Dissertations, Theses, and Capstone Projects

Date of Degree


Document Type


Degree Name





Virginia Valian

Committee Members

Martin Chodorow

Natalie Kacinik

Debra Titone

Subject Categories

Cognitive Psychology | Cognitive Science | Psycholinguistics and Neurolinguistics


bilingualism, foreign language effect, emotional blunting, availability heuristic, anchoring heuristic


Bilinguals sometimes make decisions in verbal tasks differently in their first (L1) and second (L2) language. This phenomenon is known as the foreign language effect (FLE), and it suggests strong connections between language and cognition. On the one hand, it is possible that L2 “blunts” emotional language. However, the FLE can be observed in non-emotional tasks. Therefore, it is possible that L2 requires more deliberate processing due to increased cognitive load, leading to more rational decisions. The support for each explanation is mixed.

In this thesis we propose looking for a single explanation for all instances of the FLE. After reviewing the existing research, we believe that we observe the following pattern: the FLE emerges in association-based tasks. Therefore, we suggest that the nature of the FLE is semantic. That is, association strength in L2 is weaker for all concepts, not just the emotional ones. Unlike L1, L2 is often acquired in an experientially impoverished classroom environment and is used less frequently, resulting in weaker links across the entire lexical system (Gollan et al., 2008).

The association strength argument can help address the relationship between emotion and cognition, as well as tying cognition to language. Research focused on bilingual participants and the FLE could help to first dissociate cognitive and emotional processes, and second to show that performance in non-emotional contexts can be affected by association strength between concepts in dominant and non-dominant languages. To achieve this two-fold goal, we carried out a series of experiments, manipulating emotional context and type of task.

This dissertation consists of four studies with the underlying goal of developing the idea that the FLE manifests in tasks involving semantics and association strength. In Study 1, we use a rating task to assess the subjectively perceived emotionality of words with different emotional valence in mono- and bilinguals. We found that bilinguals provide higher ratings for negative words and lower ratings for positive words compared to monolinguals irrespective of the language of the task. We suggest that associations may be weaker not just in bilinguals’ L2, but also in their L1, which they use relatively less frequently than monolinguals use their only language.

Since associations are a function of memory, Studies 2 and 3 examine the emotion-memory effect in monolinguals and in bilinguals in their L1 and L2. In Study 2 participants complete a recognition task containing emotional words. We hypothesize that the emotional words will be overrepresented in the responses of monolinguals and in bilinguals’ L1. Thus, we use Signal Detection Theory to account for both accuracy and potential false recognition. The results show a significant difference in bias for emotional and neutral words suggesting that participants were more liberal in judging emotional words. However, the bias did not vary across language groups, suggesting similar recognition effects in mono- and bilinguals.

Study 3 examines the emotion-memory aspect of FLE through the accessibility heuristic. Participants of different language backgrounds are presented with lists of words and asked to estimate the frequency of emotional words. We hypothesize that frequency estimates would be biased in monolinguals and in the L1 of bilinguals. Similar to Study 2, while the estimates of different types of words varied significantly, they did so in mono- and bilinguals alike. Emotional words appear to have a similar effect on memory in mono- and bilinguals and in bilinguals’ L1 and L2. We suggest that either the FLE may be task-dependent and could manifest in decision-making, but not in memory tasks. Alternatively, emotional words could have lower activation thresholds and do not require strong associations in order to affect cognitive processes.

Different representation of emotional words (and only emotional words) does not explain why the FLE sometimes emerges in non-emotional contexts. Therefore, in Study 4 addresses the FLE in a non-emotional association-based decision-making task by assessing the extent to which mono- and bilinguals are affected by the anchoring bias. Our results show that anchoring affects mono- and bilinguals (in L1 and L2) alike. However, anchoring may differ as a function of proficiency when confidence is factored in.

FLE research, by virtue of being a relatively new area of interest, is not devoid of inconsistencies and unresolved issues. From a theoretical perspective, research on the FLE can expand our knowledge in several fields by 1) providing new information about semantic organization in bilinguals’ L1 and L2, and in comparison to monolinguals; 2) expanding the field of decision-making by using L2 to dissociate cognition and emotional processing; and 3) since we argue that FLE is association-based (even for non-emotional stimuli), it can provide information on how knowledge representation overall affects decision-making. Therefore, it allows investigating the nature of the relationship between emotion, decision-making and a second language.