Does thinking in a second language change the way people make moral decisions? Research has often suggested that it might. Bilingual speakers sometimes appear more willing to endorse outcome-focused choices when confronting difficult moral dilemmas in a foreign language, a pattern known as the moral foreign language effect.
A new eye-tracking study adds an important qualification. The clearest language-related differences may not always appear in the final choice. They may instead emerge in where people direct their attention while deciding.
Researchers studied 89 Chinese-English bilingual adults living in the United Kingdom as they responded to five moral dilemmas in either Chinese, their first language, or English, their second language. An eye tracker recorded moment-by-moment visual attention while participants rated their emotional response and chose between deontological and utilitarian options.
The results revealed different patterns of association across the two language contexts. In participants’ first language, greater attention to the moral question was associated with stronger self-reported emotion, while longer initial viewing of that question was associated with a greater likelihood of choosing the deontological option. In the second language, broader fixation patterns across several parts of the task were associated with utilitarian choices.
Yet the study did not find a clear overall behavioural moral foreign language effect in the final choices. That distinction matters. Language may alter how people visually and cognitively engage with a moral problem even when it does not reliably change which answer they ultimately select.
Watching moral reasoning unfold
The study recruited 90 Chinese-English bilinguals, with one participant excluded because of technical problems with the eye-tracking equipment. The final sample therefore included 89 adults, 75 women and 14 men, aged 22 to 40 years. Their average age was 25.85 years.
Participants had begun learning English at an average age of 7.89 years and were living in the UK when the research was conducted. Eligibility required at least intermediate English proficiency. On 10-point self-ratings, average English reading proficiency was 7.42, listening was 7.12, speaking was 6.06 and writing was 5.92.
Each participant encountered five moral dilemmas, including three unrealistic sacrificial dilemmas and two situations designed to resemble dilemmas that could occur in real life. After reading each scenario, participants rated their emotional intensity on a 10-point scale and then made a binary moral choice.
The researchers distinguished between a deontological response, which prioritises moral duties or constraints, and a utilitarian response, which prioritises outcomes such as maximising overall welfare. This distinction is commonly used in experimental work involving sacrificial dilemmas, although real moral reasoning is more complex than a binary laboratory choice.
Eye movements were recorded at 1,000 Hz using an EyeLink 1000 Plus system. The researchers focused on three areas of interest shown after each dilemma: the emotional-intensity scale, the moral question and the moral-choice scale. They measured initial dwell time, total dwell time, fixation count and regression-path duration. Mixed-effects linear and logistic regression models were then used to examine how these attention measures related to reported emotion and moral choice.
First-language attention tracked emotional intensity
In the Chinese first-language condition, a higher number of fixations on the moral question was associated with greater self-reported emotional intensity. The result suggests that, within the first-language trials, stronger visual engagement with the morally decisive question accompanied stronger conscious emotional responses.
The researchers caution against treating fixations as a direct measurement of emotion. Eye movements can reflect several overlapping processes, including comprehension, re-reading, uncertainty and response preparation. The finding is therefore an association between attention and reported emotional intensity, not evidence that additional fixations caused stronger emotion.
A second first-language pattern concerned the type of moral choice. Longer first-run dwell time on the moral question was associated with a greater likelihood of selecting the deontological option. First-run dwell time captures the initial period of continuous attention before the eyes leave an area, making it an earlier measure than total viewing time that includes later revisits.
Together, these first-language results are consistent with a relatively close relationship between attention to morally salient wording, emotional evaluation and duty-based judgement. They do not establish that first-language reasoning is simply emotional or intuitive. The authors note that deontological decisions themselves can involve deliberation and conflict.
Second-language utilitarian choices involved broader visual engagement
The pattern was more distributed in English, the participants’ second language. A greater number of fixations on the emotion scale, the moral question and the moral-choice area was associated with a higher likelihood of choosing the utilitarian option. Longer regression-path duration on the emotion scale was also associated with utilitarian decisions.
The researchers interpret this cautiously as a pattern consistent with broader attentional engagement. Utilitarian reasoning requires comparing consequences and weighing alternatives, so repeated inspection of multiple task components could reflect more distributed evaluation. However, eye tracking cannot establish the precise mental process behind each fixation.
Another second-language result moved in the opposite direction. Longer total dwell time on the emotion scale was associated with a greater likelihood of choosing the deontological option. This suggests that different eye-movement measures may capture different routes through the decision. More numerous fixations across the task accompanied utilitarian choices, while sustained attention to emotional evaluation accompanied deontological choices.
This is an important reminder that attention is not a single quantity. Looking repeatedly across several areas and remaining focused on one area for a longer period can reflect different forms of processing, even when both increase the total amount of visual engagement.
The final decisions did not show a clear foreign-language effect
Perhaps the most important qualification is that descriptive patterns in participants’ final choices did not reveal a clear behavioural moral foreign language effect. The second-language condition therefore did not simply produce a robust overall shift toward utilitarian decisions.
The study instead suggests a subtler possibility. People can arrive at broadly similar distributions of final choices while allocating attention differently during the process that produces those choices.
That helps reconcile a mixed research literature. Earlier studies have sometimes found that using a foreign language increases utilitarian responses, particularly in emotionally charged personal dilemmas. Other studies have failed to reproduce the effect or have found that its size depends on language pair, proficiency, cultural background, immersion and the type of dilemma used.
The new findings do not settle that debate. Instead, they show why relying only on the final yes-or-no response can conceal meaningful differences in the route to a decision.
Why a second language might change the route rather than the destination
A first language is usually embedded in a person’s earliest emotional, social and cultural experiences. Moral words encountered in that language may therefore connect more readily with autobiographical memories, learned norms and affective associations. A second language can acquire many of the same meanings, especially through long-term immersion, but those associations may be organised differently.
The present study is compatible with that possibility without proving it. In the first language, attention to the moral question was linked to emotional intensity. In the second language, utilitarian decisions were associated with attention spread across several components of the task. These patterns could reflect different ways of accessing and integrating morally relevant information.
They should not be reduced to the claim that people feel in one language and reason in another. The second-language trials also showed a relationship between sustained attention to the emotion scale and deontological choices. Likewise, first-language deontological decisions need not be automatic or unreflective.
The more defensible conclusion is that language context may influence how attention, emotion and evaluation interact during moral judgement.
Important limitations
The sample was moderate in size and included only one language pair. All participants were Chinese-English bilinguals living in the UK, where frequent use of English could change the emotional resonance and processing of their second language. Results should not automatically be generalised to other bilingual populations or language combinations.
The study also analysed the first- and second-language conditions in separate statistical models rather than testing all language differences within one unified model. The apparent contrasts between languages should therefore be treated descriptively. A significant predictor in one language and a non-significant predictor in another does not by itself prove that the predictor’s effect differs significantly between the languages.
The binary moral-choice format simplifies decisions that can involve competing duties, intentions and consequences. Five dilemmas cannot represent the full range of real moral life, and participants knew they were completing a laboratory task rather than facing actual consequences.
Finally, eye movements are not process-pure measures. A fixation can indicate attention, difficulty, uncertainty, reanalysis or preparation to respond. Emotional intensity was also measured through self-report after the dilemma. The design cannot establish whether attention changed emotional evaluation, emotional evaluation redirected attention, or both processes influenced one another.
Moral language may shape the process before the choice
The study’s contribution lies in moving the foreign-language debate away from a single behavioural outcome. Asking whether bilinguals become more utilitarian in a second language may be too narrow if language changes the processing route even when the eventual answer remains similar.
For these 89 bilingual adults, first-language moral attention was more closely associated with reported emotional intensity, while second-language utilitarian choices were associated with a broader pattern of visual engagement. At the same time, there was no clear overall foreign-language shift in final moral decisions.
That combination makes the result more informative than a simple claim that one language makes people more rational or more emotional. Moral judgement appears to emerge from interacting attentional and evaluative processes, and the language in which a dilemma is encountered may alter how those processes are organised.
Source Information
Study: Eye-tracking emotional responses and moral decisions in bilingual contexts
Authors: Irini Mavrou, Andrea Révész, Andreas Kyriakou and Sin Manw Sophia Lam
Journal: Cognition, Volume 275, Article 106600
Publication: October 2026
Method: Preregistered laboratory eye-tracking study using five moral dilemmas and mixed-effects regression models
Sample: 89 Chinese-English bilingual adults living in the United Kingdom
DOI: 10.1016/j.cognition.2026.106600
