Bilinguals’ Recall and Recognition of Emotion Words

Recall of emotion words is superior to neutral words. Prior work reported in this journal (Anooshian & Hertel, 1994) found that this effect was absent in a second language. Words in a second language may thus lack the emotional associations of words acquired in childhood. To determine whether memory probes may be generally useful for assessing emotionality effects in a first vs. a second language, Anooshian and Hertel’s paradigm was extended in several ways. Recall was compared to recognition, and a variety of types of emotion words were studied, including taboo terms, and phrases likely to be learned in childhood (reprimands). Superior memory for emotion words was obtained in both the recall and recognition tasks, but this occurred in both the first and second language and indeed was stronger, for some stimuli, in the second language. This suggest that, even for bilingual speakers who acquire their second late (after age 12), words in the second language retain rich emotional associations. Emotion Word Recall and Bilingualism Page 3 Bilinguals’ Recall and Recognition of Emotion Words Ay e Ayçiçe i, Istanbul University Catherine L. Harris, Boston University Diverse reports over the last century support the idea that the first language is bilingual speakers’ choice for expressing positive emotions (Javier, Barroso & Muñoz, 1993). Sechrest, Flores and Arellano (1968) reported that married Filipinos used Tagalog for intimate expression, even though English was habitually spoken at home. Anooshian and Hertel (1994) noted that age of acquisition appears more important than proficiency in emotional expression, citing the example of a woman who grew up in a Spanish-speaking home and learned English after 8 years. Although English was her dominant language, she prayed in Spanish, because praying in English never “felt right.” These authors suggest that language learned in childhood may have greater emotional expressiveness. This is compatible with the view that emotional systems co-develop with early language use (Bloom & Beckwith, 1989). If the first language is the language of emotional expressiveness, the second language may be the language of emotional distance (Dewaele & Foth, 2002; Dewaele & Pavlenko, 2002; Marcos, 1976). Bilingual speakers feel freer to discuss embarrassing topics in their second language (Bond & Lai, 1986). Studies of code-switching describe how using a second language often serves a distancing function (Gumperz & Hernandez, 1971; Javier & Marcos, 1989). GonzalezReigosa (1976) discussed psychotherapy with Spanish-English bilinguals. Patients retreated into English to discuss anxiety-arousing topics, and used English for portraying a persona of selfconfidence, calm, and emotional reserve. Second language users commonly acknowledge that obscene and taboo words generate less anxiety when pronounced in a foreign language. SpanishEnglish bilinguals reported more anxiety after reading a list of 10 Spanish taboo words compared to reading a list of 10 English taboo words (Gonzalez-Regiosa, 1976). The current study investigated differences in emotional resonance of a first vs. a second language by adopting the methodology of Anooshian and Hertel (1994). These authors compared recall of emotion and neutral words which had been presented, in a rating task, in either a first vs. a second language. Recall of words is known to be influenced by emotionality (Rubin & Friendly, Emotion Word Recall and Bilingualism Page 4 1986). Anooshian and Hertel (1994) hypothesized that emotion words in the second language lack the emotional connotations which enhance recall. This hypothesis was confirmed with their study of bilingual speakers of Spanish and English, who differed in whether English or Spanish was the native language, and who acquired their second language after age 8. These authors found that emotional words presented in the first language were recalled more frequently than neutral words, but in the second language, emotion and neutral words were equally recalled. If the enhancing effect of emotion is language-specific, then words may be represented differently in different languages, as proposed by several researchers (Kolers & Roediger, 1984; Watkins & Peynircioglu, 1983). The current paper investigates the generality of Anooshian and Hertel’s results by using a more diverse set of emotion terms and varying the procedure. Anooshian and Hertel (1994) only compared emotional words and neutral words, and 13 of their 18 emotion words were positive (see Appendix A). Positive emotion may be more connected to early language learning, and thus the emotion-word advantage may be restricted to positive words in the first language. Indeed, not all theorists agree that a first language, but not the second language, shows enhanced recall of emotional words. Overall better memory for neutral words compared to anger words was observed in a recent study by Martins, Oliveira, and de Sousa (2003). These authors also found greater recall of anger words by bilingual speakers compared to monolingual speakers. This is a very different result from the superior recall of emotion words in L1 found by Anooshian and Hertel (1994), but the authors still attribute their results to heightened emotional processing of L1 compared to L2. They note that processing anxiety-provoking stimuli may induce a mood change. To avoid experiencing negative feelings, participants may minimize processing of aversive stimuli (Foa & Kozak, 1986; Mathews et al., 1989). If unpleasant stimuli are given less elaborate processing, they will be less accurately recalled, meaning a memory disadvantage for negative words compared to neutral words. If the emotional connotations of words in a second language are weaker than in the first, then the memory disadvantage for negative words will be weaker in the second language. Integrating the results of Anooshian and Hertel (1994) and Martins et al. (2003) suggests that L1 vs. L2 effects may differ for positive and negative words. We thus differentiated emotional words into separate categories of positive words (joy, mother) and negative words (anger, pain), and compared these to neutral words (table, column). We predicted that we would find a greater Emotion Word Recall and Bilingualism Page 5 emotion-memory advantage for positive words in the first language, but that negative words would have superior recall in the second language. Two additional emotion categories were included for exploratory purposes: childhood reprimands and taboo words (sexual and socially stigmatized terms). Both categories are anxiety-arousing. Childhood reprimands are phrases customarily used by parents to control children’s behavior, such as “Don’t do that” and “Shame on you!” Although Bloom and Beckwith (1989) have emphasized the connection between positive affect and very early language learning, childhood is of course also a time of negative emotion, particularly apprehension of parental censure. If a childhood context of emotion learning is the important factor for the recall advantage of emotion words, then childhood reprimands should show a recall advantage in the first, but not the second language. We included taboo words because these items typically generate a strong visceral response (Jay, 2000). They can be considered “super-emotion” words in terms of the diversity and strength of associated contexts and emotions. Indeed, researchers interested in the brain systems which mediate emotion have employed taboo words because of their ability to activate the amygdala, known to be a key subcortical structure for threat-detection (LaBar & Phelps, 1998). Prior work with monolinguals has found that taboo words are recalled better than neutral words. MacKay et al. (2002) argued that the superiority of recall for taboo words occurs because emotional reactions during encoding facilitate binding of the taboo words to its context. The current study extended that of MacKay et al (2002) by comparing recall of taboo words in a first vs. a second language. It could also be useful to compare childhood reprimands to taboo words. Many taboo terms are learned in middle childhood or adolescence. The concepts of sexual stigma and interpersonal slurs are highly relevant to adolescents and young adults, and may thus be equally relevant when learning a first vs. a second language. The procedure used by Anooshian and Hertel (1994) was varied in several ways beyond our examination of additional stimulus types. Half the participants received a recognition task instead of a recall task. For monolingual speakers, recall and recognition tests are influenced similarly by emotionality, with emotion words showing an advantage compared to neutral words (Rubin & Friendly, 1986). We thus expected similar emotionality effects for both recognition and recall tasks. However, the pattern may vary because recognition tasks benefit from familiarity, and Emotion Word Recall and Bilingualism Page 6 recall and recognition tests are impacted differently by frequency and distinctiveness (MacLeod & Kampe, 1996). Anooshian and Hertel (1994) presented words visually to subjects, via printed lists. Auditory presentation may lead to greater elaborative encoding, and may bring to mind more emotional associations because the interpersonal contexts of speech are plausibly more emotional than print. It thus seems worthwhile to determine if memory effects are the same regardless of modality. We thus varied auditory and visual presentation, using a mixed format for this factor as well as the factor of first/second language.