19(4), 25–42. http://dx.doi.org/10.1257/089533005775196732.
Gawinkowska, M., Paradowski, M. B., & Bilewic, M. (2013). Second language as an
exemptor from sociocultural norms. Emotion-related language choice revisited.
PLoS ONE,8,e81225.http://dx.doi.org/10.1371/journal.pone.0081225.
Geipel, J., Hadjichristidis, C., & Surian, L. (2014, May). Moral judgments: Foreign language
promotes utilitarian judgments. Poster session presented at the 26th Convention of the
Association for Psychological Science, San Francisco, CA.
Graham, J.,Haidt, J., & Nosek, B. A. (2009). Liberalsand conservatives relyon different sets
of moral foundations. Journal of Personality And Social Psychology,96, 1029–1046.
http://dx.doi.org/10.1037/a0015141.
Greene, J. (2013). Moral tribes: Emotion, reason, and the gap between us and them. New
York, NY, US: Penguin Press.
Greene, J. D. (2014). Beyond point-and-shoot morality: Why cognitive (neuro)sci-
ence matters for ethics. Ethics,124,695–726. http://dx.doi.org/10.1086/
675875.
Greene, J. D., Morelli, S. A., Lowenberg, K., Nystrom, L. E., & Cohen, J. D. (2008). Cognitive
load selectively interferes with utilitarian moral judgment. Cognition,107,
1144–1154. http://dx.doi.org/10.1016/j.cognition.2007.11.004.
Greene, J. D., Sommerville, R., Nystrom, L. E., Darley, J. M., & Cohen, J. D. (2001). An fMRI
investigation of emotional engagement in moral judgment. Science,293,2105–2108.
http://dx.doi.org/10.1126/science.1062872.
Guerra, V. M., & Giner-Sorolla, R. (2010). The Community, Autonomy, and Divinity Scale
(CADS): A new tool for the cross-cultural study of morality. Journal of Cross-Cultural
Psychology,41,35–50. http://dx.doi.org/10.1177/0022022109348919.
Hadjichristidis, C., Geipel, J., & Savadori, L. (2015s). The effect of foreign language in judg-
ments of risk and benefit: The role of affect. Journal of Experimental Psychology:
Applied (in press).
Haidt, J. (2001). The emotional dog and its rational tail: A social intuitionist approach to
moral judgment. Psycholo gical Review,108,814–834. http://dx.doi.org/10.1037/
0033-295X.108.4.814.
Haidt,J.,&Joseph,C.(2008).Themoralmind:Howfive sets of innate intuitions guide the
development of many culture-specific virtues, and perhaps even modules. In P.
Carruthers, S. Laurence, & S. Stich (Eds.), The innate mind Volume 3: Foundations and
the future (pp. 367–391). New York, NY, US: Oxford University Press. http://dx.doi.
org/10.1093/acprof:oso/9780195332834.003.0019.
Haidt, J., Koller, S., & Dias, M. G. (1993). Affect, culture, and morality, or is it wrong to eat
your dog? Journal of Personality and Social Psychology,65,613–628. http://dx.doi.org/
10.1037/0022-3514.65.4.613.
Hare, R. M. (1981). Moral thinking:Its levels, method, and point. Oxford: Oxford University
Press.
Harris, C. L., Ayçiçeĝi, A., & Gleason, J. B. (2003). Taboo wordsand reprimands elicit great-
er autonomic reac tivity in a first language than in a second language. Applied
PsychoLinguistics,24,561–579. http://dx.doi.org/10.1017/S0142716403000286.
Harris, C.L., Gleason, J. B., & Ayçiçeği, A. (2006). Whenis a first language more emotional?
Psychophysiological evidence from bilingual speakers. In A. Pavlenko (Ed.), Bilingual
minds: Emotional experience, expression, and representation. Clevedon, United
Kingdom: Multilingual Matters.
Kahane,G.,Everett,J.A.C.,Earp,B.D.,Farias,M.,&Savulescu,J.(2015).‘Utilitarian’
judgments in sacrificial moral dilemmas do not reflect impartial concern for
the greater good. Cognition,134,193–209. http://dx.doi.org/10.1016/j.cogni-
tion.2014.10.005.
Kahneman, D., & Frederick, S. (2002).Representativeness revisited: Attribute substitution
in intuitive judgment. In T. Gilovich, D. Griffin, & D. Kahneman (Eds.), Heuristics and
biases: The psychology of intuitive judgment (pp. 49–81). New York, NY, US: Cam-
bridge University Press.
Keysar, B., Hayakawa, S. L., & An, S. (2012). The foreign-language effect: Thinking in a for-
eign tongue reduces decision biases. Psychological Science,23,661–668. http://dx.doi.
org/10.1177/0956797611432178.
Khemiri, L.,Guterstam, J., Franck, J., & Jayaram-Lindström, N. (2012). Alcohol dependence
associated with increasedutilitarian moral judgment: A case control study. PLoS ONE,
7,e39882.http://dx.doi.org/10.1371/journal.pone.0039882.
Koenigs, M.,Young, L., Adolphs, R., Tranel, D., Cushman, F., Hauser, M., et al. (2007). Dam-
age to the prefrontal cortex increases utilitarian moral judgements. Nature,446,
908–911. http://dx.doi.org/10.1038/nature06785.
Lucas, B. J., & Livingston, R. W. (2014). Feeling socially connected increases utilitarian
choices in moral dilemmas. Journal of Experimental Social Psychology,53,1–4. http://
dx.doi.org/10.1016/j.jesp.2014.01.011.
Marian, V., & Neisser, U. (2000). Language-dependent recall of autobiographical memo-
ries. Journal of Experimental Psychology: General,129,361–368. http://dx.doi.org/10.
1037/0096-3445.129.3.361.
Mata, A., Ferreira, M. B., & Sherman, S. J. (2013). The metacognitive advantage of deliber-
ative thinkers: A dual-process perspective on overconfidence. Journal of Personality
and Social Psychology,105,353–373. http://dx.doi.org/10.1037/a0033640.
McFarland, S., Webb, M., & Brown, D. (2012). All humanity is my ingroup: A measure and
studiesof identification withall humanity. Journalof Personality and SocialPsychology,
103,830–853. http://dx.doi.org/10.1037/a0028724.
Mendez, M. F., Anderson, E., & Shapira, J. S. (2005). An investigation of moral judgement
in frontotemporal dementia. Cognitive and Behavioral Neurology,18,193–197. http://
dx.doi.org/10.1097/01.wnn.0000191292.17964.bb.
Pavlenko, A. (2004 ). “Stop doing that, “la komu skazala!””: Language choice and emotions
in parent–child communication. Journal of Multilingual and MulticulturalDevelopment,
25,179–203. http://dx.doi.org/10.1080/01434630408666528.
Pavlenko, A. (200 8). Emotion and emotion-laden words in the bilingual lexicon.
Bilingualism: Language and Cognition,11,147–164. http://dx.doi.org/10.1017/
S1366728908003283.
16 J. Geipel et al. / Journal of Experimental Social Psychology 59 (2015) 8–17

Pavlenko, A. (2012). Affective processing in bilingual speakers: Disembodied cognition?
International Journal of Psychology,47,405–428. http://dx.doi.org/10.1080/
00207594.2012.743665.
Pellizzoni, S., Siegal, M., & Surian, L. (2010). Contact principle and utilitarian moral judg-
ments in young children. Developmental Science,13, 265–270. http://dx.doi.org/10.
1111/j.1467-7687.2009.00851.x.
Preacher, K. J., & Hayes, A. F. (2004). SPSS and SAS procedures for estimating indirect ef-
fects in simple mediation models. Behavior Research Methods, Instruments, &
Computers,36,717–731. http://dx.doi.org/10.3758/BF03206553.
Reder, L. M., & Kusbit, G. W. (1991). Locus of the Moses illusion: Imperfect encoding, re-
trieval or match? Journal of Memory and Language,30,385–406. http://dx.doi.org/10.
1016/0749-596X(91)90013-A.
Schrauf, R. W., & Ru bin, D. C. (2000). Internal languages of retrieval: The bilingual
encoding of memo ries for the personal past. Memory & Cognition,28,616–623.
http://dx.doi.org/10.3758/BF03201251.
Schrauf, R. W., & Rubin, D. C. (2004). The ‘Language’and ‘Feel’of Bilingual Memory:
Mnemonic Traces. Estudios De Sociolingüística: Linguas, Sociedades E Culturas,5,
21–39.
Shafir, E., Simonson, I., & Tversky, A. (1993). Reason-based choice. Cognition,49,11–36.
http://dx.doi.org/10.1016/0010-0277(93)90034-S.
Shiv, B., & Fedorikhin, A. (1999). Heart and mind in conflict: The interplay of affect and
cognition in consumer decision making. Journal of Consumer Research,26, 278–292.
http://dx.doi.org/10.1086/209563.
Shweder, R. A., Much, N. C., Mahapatra, M., & Park, L. (1997). The ‘big three’of morality
(autonomy, community, divinity) and the ‘big three’explanations of suffering. In A.
M. Brandt,& P. Rozin (Eds.), Morality and health (pp. 119–169). Florence,KY, US: Tay-
lor & Frances/Routledge.
Slovic, P., Finucane, M., Peters, E., & MacGregor, D. G. (2002). The affect heuristic. In T.
Gilovich, D. Griffin, & D. Kahneman (Eds.), Heuristics and biases: The psychology of in-
tuitive judgment (pp. 397–420). New York, NY, US: Cambridge University Press.
Song, H., & Schwarz, N. (2008). Fluency and the detection of misleading questions: Low
processing fluency attenuates the Moses illusion. Social Cognition,26,791–799.
http://dx.doi.org/10.1521/soco.2008.26.6.791.
Sunstein, C. R. (2005). Moral heuristics. Behavioral and Bra in Sciences,28,531–573. http://
dx.doi.org/10.1017/S0140525X05000099.
Sutton, T. M., Altarriba, J., Gianico, J. L., & Basnight-Brown, D. M. (2007). The automatic access
of emotion: Emotional stroop effects in Spanish–English bilingual speakers. Cognition
and Emotion,21, 1077–1090. http://dx.doi.org/10.1080/02699930601054133.
Thomson, J. (1985). The trolley problem. Yale Law Journal,94,1395–1415. http://dx.doi.
org/10.2307/796133.
Tversky, A., & Kahneman, D. (1981). The framing of decisions and the psychology of
choice. Science,211,453–458. http://dx.doi.org/10.1126/science.7455683.
Watson, D., Clark, L. A., & Tellegen, A. (1988). Development and validation of brief mea-
sures of positive and negative affect: The PANAS scales. Journal of Personality and
Social Psychology,54, 1063–1070. http://dx.doi.org/10.1037/0022-3514.54.6.1063.
17J. Geipel et al. / Journal of Experimental Social Psychology 59 (2015) 8–17