Some things are always under construction.

Empirie

There are many anecdotal examples described in the literatire, testifying that the phenomenon of interest is not a single impression of the curent study's authors. Indeed, several researchers have noticed and described the subjectively experienced difference (see Francis, 1999, Anooshian & Hertel, 1994, for reviews). For example Anooshian & Hertel (1994) cite an example of a woman who grew up in a Spanish-speaking home and learned English after eight years of age. As an adult, she used mostly English at home and at work but she prayed in Spanish because praying in English never “felt right”. Sechrest et al. (1968) reported that married Filipinos used Tagalog for intimate expression, even though English was habitually spoken at home.

Besides, there are also studys that were designed to demonstrate and thus further investigate this issue, just like the present study. They focused mainly on the difference in processing taboo-words in bilinguals, as taboo-words are rather emotionally chargerd. For example, Gonzalez – Reigosa (1976) measured state anxiety per questionnaire after Spanish-English bilinguals (L1 – Spanish) read aloud lists of neutral Spanish, of taboo English and taboo Spanish words. Subjects reported more state anxiety after saying taboo-words in their mother tongue. The author interprets the finding in terms of highened emotional arousal evoked by taboo-words in L1. However, participants were only Spanish native speakers, i.e. the emotional arousal that the English taboo-words would evoke in English native speakers was not assesed. Besides, questionnaires are a rather indirect measure of emotional experience. Thus, due to the methodical deficits, the Gonzalez' study could be regarded as an evidence only for the fact that the notion of lessend emotionality is common to many researchers.

Another operationalisation of the same idea was implied by Anooshian & Hertel (1994). They based their research on the idea that emotional words are better recalled, because they are processed more intensively. Thus, they hypothesized that emotional words would be better recalled than neutral words only in the first language of participants, i.e. that in L2, there wouldn't be a difference between emotional and neutral words, because they are equally unemotional, thus to equal extents elaborized during processing. Participants were Spanish-English bilinguals (Half were Spanish and half – English native speaking). They saw half of the words in Spanish, half in English; a half of the words was affectively neutral, and a half was emotional. Their task was to rate each word for ease of pronunciation, implied activity, or emotionality. Then, subjects completed an unexpected free-recall test.
To analyze the results, Anooshian & Hertel compared the recall rates within each language for the emotional and for the neutral words. According to their expectations, they obtained a significant difference only in the native language of participants. Unfortunately, the authors conduct analysis only within and not between the two languages, “because it is difficult to interpret native versus second language difference” (Anooshian & Hertel, 1994, p. 509). However, the experimental results reveal a confusing difference in the recall rates of neutral words between the two languages. If one would compare the recall rates for neutral words between the two languages, it seems that subjects recalled neutral words in their mother tongue worse than in their second language. If the difference is significant, it would mean that neutral words in the second language are more emotional than neutral words in the first language, when one uses the study's main notion. Or, alternatively, that there were other processes involved, besides words' emotionality, that made up the differences in the whole study.

The latter study was replicated and extended by Aycicegi & Harris (2004), who compared recall to recognition and included different types of emotion words (positive, negative and neutral items, taboo-words and reprimands). The stimulus presentation modality (visual vs. auditory) was also varied. Participants (42 Turkish-English bilinguals, L1 – Turkish) rated each stimulus word for unpleasantness. Immediately afterwards, a half completed a surprise recall test and a half – a surprise recognition test. Emotionality effects were present, but in both L1 and L2. The authors propose that in the recall and recognition paradigms, the novelty and unusualness of the stimuli were the main factors influencing performance.
Indeed, maybe recall or recognition paradigms are not the best means of assessing the emotional power of words in different languages. For example Eysenck & Keane (1999) mark that “there are various strategies that can be used in order to recall or recognize stored information” (p. 114).

A further investigation on the same idea of greater emotionality of words in L1, but with a quite different methodology, was made by Harris, Aycicegi & Gleason (2003). Namely, this time a psychophisiological measure, skin conductance measures, was implemented. Participants were again native Turkish speakers with English as a second language. Stimuli were positive, negative and neutral words, taboo words and reprimands (e.g. "Shame on you!", "Go to your room"). They were presented visually (on a computer screen) or auditorially (via the computer loudspeaker).
The best results were found for the auditory condition: both reprimands and taboo words produced greater SCR's in Turkish than in English, being in line with the research hypothesis. The authors propose that “auditory language may be more closely tied to emotional arousal than visual language" (p. 565). However the authors note that “the auditory qualities of the stimulus, such as tone, cadence, accent, and participants' knowledge that the voice is speaking their L1, may be the critical practice in eliciting high autonomic arousal to auditory stimuli in the L1”. Indeed, long auditory reprimands had higher SCR's than short auditory reprimands when presented in Turkish (but not in English). This difference suggests that non-lexical aspects of auditory stimuli may actually have played an important role in the study. The auditory items were recorded by a female native speaker of each language. For most items a neutral tone was employed. For reprimands, a lightly admonishing tone was used, appropriate to the meaning of the reprimand (p. 568). Unfortunately, Harris at al. don't report any pretests on their auditory materials, that would grant for equal quality of the stimuli. It could be, for example, that the native speaker that read aloud the Turkish taboos and reprimands used a more expressive intonation than the native speaker of the English stimuli. Besides, all participants in the study were Turkish native speakers, thus we have no information if the reverse effects would be evident for English natives. Hence, as there is no evidence for comparable emotional strengh of the stimuli in the two languages, the results from the auditory condition could not be interpreted in terms of the hypotheses.

In the visual modality condition, significant SCR-difference between L1 and L2 was obtained only for reprimands, i.e. (surprisingly) not for aversive words such as "cancer" and "war". The authors suggest that in the context of expletives and sexual terms, the arousing effects of negatively valenced words were reduced. Hence, further investigation is needed to extend the finding about the differential influence of reprimands upon other lexical stimuli. As reprimands “resemble taboo words in being associated with personal threat but are unambiguously acquired in early childhood” (p. 564), it is not clear whether the obtained results don't reveal a genuine phenomenon concerning only that specific type of emotional expressions.

In a subsequent study with Spanish-English bilinguals, Harris (2004) found again the same result pattern only for reprimands, but solely for late English learners. None of the other stimulus categories included in the experiment elicit different SCR’s in the two languages. Harris (2004, 2006) draws from these results a conclusion of increased emotional arousal in L2 for late bilinguals (and overall similarity of the emotionality of the two languages for early bilinguals). Again, it is doubtful if such a general conclusion is feasible, as the differences in emotionality were obtained solely for reprimands.
However, the main effect for language (Harris et al., 2003) indicates stronger skin conductance in Turkish than in English. The authors discuss this, pointing that this difference may reflect the greater emotional associations of Turkish words.

posted by gargichka @ 9:20 AM