Taboo words and reprimands elicit greater autonomic reactivity in a first language than in a second language

Second language speakers commonly acknowledge that taboo terms can be uttered with greater ease in their second language (L2) than in their first language (L1). To investigate this phenomenon psychophysiologically, 32 Turkish–English bilinguals rated a variety of stimuli for pleasantness in Turkish (L1) and English (L2) while skin conductance was monitored via fingertip electrodes. Participants demonstrated greater autonomic arousal to taboo words and childhood reprimands (“Shame on you!”) in their L1 compared to their L2. This finding provides quantifiable support for the subjective experiences of L2 speakers. Bilingual speakers have reported that it is easier to say some kinds of emotionally charged words or expressions in a language that is not their native language. In this study, we explored the possibility that this difference between a first language (L1) and a second language (L2) has psychophysiological components. In particular, we hypothesized that taboo words in a native language would elicit stronger skin conductance responses than similar taboo words learned later in life in an L2. We also hypothesized that emotional expressions learned early in life in an L1 (such as the kinds of reprimands young children hear) would evoke greater physiological responses, whether the individuals experiencing these responses were fully aware of them or not. A number of authors have remarked that taboo words appear to generate less anxiety when spoken in a foreign language (Ferenczi, 1916; Greenson, 1950; Javier, 1989). Bilingual speakers also feel freer to discuss embarrassing topics  2003 Cambridge University Press 0142-7164/03 $12.00 Applied Psycholinguistics 24:4 562 Harris et al.: Autonomic reactivity in L1 and L2 in their L2 (Bond & Lai, 1986; Sechrest, Flores, & Arellano, 1968). Anooshian and Hertel (1994) noted that age of acquisition of a language appears to be more important than proficiency for at least some cases of how the emotional resonance of a language is experienced. They cite the example of a woman who grew up in a Spanish-speaking home and learned English after 8 years of age. Although as an adult she was most proficient in English, she prayed in Spanish, because praying in English never “felt right.” Studies of code-switching report that the L2 is often used to increase emotional distance (Altarriba & Rivera– Santiago, 1994; Bond & Lai, 1986; Gonzalez–Reigosa, 1976; Gumperz & Hernandez, 1971; Javier & Marcos, 1989). In contrast, the native language is more often the language of emotional expressiveness (Dewaele & Pavlenko, 2002). For example, Sechrest et al. (1968) reported that married Filipinos used Tagalog for intimate expression, even though English was habitually spoken at home. One way to measure the emotional impact of words is through their effect on autonomic reactivity. Electrodermal monitoring (the psychophysiological technique that records skin conductance responses, formerly referred to as galvanic skin response) has long been used to measure differential responsiveness to words. Studies have shown that among monolinguals, emotion and taboo words elicit greater skin conductance responses compared to neutral words (Gray, Hughes, & Schneider, 1982; Dinn & Harris, 2000; McGinnies, 1949). Electrodermal recording has not been previously used to study differential reactivity in the languages spoken by bilingual individuals. The goal of the current study was to determine if electrodermal recording would demonstrate that words with equivalent semantic content in an L1 and L2 generate different degrees of autonomic reactivity. Relevant research literature includes studies of taboo words and emotion terms in monolinguals and bilinguals. Taboo words are known to be the most emotionally evocative of language stimuli (Gray, Hughes, & Schneider, 1982; Jay, 1992; Matthews & MacLeod, 1985; McGinnies, 1949; Nothman, 1962; Van Lancker & Cummings, 1999). Researchers have used taboo-word presentation to study the working of the brain’s emotional systems, because processing of taboo words is believed to activate the amygdala, a structure in the limbic system involved in emotion processing and emotion-mediated learning (Labar & Phelps, 1998). MacKay et al. (2002) used taboo words in a Stroop task. Taboo words were displayed in a salient color, the task being to name the ink color and ignore the word. Participants were slower to name the color of the taboo words compared to the color of the neutral words. This “Stroop interference effect” presumably occurs because the anxiety associated with taboo words (or people’s understanding of their status as socially stigmatized words) interferes with attending to the colornaming task. The Stroop interference effect was much stronger than the interference effect found for neutral words, and it persisted for over 100 trials. Aversive words (such as cancer, death, stabbed) resemble taboo words in that they produce orienting and attentional effects (Hill & Kemp–Wheeler, 1989; Wischner & Gladis, 1969), although the magnitude is typically less than that of taboo words or words of strong personal significance (Crane, Dieker, & Brown, 1970). Compared to neutral words, aversive words are easier to identify as Applied Psycholinguistics 24:4 563 Harris et al.: Autonomic reactivity in L1 and L2 words in a lexical decision task (Hill & Kemp–Wheeler, 1989). Both clinical and nonclinical subjects were better able to form associations to aversive words than to neutral words (Wischner & Gladis, 1969). Comparing emotional responsiveness to words in an L1 versus an L2 has been examined in relatively few laboratory studies, and only one used taboo words. Gonzalez–Regiosa (1976) asked Spanish–English bilinguals to read lists of 10 Spanish taboo words and 10 English taboo words. Following the reading of each list, they rated themselves on their degree of anxiety using a questionnaire sensitive to “state anxiety” (a transient state). Participants also completed a questionnaire designed to assess their level of “trait anxiety” (i.e., whether they had an anxiety-prone personality). Participants rated themselves as more anxious after reading the taboo words in their L1. This result held for both highand low-anxiety participants. Anooshian and Hertel (1994) examined differential emotional connotations in an L1 versus an L2 using a recall test. Recall of words is known to be influenced by emotionality (Rubin & Friendly, 1986). Anooshian and Hertel (1994) hypothesized that emotion words in the L2 would lack the emotional connotations that render words easy to recall. Participants first rated blocks of words for emotional intensity, pronounceability, and how much activity was inherent in the meaning of the word. After the rating task, participants received a surprise recall task. Emotionality effects were strongest for words when they had been rated for emotionality. More emotion words than neutral words were recalled in the L1, and an equivalent number of emotion and neutral words were recalled in the L2. This supports the authors’ hypothesis that only words in the L1 would show an emotionality effect. This supports the thesis of the current paper, which is that words in an L1 have greater emotional resonance than words in an L2. However, in an extension and replication of Anooshian and Hertel’s (1994) paradigm, two of the current authors found greater recall for emotion words in both the L1 and L2 (Ayçiçeği & Harris, in press). Indeed, a goal of the current paper was to determine if a psychophysiological measure would be a useful methodology for quantifying intuitions about the greater emotionality of words heard or read in the L1. Although no prior study has employed electrodermal recording to gauge emotional responsiveness, one study used event-related potentials (ERPs). In this psychophysiological technique, electrodes fixed to the scalp measure brain wave activity corresponding to stimulus presentation. It was interesting that the underlying rationale and design for this study was similar to our own: would emotional responsiveness vary for words in an L1 and an L2? Kim (1993) used P300 amplitude as her dependent measure, following the literature that this ERP component is sensitive to the incentive or emotional value of a stimulus. Kim recorded ERPs to neutral, positive, and negative English words from monolingual English speakers and Korean speakers who had varying degrees of Englishlanguage competency. However, no differences were found in P300 amplitude as a function of words’ emotional valence or participants’ English proficiency. Kim concluded that her emotion words (words with negative connotations such as steal and positive connotations such as truth) had probably been insufficiently arousing, because prior studies documenting the sensitivity of the P300 ampliApplied Psycholinguistics 24:4 564 Harris et al.: Autonomic reactivity in L1 and L2 tude to emotional stimuli used highly evocative stimuli such as pictures and slang expressions (Vanderploeg, Brown, & Marsh, 1987). Kim’s (1993) study motivates the current project by describing an area in which electrodermal monitoring may be a superior psychophysiological technology in comparison to ERPs. ERPs are sensitive to millisecond changes in brain electrical activity and change predictability in response to both lexical and grammatical factors (Weber–Fox & Neville, 1999). However, Kim (1993) found that ERP waveforms did not vary by emotional valence for monolingual or bilingual participants and so concluded that ERPs are not sensitive to the emotional valences of words. Skin conductance responses are known to be sensitive to differences in the same types of words Kim used in her ERP study (e.g., Dinn & Harris, 2000). Words’ emotionality may be a case where electrodermal recording is preferred to ERPs, despite the latter’s superior temporal resolution. Recent research has begun to elucidate hypotheses for why emotional resonance may differ between two languages (Bond & Lai, 1986; Dewaele & Pavlenko, 2002; Dewaele & Regan, 2001; Durst, 2001; Pavlenko, 1999, 2002; Schrauf, 2000; Wierzbicka, 1999). An L1 and L2 are frequently learned in different contexts. The L1 is learned in the context of family life and thus generally includes emotional extremes and usually also an individual’s earliest encounters with the gamut of human emotions (Schrauf, 2000). An L2 is frequently the language of schooling, work, and professional achievement and thus comes to be associated with emotional control, autonomy, and achievement (Bond & Lai, 1986; Dewaele & Pavlenko, 2002; Ervin, 1964). Bond and Lai (1986) reasoned that, because the L2 is usually mastered in more emotionally neutral settings than the L1, less arousal would be conditioned to L2 words. These authors predicted that bilingual interviewees would be more comfortable discussing embarrassing topics in their L2. Indeed, it was found that the interviewees spoke at greater length about embarrassing topics (but not neutral topics) when they were instructed to respond in their L2. Differential emotional context of language acquisition appears to be a sufficient explanation for greater emotionality of words and expressions in the L1. However, the implications for taboo words are not obvious. Historically, the anxiety-arousing effect of taboo words has been attributed to punishment from parents for saying these words during childhood (Ferenczi, 1916). This seems an overly simplistic view from a contemporary standpoint. In some households and some cultures, children acquire taboo words outside the home and learn their stigmatized status from peers. The broader concept of societal disapproval may be more important than punishment by parents. Taboo words learned in an L2 may actually be learned in a similar environment to taboo words in an L1 (i.e., both learned from peer or street culture) and thus may have similar emotional associations. According to this view, elecrodermal recording could reveal high reactivity for taboo words in both an L1 and an L2, given that reasonably proficient L2 learners will be aware of the stigmatized nature of these words. It thus seemed reasonable to identify emotional expressions that resemble taboo words in being associated with personal threat but are unambiguously acquired in early childhood as part of daily family life. Inspired by prior work on language routines Applied Psycholinguistics 24:4 565 Harris et al.: Autonomic reactivity in L1 and L2 and childhood language socialization (Gleason, 1985), we selected the category of childhood reprimands, including expressions like Go to your room! and Shame on you! The current study is the first attempt to examine autonomic reactivity to emotional expressions in the L1 and L2. We thus felt it was acceptable to include an exploratory variable, the variable of presentation modality. Comparing visual versus auditory presentation can be useful for future research for two reasons. Greater skin conductance responses might occur for auditory stimuli because the spoken language likely has more diverse and more numerous emotional associations than written language. Spoken words and phrases may activate specific memories. Spoken language is acquired before visual language (for L1 acquisition). To the extent that linguistic representations that are learned early become connected with emotional regulation systems (Bloom & Beckwith, 1989), auditory language may be more closely tied to emotional arousal than visual language. We also varied the modality of presentation for exploratory methodological purposes. For investigating languages that do not use the Latin alphabet, it could be convenient to restrict data collection to the auditory realm. For languages that use the Latin alphabet, typing in a list of words for visual presentation is less labor intensive than recording and digitizing auditory stimuli, thus reducing time to pilot and test a hypothesis.