The authors have declared that no competing interests exist.
Conceived and designed the experiments: MG MBP. Performed the experiments: MG. Analyzed the data: MG MB. Wrote the paper: MG MBP MB.
Bilinguals often switch languages depending on what they are saying. According to the Emotion-Related Language Choice theory, they find their second language an easier medium of conveying content which evokes strong emotions. The first language carries too much emotional power, which can be threatening for the speaker. In a covert experiment, bilingual Polish students translated texts brimming with expletives from Polish into English and vice versa. In the Polish translations, the swear word equivalents used were weaker than in the source text; in the English translations, they were stronger than in the original. These results corroborate the ERLC theory. However, the effect was only observed for ethnophaulisms, i.e. expletives directed at social groups. It turns out that the main factor triggering the language choice in bilinguals is not necessarily the different emotional power of both languages, but social and cultural norms.
Over centuries, language and its influence on the functioning of human beings has evoked wide interest and fascination. Whorf’s famous idea of linguistic relativity assigns to language a decisive role in affecting human thinking, resting on the assumption that speakers of different languages will perceive the world differently: “all observers are not led by the same physical evidence to the same picture of the universe, unless their linguistic backgrounds are similar”
Those who have contact with bi- or multilinguals often notice that it is easier for the latter group to express certain content in one language, while other content is more willingly conveyed in the other. What does this preference depend on? It has been noticed and proven that it is mostly emotional topics that cause slipping from one language to another. Kim and Starks
An interesting real-life proof for ERLC being triggered by the greater emotionality of L1 can be found in the diary of famous bilingual writer Eva Hoffman, who describes her dilemmas concerning the choice of language for writing her memoirs – her native Polish, or English, the language of the environment she was in at the time when the events she wanted to account for were happening. She finally chooses English, as ‘the language of the present’, but she emphasizes that it is not ‘the language of the self’, so for her writing the diary will be ‘an impersonal exercise’
ERLC is also present in bilinguals’ everyday lives. Scenes from multicultural marriages serve here as a good example, as it was repeatedly observed that in disagreements the spouses automatically switch to their L1 as a language more natural to express their emotions (
The existence of ERLC has been proven in numerous studies. The overall conclusion that follows from research on language-switching behaviour is that code alternation in emotional contexts may serve a distancing function, as expressing emotions in L1 would be too burdensome and anxiety-provoking, especially for bilinguals who learned their second language later than in early childhood
The greater emotional overtone of L1 might be connected with the first emotion experience. Altarriba
Psycholinguistic explorations, psychoanalytic case studies, and finally their own real life experiences unequivocally indicate that for bilinguals the two languages differ in their emotional impact (especially when the second language has been learnt after puberty), the first being the language of personal and emotional involvement, and the second the language of distance and detachment (or at least the language having less emotional influence on the individual). The phenomenon of ERLC is not limited to highly reflective, sensitive bilinguals who are greatly sensitive to emotional context, but to some extent is experienced by all.
A particularly interesting kind of language behaviour is swearing and using taboo words. Not only does it evoke a lot of excitement and strong feelings, but it also enjoys much more attention from language learners than other vocabulary items, or grammatical structures – sometimes swear words are the only lexis they know in the language. In all languages swear and taboo words are a rather delicate and controversial field, which even native speakers may sometimes find hard to navigate. It is because the rules of using these expressions may differ depending on the context. Also attitudes towards these speech acts vary according to the situation and the speakers. In some circles and subcultures they are tolerated, whereas in other milieus and more formal contexts they may be considered unacceptable. Some countries (e.g. Australia or the United Arab Emirates) have banned swearing in public and impose fines for doing so. Taking all this into account, it is not surprising that non-native speakers face an especially challenging task entering this vocabulary domain. Even though they have acquired the rules of using strong language in their L1–their point of reference–these rules are not universal and cannot be straightforwardly transferred from one language to another. While non-native speakers acquire foreign-language swear words relatively fast
This rule does not apply to highly proficient multilinguals. According to quantitative and qualitative data gathered by Dewaele
In subsequent research on pentalinguals, Dewaele
Dewaele
The above-described research on ERLC while swearing reveals that in the case of expletives, the language change does not have to be motivated by differences in the emotional power of both languages, but by the social norms regulating the use of such expressions.
The term ‘ethnophaulism’ was introduced by Roback
Based on existing literature one could derive two alternative hypotheses about the easiness of swearing in L1 and L2. According to the traditional formulation of ERLC, using more explicit emotional language in L1 should be much easier than in L2, due to the higher emotional spontaneity that bilingual language users express in their native language. Alternatively, as we discussed above, long cultural learning and socialization make expressions in L1 extremely prone to normative influences. Thus, expressing any socially unaccepted words and utterances (such as swearing) should be more difficult in L1 than in L2. In order to test whether it is norms or emotional spontaneity that have a greater impact on the willingness to swear among bilingual speakers, we designed an experimental study in which participants (Polish-English bilingual students) were asked to translate two equally offensive texts (from L1 to L2 or from L2 to L1), and to assess their offensiveness.
In order to capture the normative aspect of ERLC, we decided to confront the participants with words that are subject to higher (ethnophaulisms) and lower (general swearwords) norms of political correctness. Ethnophaulisms are a specific type of swearwords, as they relate to specific social groups and are subject to a greater number of norms and risk of disapproval. Their use means flouting political correctness and reflects not only the speaker’s lack of manners, but also their attitude towards other social groups. Therefore, in the case of such utterances, the hypothesised ERLC effect could be different than in the case of regular swear words.
Thus, the design was a 2×2×2 experiment with one between-participants factor (direction of translation: L1 to L2 vs. L2 to L1), and two within-subjects factors (rating of source material vs. rating of translated material; general swear word rating vs. ethnophaulism rating). The dependent variable was the ratings of offensiveness. If the emotional spontaneity hypothesis is true, in translations from L2 into L1 the target words should be more offensive than the source words, and the reverse should occur in L1 into L2 translations (Hypothesis 1). If the normative influence hypothesis is true, then in translations from L2 into L1 the target words should be rated as less offensive than the source words, and the reverse should occur in L1 into L2 translations (Hypothesis 2a). Additionally, if the normative influence hypothesis is true, L2 would make it easier for bilinguals to use normatively constrained words (such as ethnophaulisms), but should not influence the use of other expletives (Hypothesis 2b).
The participants of the present study were 61 students of Applied Linguistics and English Philology at the University of Warsaw, 11 men and 50 women at the age from 21 to 24. The subjects were picked randomly, belonged to various groups, majors and years of study. The students of both faculties were treated as Polish-English bilinguals, as their command of English is advanced and fluent. An advanced level of English had been required in the recruitment process; secondly, nearly all their courses take place in this language, and they use it on an everyday basis in various ways. Additionally, to verify their level of English the questionnaire included a measure of vocabulary knowledge.
The experiment is different from previous studies on swearing from the methodological point of view. To date the hypothesis that swearing is easier in L2 has been based chiefly on self-reports of bilinguals, bilingual writers and observers of bilingual people. In the present study the subjects did not know what was being investigated, so they did not take note of the fact that language change might affect their expression of emotions. There is often discrepancy between human behaviour and the perception of this behaviour, which is why the current study was designed as a covert experiment. This helped gather direct data, unbiased by the subjects’ willingness to present themselves in a positive light.
The study was conducted during two weeks of classes at the university and consisted of three stages. In the first part, the subjects were given a task: a short text which they were to translate. They were randomly divided into two groups; one translated the text from L1 Polish into L2 English, and the other from L2 English into L1 Polish (both versions of the source text are included in
Under the text there was a question included to verify the subject’s knowledge of vocabulary, to make sure that it would not influence the translation. The time to do the first part was not limited, but the majority of the participants finished it in under 20 minutes.
In the next stage, the subjects were given a list of the swear words which had appeared in the
The participants provided verbal informed consent to take part in the study. Such a form of consent is customarily used in Poland in studies on adult student samples. The consent procedures were detailed in a description submitted to the institutional review board (Ethics Committee of the Faculty of Psychology, University of Warsaw), where they were granted final approval in June 2012. The participants were granted full anonymity of the data gathered for the analyses. The final results were subsequently available to them by request (a possibility of which they had been informed before taking part in the study).
The first step was to calculate the mean offensiveness of the source and target words in both groups. Then, using Repeated Measures ANOVA, the levels of offensiveness between the source and the target words were compared for both groups. The within-subject variable was the words (source vs. target) and the between-subject moderator was the direction of translation (L1→L2 vs. L2→L1). The dependent variable was the level of offensiveness. The analysis showed a significant interaction between the source and target words and the direction of translation,
To verify Hypothesis 2, the words were split into “ethnophaulisms” and “other swear words”. Ethnophaulisms included all and only those swear words from the text that referred to social or ethnic groups. The interaction was investigated with the use of Repeated Measures ANOVA. The within-subject variables were i) the words (source vs. target) and ii) the type of insults (ethnophaulisms vs. other swear words); the between-subject moderator was the direction of translation (L1→L2 vs. L2→L1). The dependent variable was the level of offensiveness. The analysis showed a significant interaction between the source and target words, direction of translation, and type of words,
in the translations from L1 into L2, the target ethnophaulisms were more offensive (M = 2.79, SD = .42) than their source words (M = 2.43, SD = .59),
in the translations from L2 into L1, the target ethnophaulisms were less offensive (M = 2.12, SD = .53) than their source words (M = 2.87, SD = .51),
These effects did not appear in the case of other swear words. The results are presented in
Due to their explicit language which may have the potential of offending some readers, the most extreme examples of the effects observed are presented separately in
Interestingly, there was a similar case among the English translations (L1→L2), in which some of the swear words were softened, at odds with Hypothesis 1 (reproduced in point 1 of Translation Samples S2). It can be a premise that some people feel too limited by social norms and they are not able to breach them.
In the remaining translations into English, the swear words supplied by the subjects were more offensive than in the source texts. In some cases not only were the existing swear words strengthened, but also some more expletives were added. Some representative examples of the English translations are presented in point 2 of Translation Samples S2, with the relevant expressions in bold.
The ERLC theory in the case of swearing implicates two different hypotheses: it is easier to swear in a foreign language either because it distances the speaker emotionally from the expressed content or it exempts them from the social constraints concerning using swear words. The research confirmed the normative influence hypothesis. Bilinguals find it easier to swear and to offend out-groups in L2. Even though the participants were bilingual, the results of the study can reasonably be extrapolated to all people who know a foreign language at a communicative level. It transpires that the foreign language exempts us from our own or socially imposed norms and limitations and makes us more prone to swearing and offending others.
On the one hand, our findings are consistent with those cited in the literature overview. On the other, they bring a fresh perspective into research on language choice in bilinguals. They prove the distancing function of L2 claimed by Bond & Lai
In relation to the ERLC study, the question arises whether the results were not caused by the context of the experiment–it was conducted in Poland by a Polish experimenter. Could it have been the fact that the subjects were in their homeland–the place where L1 is spoken – that made them too constrained by the applicable norms and as a result they softened the words when translating into Polish? Given that the experiment was carried out at university in a class which would normally take place in the English language, this concern can be dismissed. Still, it is worth investigating whether, had the study been conducted in a different context, the subjects would have felt equally or more comfortable swearing in their L1.
There is also the problem of different constraints present in the Polish and English cultures, which are conveyed through the language. One may argue that in English-speaking countries it is more common to hear swear words in public, e.g. in the media. On the other hand, political correctness emanating from the Anglo-Saxon culture is definitely more rooted in English than in Polish. The problem was explored in a study by Bilewicz and Bocheńska
The normative influence discovery undoubtedly opens new perspective in the research on language choice in bilinguals. It proves that there is something more to it than the straightforward emotional distancing explanation. It implies that the use of language does not depend on the speaker’s inner experiences and feelings, but, at least in the case of swearing, rather on the social reality in which they function.
(RTF)
(RTF)
(RTF)
The authors would like to thank all those who contributed to the completion of this article. We express our thanks to Barry Keane, Ph.D., Zbigniew Możejko, Ph.D., Anna Pochmara-Ryżko, Ph.D., and Izabela Szymańska, Ph.D. from the Institute of English Studies of the University of Warsaw for allowing us to carry out the experiment in their classes. Finally, our deep sense of gratitude goes to the students for their agreement to participate in the study.