Ask most people which language they think best in, and they won't hesitate: their mother tongue, the one that feels like bare skin rather than clothing. A foreign language feels clunky by comparison, a medium that forces you to stumble over grammar and lose the shading you take for granted.
The assumption follows naturally: high-stakes decisions deserve your sharpest cognitive tools, and that means your native language. Boaz Keysar, a psychologist at the University of Chicago who had spent years studying how context reshapes judgment, suspected the opposite was true, because he knew that decision-making relies on two intertwined systems and that the emotional one activates far more powerfully in a first language than in a second one learned under fluorescent lights at age fourteen.
What if that emotional distance was not a deficiency but a cognitive advantage that could strip away the very biases researchers have spent decades documenting?
The Experiments
In 2012, Keysar, doctoral student Sayuri Hayakawa, and Sun Gyu An ran six experiments testing this idea. They used one of the most famous demonstrations of irrationality: the Asian Disease Problem, designed by Kahneman and Tversky in 1981.
The problem presents two identical options framed differently: in the "gain frame," Option A saves 200 out of 600 people, and in the "loss frame," the same option means 400 will die. Same math, but opposite feelings. People are risk-averse when thinking about gains and risk-seeking when thinking about losses, a gap psychologists call the framing effect.
Keysar's team presented these problems to bilingual participants across four language pairs: Korean-English, English-Japanese, English-French, and English-Korean. In their native tongue, the framing effect appeared exactly as predicted. In the foreign language, it vanished. People chose identically regardless of how the options were framed, as if the emotional loading of gain-versus-loss language had been surgically removed from their decision-making process.
Two additional experiments tested loss aversion with real money, offering participants coin flips with positive expected value but lopsided payoffs (lose $5 or win $8, for instance). In their native language, they rejected many of these advantageous bets, but in a foreign language they accepted significantly more, behaving closer to what economists would call rational.
Why It Works
If a foreign language simply slowed people down, you'd expect more deliberation to produce different biases or new errors. That's not what happened. Instead, it selectively muted the emotional reactions that fuel irrationality.
In 2017, Hayakawa, Keysar, and colleagues at Universitat Pompeu Fabra published a follow-up in Psychological Science using process dissociation to separate two explanations. Did a foreign language make people think harder about consequences? Or did it dampen the gut-level horror of violating moral rules? Across six experiments the answer was the second: foreign-language use didn't boost deliberation but instead blunted the emotional alarm bells that normally override rational calculation.
This aligns with psycholinguistic research showing that a second language is processed more semantically and less affectively: skin conductance dips, mental imagery fades, and profanity that makes you flinch in English barely registers in your high-school French.
A Meta-Analysis Confirms the Pattern
A single lab can be wrong, which is why it matters that by 2021 enough independent groups had tested the foreign language effect for Circi, Gatti, Russo, and Vecchi to publish a meta-analysis in Psychonomic Bulletin & Review pooling 17 studies and 47 experiments. In the moral decision-making domain (38 experiments), the pooled effect was Hedges's g = 0.22, p < .0001, and in the risk-aversion domain (9 experiments) it was slightly larger at g = 0.28, p < .0001, with Egger's test finding no publication bias in either domain. The fail-safe N was 1,175: you'd need that many drawer-filed null studies to erase the moral-domain result.
One moderator mattered. When the native and foreign languages came from the same linguistic branch (German and English, for example), the effect disappeared entirely (g = 0.06, not significant), but when they came from different branches (Spanish and English, Korean and English), it was robust at g = 0.30, suggesting that linguistic distance amplifies the emotional distance that drives the bias reduction.
The Strongest Counterargument
The most serious challenge came from Oganian, Korn, and Heekeren in a 2016 paper in Journal of Experimental Psychology: Learning, Memory, and Cognition. Their claim: in two experiments with over 1,500 participants, the framing effect was reduced not by foreign-language use per se but by the act of switching languages, with participants showing reduced bias regardless of whether they switched into a foreign or native tongue. If true, the hero of the story is cognitive control, not emotional distance.
The field hasn't settled this. Corey et al. (2017) and Driver (2020) failed to replicate the language-switching explanation, and the Circi et al. meta-analysis excluded mixed-language designs to isolate the foreign-language variable, finding the effect held across dozens of experiments. Two mechanisms may coexist: emotional blunting and cognitive disfluency. Theorists care which one dominates. The practical result is the same.
What We Didn't Prove
Most participants in these studies were young university students with intermediate-to-advanced foreign-language proficiency. The question of whether the effect holds for older adults, low-proficiency speakers, or heritage bilinguals who grew up with two emotional languages remains open. A 2023 study of early Catalan-Spanish bilinguals found no foreign-language effect at all, which makes sense: both languages were acquired in emotionally rich environments, so neither one carried the clinical detachment of a classroom-learned tongue.
The effect sizes are small-to-medium (g = 0.22–0.30), which means a foreign language won't turn you into a Bayesian rationalist; it offers a nudge away from the emotional tail of the distribution rather than a total recalibration of your judgment.
And there are limits. Outcome bias and the representativeness heuristic appear undiminished in a foreign language (Vives, Aparici, & Costa, 2018). The foreign-language effect seems specific to emotionally-driven biases, not all cognitive shortcuts.
The Bottom Line
About 1.5 billion people worldwide speak a second language, many of them making consequential decisions in that tongue daily, and the intuition that they're thinking less clearly turns out to be wrong. For emotional decisions, they may think better.
United Nations delegates negotiate in second languages, immigrant entrepreneurs evaluate business risks in adopted tongues, and EU Parliament members draft policy in English, French, or German that isn't their own. The meta-analysis suggests these decisions may be partially inoculated against the biases that infect native-language thinking.
What You Can Do
If you speak a second language, try writing out the pros and cons of a major decision in that language before committing, because the translation effort itself isn't the point but the emotional cooling it creates is.
If you manage multilingual teams, reconsider the assumption that everyone should present and decide in their strongest language. For decisions involving risk assessment, sunk costs, or emotionally charged trade-offs, a shared second language may produce less biased group judgment.
If you're a monolingual decision-maker, the underlying principle still applies: anything that increases psychological distance from a choice reduces emotional bias. Writing about yourself in the third person, imagining you're advising a friend, or waiting 24 hours before deciding are all low-tech versions of the same cognitive distancing the foreign-language effect exploits.