Does Thinking in a Foreign Language Lead to Riskier Decisions? An Empirical Study of Decision-Making in Moral Dilemmas

Contents

Abstract

Thinking in a Foreign Language Leads to Riskier Decisions

Method
Participants
Procedure

Results and Discussion

Risk scenario 1: Ebola
Risk scenario 2: Lottery
Risk scenario 3: Fire in building
Risk scenario 4: Africa Safari
Moral Dilemmas
Risk Matrix

General Discussion

References

Appendix A: Risk Scenarios

Appendix B: Moral dilemmas

Appendix C: Risk Matrix

Abstract

Millions of people around the world use foreign languages for communication on a daily basis. Not being on the same level of proficiency as the mother tongue, thinking in a foreign language can lead to changes in the decision-making process. We tested to see whether thinking in a foreign language influences risk and moral judgments. German and English natives completed an online study either in their native, or in a foreign language. Results indicate that participants, answering in a foreign language, were more risk seeking. No language effect was found for people’s moral judgments. These findings support the hypothesis that a foreign language puts more cognitive load on the brain and thus leaves less capacity available for rational decision-making.

Keywords: language, risk, decision-making, moral, cognitive load

Thinking in a Foreign Language Leads to Riskier Decisions

Language plays a central role in our life. We use it as means to communicate our thoughts and ideas. Many people have knowledge of more than one language that was acquired in the course of their school life. The foreign language is most often not on the same level of proficiency as the mother tongue. This can lead to a different path of information processing and therefore to an overall change in judgment and decision-making (Keysar, Hayakawa, & An, 2011).

Thinking can be divided into two separate systems (Kahneman, 2003) that differ both qualitatively and quantitatively. The first type of thinking (System 1) is fast, automatic and does not consume much cognitive resources. This is the type of thinking we use most of the time. It is efficient and most often reliable. The second type (System 2) is slow, non-automatic, and requires much more concentration and cognitive resources. We tend to use it sparingly, because it is demanding and may cause discomfort.

When thinking and speaking in our mother tongue, we use System 1. Words come naturally to us, without putting much thought into them. This is because we are experienced and proficient. We also feel emotionally connected to our mother tongue. The language learned in childhood may have deeper emotional expressiveness than the foreign language. (Anooshian & Hertel, 1994, as cited in Ayçiçegi & Harrism, 2004). Curse and taboo words tend to make us feel awkward and cause emotional response. This does not hold true for our foreign language. Speaking in a foreign language is a much more demanding activity. An activity that requires concentration and careful consideration of words and grammar. It takes more time and puts a great load on our cognitive system. Because the foreign language is often not on the same level of proficiency as the mother tongue, differences in our thinking are likely to occur (some languages have completely different grammatical structures, which can further intensify the gap). Foreign languages are learned in more emotionally neutral settings than the first language. Consequently, less arousal would be conditioned to foreign-language words (Bond & Lai, 1986) leading to an “emotional distancing” from the foreign language. For instance, taboo and curse words do not seem to cause emotional responses or feelings of awkwardness. Some people also feel more comfortable to discuss embarrassing topics in their second language (Bond & Lai, 1986).

We assume that this emotional distancing from the foreign language can lead to a more rational way of thinking. We argue that subjects who assess risk and moral situations would make more rational decisions when thinking in a foreign language. We suspect that subjects would be more risk avoiding and also more responsible and democratic in their moral judgments.

Method

Participants

Two hundred sixty-four people attempted to participate in the online study. One hundred sixty-three successfully completed the entire study. Of the one hundred sixty-three participants, a total of forty-three were excluded, because they did not satisfy the requirements: Eight subjects were excluded, because their native language was neither German nor English. Twenty were excluded, because their second best language was not German (for English natives) or English (for German natives). A further criterion for exclusion was introduced, in which participants were asked to report whether they had answered conscientiously, and whether we should use their data for scientific purposes. No negative consequences followed admitting the truth, so we assume that participants answered this question honestly. Nine participants reported not having answered honestly, so their data was left out. We also considered the fact that an online study cannot provide the rigorous control conditions required for optimal experimentation and thus added the criterion “completion time” as a measure for accuracy. We calculated the minimum time required for reading all texts in the study conscientiously. Three participants were excluded, because it was physically impossible for them to have read all the texts for such a small amount of time. (e.g. the mean time of completion was 10 min; these subjects completed it in only 1.5 min).

We were left with one hundred twenty participants for the data analysis. Twenty-eight male and ninety-two female participants with a mean age of (M = 26.62, SD = 14.07). Subjects were either native German (n = 104) or native English speakers (n = 16). A measure for language fluency was introduced: We asked participants to rate their language proficiency on a 10-point scale (10: full fluency). German Natives scored 8.1, English Natives 6.5.

Procedure

We constructed an online survey using the software Unipark. It was written in both English and German, and subjects were randomly assigned to fill out the survey in either of the languages. We used the date they were born as a random generator: those born on an odd date were assigned to the English version; those born on an even date to the German version, respectively. All subjects were recruited either per email or personally. We consulted a bilingual speaker to ensure that instructions and scenarios in both languages conveyed the same meaning.

All participants were presented with 4 risk scenarios (Appendix A). The first is a modified version of the Asian disease (Kahneman & Tversky, 1979). The second scenario is a money-involved risk situation, in which higher risk can bring more potential profit. The third and fourth scenarios are hypothetical situations, in which outcomes had either a firmly determined (50%) or an unknown probability. We also presented subjects with 2 moral dilemmas (Appendix B) and a probability matrix, in which they had to report the likelihood of a particular event happening (Appendix C).

Subjects were unaware of the different experimental conditions and believed they were simply filling out a risk survey. After data collection, we asked subjects to share their experience with the survey. None of them reported having identified the true purpose of the study. Therefore, we assume that their answers were unbiased.

Results and Discussion

Risk scenario 1: Ebola

German Natives showed no significant difference χ2 (1, n = 104) = .277 p = .59 in their responses to the modified version of Kahneman and Tversky’s (1979) Asian disease (Fig. 1a). Data show that the language they were using did not influence their decisions. English natives responded differently to the same risk scenario. We found a significant effect χ2 (1, n = 16) = 4,267 p = .039, indicating that English natives, answering in German, were significantly more risk-seeking than their counterparts that were answering in English (Fig. 1b). When filling out the survey in German, English natives responded in a more risk-seeking manner, choosing the riskier outcome 70% of the time. These results contradict the findings of a recent similar study (Keysar, et al.). They speak against our initial hypothesis that thinking in a foreign language makes you more rational and emotionally distanced.

Abbildung in dieser Leseprobe nicht enthalten

Fig. 1a. German natives’ (n=104) responses to risk scenario 1: Ebola Fig. 1b. English Natives’ (n=16) responses to risk scenario 1: Ebola

Risk scenario 2: Lottery

In the money-related risk scenario, German Natives were significantly more risk seeking when they were answering in English: χ2 (2, n = 104) = 6.226, p = 0.4. As Fig. 2a indicates, German Natives were more conservative and risk-averse in their mother tongue, and chose only between the $750 and the $500 option. This tendency does not hold true for the foreign language: When answering in English, they chose the higher-risk options more often. These findings are consistent with the results from Risk scenario 1 (see above).

English Natives chose very often the $750 outcome (moderate risk), when filling out the survey in German. However, unlike German Natives, they also chose the higher-risk outcome in their native English (Fig. 2b). Because of the small sample size, their answers did not come out significant: χ2 (2, n = 16) = 2.424, p = .298.

Abbildung in dieser Leseprobe nicht enthalten

Fig. 2a. German Natives’ responses to risk scenario 2: Lottery Fig. 2b. English Natives’ responses to risk scenario 2: Lottery

[...]