1.Introduction

Parliaments are supposed to feature ‘parliamentary’ (i.e. polite, respectful) language, which is often enshrined in their rules of procedure. However, even casual observers of the political scene realize that political struggle tends to produce aggressive verbal behaviours. This is true of the British Parliament (see, e.g. Harris 2001Harris, Sandra 2001 “Being Politically Impolite: Extending Politeness Theory to Adversarial Political Discourse.” Discourse and Society 12 (4): 451–472. ; Pérez de Ayala 2001Pérez de Ayala, Soledad 2001 “FTA and Erskine May: Conflicting Needs? Politeness in Question Time.” Journal of Pragmatics 33 (2): 143–169. ; Chilton 2004Chilton, Paul 2004 Analysing Political Discourse: Theory and Practice. London and New York: Routledge. ), this is true of the Polish Parliament (see, e.g. Kamińska-Szmaj 2007Kamińska-Szmaj, Irena 2007 Agresja językowa w życiu publicznym. Leksykon inwektyw politycznych 1918–2000 [Linguistic aggression in public life. A lexicon of political invectives 1918–2000]. Wrocław: Wydawnictwo Uniwersytetu Wrocławskiego.; Polkowska 2014Polkowska, Laura 2014 “Naruszanie zasad etyki poselskiej w latach 2001–2012 [Breaking the rules of MP ethics in the years 2001–2012].” Poradnik Językowy 1/2014: 61–70.), and this is also true of the European Parliament (Plug 2010Plug, H. José 2010 “Ad Hominem Arguments in the Dutch and the European Parliaments. Strategic Manoeuvring in an Institutional Context.” In European parliaments under scrutiny: Discourse strategies and interaction practices, ed. by Cornela Ilie, 305–328. Amsterdam: John Benjamins. ), which, in a sense, constitutes a hybrid of parliamentary traditions of many Member States. What makes debates in the EP special is that they are carried out in 24 languages, with numerous teams of simultaneous interpreters ensuring effective communication. The interpreters’ task is far from easy, the typical problems including breakneck speech rates, high information density and poor, non-native use of some languages, notably English (e.g. Cosmidou 2013Cosmidou, Olga 2013 “The European Parliament: A Temple of Multilingualism, a Pioneer in Interpreting ‘exploits’.” Gramma 19: 129–132.; Seeber 2017Seeber, Kilian 2017 “Interpreting at the European Institutions: Faster, Higher, Stronger.” CLINA 3–2: 73–90.).

What attracted my attention as a potential interpreting problem was extreme impoliteness, as exemplified vividly by Member of European Parliament (MEP) Nigel Farage’s speech of 24 February 2010. Farage ‘welcomed’ the newly elected President of the European Council Herman Van Rompuy saying, inter alia, that the latter had the “charisma of a damp rag and the appearance of a low-grade bank clerk”. This triggered a fairly comprehensive study (Bartłomiejczyk 2016Bartłomiejczyk, Magdalena 2016 Face Threats in Interpreting: A Pragmatic Study of Plenary Debates in the European Parliament. Katowice: Wydawnictwo Uniwersytetu Śląskiego.) examining how the pragmatic meaning of such statements is handled by interpreters, which will be outlined in the General Results. This paper is intended as a follow-up to explore a hypothesis that female interpreters are more inclined to mitigate impoliteness than their male peers.

A discussion of impoliteness in parliamentary debates will be followed by a review of possible gender-based differences related to (im)politeness and facework, zooming in on Polish. The existing, scarce research on the gender factor in conference interpreting will be presented. Afterwards, I will proceed with my own study, starting with a qualitative description of how the pragmatic meaning may be transferred by interpreters, and following with a quantitative analysis.

2.Impoliteness

The earliest politeness theory by Brown and Levinson (1987)Brown, Penelope, and Stephen C. Levinson 1987 Politeness. Some Universals in Language Usage. Cambridge: Cambridge University Press. assumed that politeness is the ‘default setting’ in human communication. Any utterances that endanger someone’s face (face-threatening acts, FTAs) are therefore typically accompanied by facework, i.e. mitigating politeness strategies such as apologies, hedges, impersonal constructions. However, social harmony is hardly a universal goal. The fact that impoliteness is not a marginal phenomenon was pointed out, among others, by Lakoff (1989)Lakoff, Robin 1989 “The Limits of Politeness: Therapeutic and Courtroom Discourse.” Multilingua 8: 101–129. and Culpeper (1996)Culpeper, Jonathan 1996 “Towards an Anatomy of Impoliteness.” Journal of Pragmatics 25: 349–367. . The latter’s model of linguistic impoliteness mirrors Brown and Levinson’s theory of politeness by presenting a number of strategies that are opposites of theirs. In Culpeper’s more recent works, his concept of impoliteness evolves to become more addressee-oriented. Consequently, the focus shifts from the speaker’s hostile intentions to the addressee’s failed expectations (Culpeper 2011Culpeper, Jonathan 2011 Impoliteness: Using Language to Cause Offence. Cambridge: Cambridge University Press. , 254).

Impoliteness may damage the addressee’s face or, alternatively, be perceived as infringing on his/her rights: the former will cause the addressee’s hurt or humiliation, the latter is more likely to result in anger (Culpeper 2013Culpeper, Jonathan 2013 “Impoliteness: Questions and Answers.” In Aspects of Linguistic Impoliteness, ed. by Denis Jamet, and Manuel Jobert, 2–15. Newcastle upon Tyne: Cambridge Scholars Publishing., 5–6). For researchers to be sure that a behaviour was actually impolite, the addressee’s emotional reaction should become manifest – by reciprocating the impoliteness, commenting on it or at least by some non-verbal symptoms (Culpeper 2016Culpeper, Jonathan 2016 “Impoliteness Strategies.” In Interdisciplinary Studies in Pragmatics, Culture and Society, ed. by Alessandro Capone, and Jacob L. Mey, 421–445. Berlin: Springer International. , 436). In his recent empirical studies, Culpeper uses the notion of “impoliteness events”, defined as “constellations of behaviours and co-textual/contextual features that co-occur in time and space, have particular functions and outcomes, and are/can be discussed and remembered by participants after the event” (2011Culpeper, Jonathan 2011 Impoliteness: Using Language to Cause Offence. Cambridge: Cambridge University Press. , 195).

Some researchers (e.g. Mills 2005Mills, Sara 2005 “Gender and Impoliteness.” Journal of Politeness Research 1 (2): 263–280. ) also point out that politeness and impoliteness should not be perceived as binary opposites, but rather as a continuum. Moreover, some believe that both politeness and impoliteness go beyond the participants’ expectations as to what is appropriate in a given situation and a third term is necessary to describe utterances and behaviours that are neutral in this respect. Lakoff (1975)Lakoff, Robin 1975 Language and Woman’s Place. New York: Harper and Row. and Leech (2014)Leech, Geoffrey 2014 The Pragmatics of Politeness. Oxford: Oxford University Press. opt for “nonpoliteness”, whereas Watts (2003)Watts, Richard J. 2003 Politeness. Cambridge: Cambridge University Press. prefers “politic behaviour”.

3.Parliamentary impoliteness

Due to its very nature (continuous competing for voters’ favour), parliamentary discourse is at times highly adversarial, as it contains numerous FTAs of varying gravity directed at speakers’ political opponents. Many utterances easily fit under the headings “seeking disagreement” and “denying association with others”, i.e. two of Culpeper’s (1996Culpeper, Jonathan 1996 “Towards an Anatomy of Impoliteness.” Journal of Pragmatics 25: 349–367. ) impoliteness strategies. On the other hand, in a democratic parliament divergent opinions and criticism are expressed as a matter of course. If face-threatening behaviour is fully expectable, should it be deemed impolite or, rather, politic? Watts claims that “competitive forms of interaction such as political debate” are likely to “sanction or neutralise face-threatening or face-damaging acts” (2003Watts, Richard J. 2003 Politeness. Cambridge: Cambridge University Press. , 131–132). Likewise, Culpeper explicitly mentions parliamentary debates as a setting in which impoliteness is “expected and sanctioned”, with the reservation that certain limits still apply (2011Culpeper, Jonathan 2011 Impoliteness: Using Language to Cause Offence. Cambridge: Cambridge University Press. , 172). As regards the British House of Commons, Harris points out that “[s]ystematic impoliteness, in the form of utterances which are intentionally designed to be face-threatening, is not only sanctioned but rewarded” (2001, 466).

In the light of the above, we may ask what qualifies as impolite in a parliamentary debate. Laskowska (2008)Laskowska, Elżbieta 2008 “Między językiem ideologii a językiem agresji [Between the language of ideology and the language of aggression].” In Reverendissimae Halinae Satkiewicz cum magna aestimatione, ed. by Grzegorz Dąbkowski, 185–193. Warsaw: Plejada. proposes to differentiate between ideological utterances that should not cause offence and aggressive utterances that target the opponent’s actions in their entirety and/or personal qualities beyond the addressee’s control (e.g. intellect, ethnicity, genetic make-up). Likewise, Ilie defines parliamentary insults as remarks that are “more challenging and more intense than reproaches, accusations and criticisms” (2001Ilie, Cornelia 2001 “Unparliamentary Language: Insults as Cognitive Forms of Confrontation.” In Language and Ideology, Vol. II. Descriptive Cognitive Approaches, ed. by René Dirven, Roslyn Frank and Cornelia Ilie, 235–263. Amsterdam: John Benjamins. , 260) and possess three characteristic features: (1) emotional force exceeding rational force, (2) reinforcement of stereotypical reasoning, and (3) inhibition of further dialogue as a result of seriously undermining the opponent’s image, position and authority. Plug, in turn, focuses on “personal attacks”, also referred to as “ad hominem arguments”, which are “directed not at the intrinsic merits of the opponent’s standpoint or doubt, but at the person himself or herself” with the primary goal of casting doubt on his/her “expertise, intelligence, character or good faith” (2010Plug, H. José 2010 “Ad Hominem Arguments in the Dutch and the European Parliaments. Strategic Manoeuvring in an Institutional Context.” In European parliaments under scrutiny: Discourse strategies and interaction practices, ed. by Cornela Ilie, 305–328. Amsterdam: John Benjamins. , 311). All three contributions offer many real-life examples illustrating what the authors consider to go beyond the threshold of politic behavior. Even more importantly, they enumerate the features that can serve as a starting point in the search for parliamentary impoliteness.

4.Gendered language and (im)politeness

‘Gendered language’ is a very complex issue comprising the idea that men and women may talk differently (i.e. in separate ‘genderlects’). In 1990s, the topic gained much exposure through popular science books that deal with communication problems between men and women (e.g. Tannen 1990Tannen, Deborah 1990 You Just Don’t Understand: Women and Men in Conversation. New York: Ballantine.; Gray 1992Gray, John 1992 Men Are from Mars, Women Are from Venus: A Practical Guide to Improving Communication and Getting What You Want. New York: Harper Collins.). Such problems are attributed to the ‘cooperative’ and ‘competitive’ conversational styles adopted by female and male speakers, respectively.

Considering more scientific approaches, differences in language use between male and female speakers have been primarily researched for English. The widespread interest in gendered language started with assumptions based mainly on introspection and anecdotal evidence made by Lakoff (1975)Lakoff, Robin 1975 Language and Woman’s Place. New York: Harper and Row.. According to her, women’s language displays many features of powerlessness, such as indirectness, question tags, hedges and euphemistic swearwords. These may often amount to politeness strategies. Lakoff’s paper has been exceptionally seminal and it triggered what Weatherall calls “the empirical avalanche” (2002Weatherall, Ann 2002 Gender, Language and Discourse. Hove: Routledge., 57).

Interestingly, the empirical research has yielded divergent results. For example, Holmes’ extensive research carried out in New Zealand (1995Holmes, Janet 1995 Women, Men and Politeness. London: Longman.) confirms many of the postulates made by Lakoff (1975)Lakoff, Robin 1975 Language and Woman’s Place. New York: Harper and Row.: in her data, inter alia, women were three times more likely to apologize than men, they also paid more compliments and used more facilitative question tags. On the other hand, Römer (2005)Römer, Ute 2005 “ ‘This seems somewhat counterintuitive, though…’ – Negative Evaluation in Linguistic Book Reviews by Male and Female Authors.” In Strategies in academic discourse, ed. by Elena Tognini Bonelli, and Gabriella Del Lungo Camiciotti, 97–115. Amsterdam: John Benjamins. found no significant differences (e.g. in the level of indirectness) in the ways female and male reviewers expressed negative evaluations of scientific papers. Similarly, Hampel (2015)Hampel, Elisabeth 2015 “ ‘Mama Zimbi, pls help me!’ – Gender Differences in (Im)politeness in Ghanaian English Advice-giving on Facebook.” Journal of Politeness Research 11 (1): 99–130. , who studied on-line advice offered by speakers of Ghanaian English, concludes that no clear gender differences are discernible. A detailed discussion is impossible here for reasons of space, however, the divergencies are not surprising given that studies such as the ones mentioned above have been conducted on a number of different languages and dialects. We should also realize that many studies suffer from methodological problems ranging from very small samples through insufficient attention paid to factors other than gender (e.g. age, status, situational context, native/non-native language use) to questionable data interpretation. Overall, (im)politeness does not lend itself easily to measurement and comparison.

More recently, as gender is seen primarily as a social construct, the perception of genderlects tends to change. They are less often discussed in terms of inherent features of men’s/women’s speech, but rather as stereotypes, assumptions and expectations that individuals may conform with or not in a particular interaction (see, e.g. Mills 2003Mills, Sara 2003 Gender and Politeness. Cambridge: Cambridge University Press. , 2005Mills, Sara 2005 “Gender and Impoliteness.” Journal of Politeness Research 1 (2): 263–280. ). Additionally, many studies (e.g. Ogiermann 2008Ogiermann, Eva 2008 “On the Culture-specificity of Linguistic Gender Differences: The Case of English and Russian Apologies.” Intercultural Pragmatics 5 (3): 259–286. ; Saeli 2016Saeli, Hooman 2016 “Persian Favor Asking in Formal and Informal Academic Contexts: the Impact of Gender and Academic Status.” Pragmatics 26 (2): 315–344. ) emphasize the culture-specificity of linguistic gender differences. Consequently, let us look at this issue in Polish, the target language11.The target language is the language into which the text is translated or interpreted, whereas the source language is the original one. The terms ‘target/source text’ are used by analogy. in this study.

Polish linguists engage with this topic much more rarely than Anglo-Saxon ones, and empirical research is very scarce. This does not prevent some scholars, however, from formulating strong opinions. Already in 1987Wilkoń, Aleksander 1987 Typologia odmian językowych współczesnej polszczyzny [Typology of linguistic varieties of modern Polish]. Katowice: Uniwersytet Śląski., Wilkoń wrote about the language of Polish women as a separate “biolect”, characterized (in decisively positive terms) by a stronger adherence to linguistic norms and traditions and also by more politeness and less aggressiveness (1987Wilkoń, Aleksander 1987 Typologia odmian językowych współczesnej polszczyzny [Typology of linguistic varieties of modern Polish]. Katowice: Uniwersytet Śląski., 103–106). As argued by Marcjanik (2007)Marcjanik, Małgorzata 2007 Grzeczność w komunikacji językowej [Politeness in linguistic communication]. Warsaw: PWN., the traditional Polish politeness model gives certain privileges to women, but in return women are expected to live up to higher politeness standards, i.e. show more “sensitivity, empathy and gentleness” (2007Marcjanik, Małgorzata 2007 Grzeczność w komunikacji językowej [Politeness in linguistic communication]. Warsaw: PWN., 29) as regards both their non-linguistic and linguistic behaviours.

Kwiryna Handke and Marta Dąbrowska are the two Polish scholars who most consistently return to genderlects in their work. The former offers highly impressionistic insights somewhat similar to those by Lakoff (1975)Lakoff, Robin 1975 Language and Woman’s Place. New York: Harper and Row., whereas the latter engages in sound empirical research using both English and Polish material, albeit relatively small samples. Handke’s contribution can be exemplified well by her description of the “feminine style” (2006Handke, Kwiryna 2006 “The Language of Polish Women.” Studia z Filologii Polskiej i Słowiańskiej 41: 83–94.). This style features high emotiveness and tendency to evaluate, both in positive and in negative terms. Therefore, women’s speech abounds in diminutives, augmentatives, exclamatives and neologisms of various types, accompanied by high redundancy resulting from the use of numerous ‘empty’ adverbs, adjectives and pronouns. The description is provided without reference to any empirical research. Although politeness is not mentioned explicitly, we may conclude that Polish women are prone to engage in more extreme pragmatic behaviours, i.e. both politeness and impoliteness, whereas men’s speech remains within the scope of nonpoliteness most of the time.

Dąbrowska explores intercultural differences between (supposed) genderlects of English and Polish, concluding that cultural differences frequently override gender differences, and the genderlects of Poles seem less divergent than those of Britons (2007Dąbrowska, Marta 2007 “Are Genderlects Universal?” Studia Linguistica Universitatis Iagellonicae Cracoviensis 124: 49–50.). Rather disappointingly for us, however, her studies that pay more attention to politeness phenomena (e.g. 2012Dąbrowska, Marta 2012 “Gender and Positive Politeness in Facebook Communication.” Armenian Folia Anglistika 1–2 (10): 7–20.) do not directly compare Polish men and women.

Interesting findings are presented by Karwatowska and Szpyra-Kozłowska (2010)Karwatowska, Małgorzata, and Jolanta Szpyra-Kozłowska 2010 Lingwistyka płci. Ona i on w języku polskim [Gender linguistics. She and he in Polish]. Lublin: Wydawnictwo UMSC., who investigated 140 Polish teenagers (aged about 16–19) by means of surveys with open questions. The respondents were supposed to produce different types of speech acts (greetings, farewells, thanks, apologies, compliments, criticisms and consolations), and they were also asked two additional questions: ‘Is politeness important?’ and ‘Why do people try to be polite?’. The production tasks revealed more similarities than differences between boys and girls, with a general strong tendency to undermine traditional, ‘adult’ politeness norms. However, boys were more likely to use vulgarisms and humour, whereas girls tended to employ more diminutives, questions and exclamatives. The direct questions about attitudes to politeness revealed that both girls and boys valued it, but to different degrees. Girls put stronger emphasis on their assertions to the effect that politeness is important and needed, whereas some boys questioned the necessity or sincerity of polite behaviours, mentioned exceptions or the acceptability of retaliating impoliteness received from someone else.

Overall, opinions based on anecdotal evidence and linguists’ intuitions suggest that Polish women should be more polite than their male compatriots. The existing empirical research on Polish genderlects, however, is clearly insufficient to either support or reject such assertions.

5.Gender factor in interpreting

Empirical research on gender-dependent politeness behaviour by interpreters is also scarce, although some authors (Mason 2008Mason, Marianne 2008 Courtroom Interpreting. Lanham: University Press of America.; Nakane 2008Nakane, Ikuko 2008 “Politeness and Gender in Interpreted Police Interviews.” Monash University Linguistics Papers 6 (1): 29–40.) claim that women do add more politeness markers in community interpreting. Among quite numerous studies that evidence interpreters’ facework mitigating original speakers’ FTAs across various interpreting modes and settings (see, e.g. Wadensjö 1998Wadensjö, Cecilia 1998 Interpreting as Interaction. London and New York: Longman.; Mason and Stewart 2001Mason, Ian, and Miranda Stewart 2001 “Interactional Pragmatics, Face and the Dialogue Interpreter.” In Triadic Exchanges. Studies in Dialogue Interpreting, ed. by Ian Mason, 51–70. Manchester: St Jerome.; Knapp-Potthoff 2005Knapp-Potthoff, Annelie 2005 “Secondhand Politeness.” In Politeness in Language: Studies in its History, Theory, and Practice, ed. by Richard J. Watts, Sachiko Ide, Konrad Ehlich, 203–218. Berlin: Mouton de Gruyter.; Monacelli 2006Monacelli, Claudia 2006 “Implications of Translational Shifts in Interpreter-mediated Texts.” Pragmatics 16 (4): 457–473. and 2009Monacelli, Claudia 2009 Self-preservation in Simultaneous Interpreting: Surviving the Role. Amsterdam: John Benjamins. ; Pöllabauer 2007Pöllabauer, Sonja 2007 “Interpreting in Asylum Hearings: Issues of Saving Face.” In The Critical Link 4. Professionalisation of Interpreting in the Community, ed. by Cecilia Wadensjö, Birgitta Englund Dimitrova, and Anna-Lena Nilsson, 39–52. Amsterdam: John Benjamins.; Duflou 2012Duflou, Veerle 2012 “The ‘first person norm’ in Conference Interpreting (CI) – Some Reflections on Findings from the Field.” In Interpreting Brian Harris: Recent Developments in Translatology, ed. by María Amparo Jimenez Ivars, and María Jesús Blasco Mayor, 145–160. Frankfurt a. M.: Peter Lang.), only a few take the gender factor into consideration.

The experiment carried out by Łyda et al. (2010Łyda, Andrzej, Alina Jackiewicz, and Krystyna Warchał 2010 “Agentless Structures in the Interpreter’s Output: Looking into the Gender Factor.” Linguistica Silesiana 31: 193–208., 2011Łyda, Andrzej, Krystyna Warchał, and Alina Jackiewicz 2011 “Managing Criticism and Praise by Trainee Interpreters: Looking for Gender Differences.” In Individual Learner Differences in SLA, ed. by Janusz Arabski and Adam Wojtaszek, 161–183. Clevedon: Multilingual Matters. ) enables the authors to compare 25 interpretations by women and 25 interpretations by men, provided by advanced interpreting trainees working in the consecutive mode from English into Polish. The study focuses on both criticism and praise directed at the audience. In set one, the interpreters belong to the group that is either praised or blamed by the speaker (students of English at the University of Silesia). In set two, the speaker is evaluating the Australian/British Olympic Team and the students are supposed to interpret for the Polish Olympic Team, so they are not among the group that is praised/criticized, furthermore, the evaluation as such is not directed at the target language audience. Although the source texts were in fact carefully designed for the needs of the study, they were camouflaged as authentic speeches. The interpretations are analyzed in terms of pre-determined ‘focal points’, i.e. fragments where criticism or praise is expressed very directly. The detected mitigation strategies include use of passive and impersonal constructions, transformation of verbal clauses into nominal ones and change of deictic perspective.

Łyda et al. 2011Łyda, Andrzej, Krystyna Warchał, and Alina Jackiewicz 2011 “Managing Criticism and Praise by Trainee Interpreters: Looking for Gender Differences.” In Individual Learner Differences in SLA, ed. by Janusz Arabski and Adam Wojtaszek, 161–183. Clevedon: Multilingual Matters. reveal marked differences in how male and female interpreters handled the deictic perspective of the texts. Overall, three types of shifts were detected: I to WE, YOU to WE and YOU to THEY, and women produced 2.6 times more deictic shifts than men. Moreover, male and female approach to praise vs. criticism was quite different: male interpreters employed deictic shifts mainly for praise, whereas female ones were more likely to use them for negative evaluation. In particular, female interpreters preferred YOU to THEY shifts, which they produced 7 times more often than their male colleagues (with the effect of deflating both criticism and praise). Both male and female participants were much more likely (4.2 times and 6 times, respectively) to reduce directness in case of the texts addressed at the audience they presumably identified with. The other paper by the same team (Łyda et al. 2010Łyda, Andrzej, Alina Jackiewicz, and Krystyna Warchał 2010 “Agentless Structures in the Interpreter’s Output: Looking into the Gender Factor.” Linguistica Silesiana 31: 193–208.) focuses on agentless structures and nominalizations. While no significant gender difference was detected for the former, female interpreters were about 3 times more likely to employ the latter when transferring the texts involving ‘in-group loyalty’, both with negative and positive evaluations. Although the scope of the study is limited, the above findings may suggest that Polish female interpreters are more sensitive to face issues and more prone to engage in facework, especially when they identify with the target audience. Among all the research discussed so far, this study gives the strongest support to the hypothesis that female interpreters from the EP Polish Language Unit might be more prone to mitigate FTAs than their male peers. However, note that the degree of the interpreter’s identification with the target audience at the EP is unknown.

On the other hand, a recent study that seems to bear the most similarity to this one (Magnifico and Defrancq 2016Magnifico, Cédric and Bart Defrancq 2016 “Impoliteness in Interpreting: A Question of Gender?” Translation & Interpreting 8 (2): 26–45.) as it also analyzes authentic interpretations from the EP in terms of FTAs (albeit in different language combinations, French-Dutch and French-English) brings results that clearly undermine this hypothesis. In their relatively small corpus, the authors have identified 92 FTAs22.Given the limited number of examples, this is rather difficult to assess, but it seems that the FTAs in Magnifico and Defrancq’s data, even the unmitigated ones, would not qualify as impolite in the parliamentary context according to the criteria described in Section 2. in the source texts (among them about 40% mitigated and about 60% unmitigated ones) and they examine their representations in the target texts. When all FTAs are analyzed together, no significant differences between the genders are discernible (i.e. both male and female interpreters omitted a similar share of FTAs). However, when the distinction between mitigated and unmitigated FTAs is taken into consideration, the proportion of unmitigated FTAs in the interpreters’ output (to the ones present in the source texts they rendered) is considerably lower for male interpreters. Applying my terminology (explained in 6.2.1) to the types of interpreting solutions present in this corpus, the women preserved the pragmatic effect of unmitigated FTAs in 68.4% cases, opted for attenuation in 25% of cases, and only sporadically resorted to elimination (2.6%) or strengthening (3.9%). The men, on the other hand, clearly preferred attenuation (in 55.9% of cases), followed by preservation of the pragmatic effect (26.5%) and elimination (17.7%). Consequently, the study shows that the male interpreters carried out more facework to tone down FTAs produced by French speakers.

6.The study

6.1Methodology

The main aim when compiling the observational corpus was to obtain possibly many grave FTAs in a manageable (both to transcribe and to analyze manually) amount of text. As the United Kingdom Independence Party (UKIP) is strongly Eurosceptic and its MEPs routinely undertake verbal attacks against prominent EU officials, I decided to focus on plenary contributions of this very party. Nigel Farage, Godfrey Bloom and John Bufton were selected as the participants, and the corpus covers all the contributions they delivered from the moment EP sessions start being available on-line (September 2008) until the end of 2012. It consists of 218 original speeches (5 h 13 min of talk, approx. 52,700 words) and the corresponding Polish interpretations. The material was downloaded from the EP website as MP4 video files (with separate audio tracks for all the official EU languages) and official verbatim reports of the original contributions. The interpretations had to be transcribed from scratch. Both the English and the Polish transcripts were verified and corrected against delivery by the researcher.

Having no access to voice recognition software, I have not been able to determine how many interpreters were recorded. Relying on an imperfect judgment ‘by ear’, there are certainly well over 20 individuals. Considering the time scope and the multitude of sessions, the corpus probably contains samples from almost every Polish interpreter who regularly worked at plenary sessions at the time. However, as at the stage of corpus compilation I was not aiming to investigate the gender factor, the share of input interpreted by men and by women is not equal. Rather, it is likely to reflect, roughly, the overall shares of male and female interpreters’ output in the Polish booth.

First of all, parts of the corpus underwent detailed qualitative analysis (in accordance with the principles of Discourse Analysis as outlined by Hale and Napier 2013Hale, Sandra and Jemina Napier 2013 Research Methods in Interpreting. A Practical Resource. London: Bloomsbury., 119), conducted bottom-up, to detect some recurring patterns in how FTAs were handled by interpreters. The very complex picture that emerged (see Bartłomiejczyk 2016Bartłomiejczyk, Magdalena 2016 Face Threats in Interpreting: A Pragmatic Study of Plenary Debates in the European Parliament. Katowice: Wydawnictwo Uniwersytetu Śląskiego., 209–211) made it clear that any quantitative analysis of the entire corpus would have to consider entities larger than single FTAs (this is where Culpeper’s 2011Culpeper, Jonathan 2011 Impoliteness: Using Language to Cause Offence. Cambridge: Cambridge University Press. concept of impoliteness events emerged as a handy solution). At the same time, it needs to be pointed out that pragmatic phenomena are difficult to research quantitatively. Source-text fragments that are deemed impolite will still display considerable variety as to the intended and perceived level of face-threat. Furthermore, the interpreters’ facework, even if going in the same direction, may modify the original to different degrees in particular cases. However, quantitative analysis, even involving oversimplifications, is the only tool that may allow us to go beyond individual interpreter’s idiolects to at least tentative generalizations.

6.2General results

6.2.1Qualitative analysis

The whole corpus was scrutinized to detect impoliteness events as defined by Culpeper (2011)Culpeper, Jonathan 2011 Impoliteness: Using Language to Cause Offence. Cambridge: Cambridge University Press. . Considering the specificity of parliamentary discourse and the resulting high threshold for impoliteness, they were supposed to meet the criteria enumerated by Laskowska (2008)Laskowska, Elżbieta 2008 “Między językiem ideologii a językiem agresji [Between the language of ideology and the language of aggression].” In Reverendissimae Halinae Satkiewicz cum magna aestimatione, ed. by Grzegorz Dąbkowski, 185–193. Warsaw: Plejada., Ilie (2001)Ilie, Cornelia 2001 “Unparliamentary Language: Insults as Cognitive Forms of Confrontation.” In Language and Ideology, Vol. II. Descriptive Cognitive Approaches, ed. by René Dirven, Roslyn Frank and Cornelia Ilie, 235–263. Amsterdam: John Benjamins. and Plug (2010)Plug, H. José 2010 “Ad Hominem Arguments in the Dutch and the European Parliaments. Strategic Manoeuvring in an Institutional Context.” In European parliaments under scrutiny: Discourse strategies and interaction practices, ed. by Cornela Ilie, 305–328. Amsterdam: John Benjamins. as described in Section 3. Admittedly, the judgment may, to some extent, still be subjective, i.e. dependent on the analyst’s sensitivity. The audience’s reactions (heckling by opponents, applause and/or laughter by supporters, subsequent admonishment by the chairperson or other speakers) were often checked in case of doubt.

A typical impoliteness event contains several identifiable FTAs (mostly to the addressee’s positive face), sometimes accompanied by politeness strategies that may be construed either as superficial adherence to the norms of parliamentary discourse or as mock politeness. An impoliteness event has a specific target, most typically an individual or many individuals (e.g. the whole audience at the EP, possibly excluding the speaker’s political supporters). It may also target an institution (e.g. the European Commission), country, or political party with which individuals are likely to identify (in line with Culpeper’s broader understanding of face as “not confined to the immediate properties of the self” (1996Culpeper, Jonathan 1996 “Towards an Anatomy of Impoliteness.” Journal of Pragmatics 25: 349–367. , 361)). The targeted persons do not have to be addressed directly (they are often mentioned in the third person) and do not even have to be present (as reachable through the broadcast).

A comparative analysis was carried out for impoliteness events detected in the source texts and the corresponding fragments of the Polish interpretations. In general, it revealed three recursive patterns, to which I refer as facework superstrategies: preservation of the pragmatic effect, mitigation and aggravation. The first one results in a target language version assessed as having the same pragmatic effect as the equivalent original event. If we ascribed to the conduit model of interpreting (see, e.g. Pöchhacker 2004Pöchhacker, Franz 2004 Introducing Interpreting Studies. London and New York: Routledge. , 194) assuming that the interpreter’s task consists in fully adopting the role of the original speaker, this would be the only acceptable option. Mitigation and aggravation modify the face-threat level present in the original impoliteness events. As these terms lack precision, I further propose to divide mitigation into elimination (where the target text is no longer face-threatening) and attenuation (where some face threat, although milder, is still present in the target text). By analogy, aggravation should be divided into strengthening and creation (although the latter was not detected throughout the corpus, i.e. there are no impoliteness events introduced by the interpreters). At this level, I refer to facework strategies, which are not equivalent to “face-saving strategies” or “mitigation strategies” as used elsewhere in the literature – the term has a broader scope here.

Moving further down, particular facework strategies are achieved by means of interpreting strategies33. Gile (2009)Gile, Daniel 2009 Basic Concepts and Models for Interpreter and Translator Training. Revised Edition. Amsterdam: John Benjamins. refers to them as “coping tactics”. (such as omission, deictic shift or literal translation). However, interpreting strategies do not inherently correspond to certain facework strategies: even the same interpreting strategy may produce divergent effects on facework, depending on the pragmatic value of the element to which it is applied. For instance, addition of a hedge will tone an FTA down, whereas addition of a booster will aggravate it.

Figure 1.Model of interlingual transfer of impoliteness

Before I proceed to illustrate the model presented here with a few examples, an important caveat is necessary. With no access to the interpreter’s strategic processing, interpreting strategies as intentional moves may only be hypothesized. As a matter of fact, an analyst relying on product-oriented methods is not in a position to determine with any degree of certainty which interpreting strategies are used for facework management and which are used to alleviate cognitive strain. Sometimes both the considerations might even conspire in the interpreter’s mind to produce the final effect of reduced impoliteness. This is perhaps best exemplified by omission. From the interpreter’s perspective, the use of omission may easily stem from other considerations than reducing danger to face. S/he may evaluate certain elements as less relevant and therefore dispensable in view of the high delivery rate and/or information density, or simply fail to hear or to understand a part of the original. Napier and Baker (2004)Napier, Jemina, and Roz Baker 2004 “Sign Language Interpreting: The Relationship between Metalinguistic Awareness and the Production of Interpreting Omissions.” Sign Language Studies 4 (4): 369–393. actually distinguish five types of interpreting omissions. Only 26% of the omissions detected in their experimental material are classified as “conscious strategic”, i.e. ones meant to “enhance the effectiveness of the interpretation” and based on the interpreter’s linguistic and cultural knowledge. Other omissions range from “conscious intentional” resulting from comprehension problems or failure to find an appropriate target language equivalent, through “conscious unintentional” resulting from short-term memory overload and the need to catch up with the speaker, “conscious receptive” caused by problems with transmission of the input, to “unconscious omissions” of information the interpreter did not even notice (2004Napier, Jemina, and Roz Baker 2004 “Sign Language Interpreting: The Relationship between Metalinguistic Awareness and the Production of Interpreting Omissions.” Sign Language Studies 4 (4): 369–393. , 377–378).

Table 1 presents several impoliteness events from the corpus as delivered in English and interpreted into Polish, the author’s back-translations of the Polish interpretations, facework strategies used to transfer each impoliteness event, and interpreting strategies serving to achieve the final effect.

Table 1.Selected impoliteness events from the UKIP corpus

No.	Source text	Target text	Back-translation	Facework strategy	Interpreting strategies
1.	[…] I happily am not a Conservative, therefore I do not have to vote blindly for complete nonsense.	Ja na szczęście nie jestem konserwatystą, więc nie muszę ślepo głosować za kompletnym nonsensem.	I happily am not a Conservative, therefore I do not have to blindly vote for complete nonsense.	preservation of the pragmatic effect	literal translation
2.	[…] it beggars belief that you and our President, Mr Buzek, can talk about the Solidarity movement, can talk about Poland getting its democracy back twenty years ago and yet here you are, surrendering the democracy and sovereignty of Poland to a failed European Union.	I pan jak i przewodniczący Buzek tutaj mówią o mówią o mówią o Solidarności, o Polsce, która odzyskała demokrację dwadzieścia lat temu. No i proszę, jednocześnie poddają się panowie i poddają pan panowie suwerenność Polski tej Unii Europejskiej, która przecież niesie niepowodzenie.	And you as well as President Buzek talk here about Solidarity, about Poland, which regained its democracy twenty years ago. And here you are, at the same time you are surrendering yourselves, and you are surrendering the sovereignty of Poland to this European Union, which, after all, carries failure.	preservation of the pragmatic effect	substitution, omission, addition
3.	You can smile, Mr Schulz, but you know nothing of financial markets or how these things work.	–	–	elimination	omission
4.	I don’t trust this place, which gives a veneer of democracy, which is largely made up of placemen.	Ja wierzę w prawdziwą demokrację.	I believe in real democracy.	elimination	parallel reformulation? /inferencing?
5.	[…] I saw for the first time even your own supporters shaking their heads. They don’t believe in what you’re saying. The European people don’t believe in what you’re saying, and I don’t really think even you now believe in what you’re saying.	Nawet miałem wrażenie, że właściwie nawet pańsk pańska pańscy zwolennicy niespecjalnie wierzyli w to co pan mówił. Nikt właściwie nie wierzył w to, co pan mówił. Nawet nie wiem, czy pan sam w to wierzy.	I even had the impression that actually even your your supporters did not particularly believe in what you were saying. Actually nobody believed in what you were saying. I even don’t know if you yourself believe it.	attenuation	addition
6.	Your henchman Olli Rehn, who is here today, I mean he dares to tell countries when they should and should not have general elections.	Dzisiaj jest z nami Olli Rehn, który śmie mówić krajom, czy powinny czy nie powinny organizować powszechne wybory.	Today with us is Olli Rehn, who dares to tell countries if they should or should not organize general elections.	attenuation	omission, deictic shift (here – z nami ‘with us’)
7.	Now they are gonna be, it would appear, subsumed by some sort of EU overseer, consisting no doubt of ignorant bureaucrats, Scandinavian housewives, Bulgarian mafia and Romanian peg-makers. You know, frankly, I think you’re gonna get on really well with each other.	I teraz wydaje się, że jakiś regulator czy organ nadzoru europejski ma przejąć jego zadania i tak naprawdę to będą tylko półgłówki, gospodynie domowe i nie wiem, kto jeszcze. I naprawdę myślę, że doskonale w swoim gronie się będą mieli.	And now it appears that a regulator or a European overseer is to take over its tasks and, as a matter of fact, these will only be halfwits, housewives and I don’t know who else. And I really think that they will get on fantastically among themselves.	attenuation	omission, generalization, deictic shift (YOU – THEY)
8.	Her appointment is an embarrassment for Britain.	To na pewno wielki wstyd dla Wielkiej Brytanii, że ją nominowano.	It is surely a huge embarrassment for Great Britain that she got appointed.	strengthening	addition, syntactic shift (noun – verbal phrase)

Examples 1 and 2 illustrate preservation of the pragmatic effect. It would seem that literal translation is the most suitable strategy (as shown by Example 1), but this is not always the case. Literalness rarely works for longer fragments of texts, and often produces results that are not pragmatically equivalent. For instance, if, in Example 2, Mr Buzek were translated literally, this would produce a face-threatening version in Polish, because, according to the Polish rules of politeness, high-ranking officials should be addressed by their positions. Therefore, przewodniczący Buzek ‘President Buzek’ is politic, whereas pan Buzek ‘Mr Buzek’ would not be – the substitution of the honorific was fully justifiable. Example 2 is a complex one where, I would argue, mitigation and aggravation cancel each other out. The mitigation is achieved through omission of the face-threatening phrase it beggars belief as well as the noun democracy as one of the values the addressees are accused of surrendering. The aggravation, in turn, results from the omission of our as a solidarity-creating pronoun and the addition of poddają się panowie ‘you surrender yourselves’ and the booster przecież ‘after all’.

Examples 3 and 4 illustrate elimination. The dominant interpreting strategy accounting for elimination is omission (Example 3), which, as mentioned before, is impossible to classify as strategic or otherwise without any access to the interpreter’s mental processes. In Example 4, however, I strongly suspect that some comprehension problems may be at play. What we see here is probably either a failed attempt at inferencing (reconstructing the rest on the basis of the only element the interpreter has grasped, i.e. democracy) or parallel reformulation (trying to contribute something in line with the speaker’s general attitude while remaining possibly noncommittal) – more details on these strategies are provided by Gile (2009Gile, Daniel 2009 Basic Concepts and Models for Interpreter and Translator Training. Revised Edition. Amsterdam: John Benjamins. , 201, 211).

Examples 5–7 account for attenuation. Example 5 is the most straightforward, as the final effect is achieved through multiple application of the same interpreting strategy, i.e. addition of a hedge. It is relatively rare for one single strategy to be employed for attenuation, usually interpreters opt for a combination of a few various strategies used on the same impoliteness event (as shown by Examples 6 and 7). Omission also plays a prominent role for attenuation, but it is applied more locally – see how the phrase your henchman is seamlessly deleted in Example 6. Likewise, some items from the highly offensive list of potential members of a new EU body in Example 7 are omitted (the interpreter takes special care not to mention their nationalities, which strongly suggests strategic omission) and replaced by the very general phrase i nie wiem, kto jeszcze ‘and I don’t know who else’. Example 7 also features a typical deictic shift from a more direct plural YOU to THEY, which somewhat dilutes the criticism.

Example 8 shows strengthening, which is relatively rare. Addition is one of the major interpreting strategies employed for strengthening. What is added here is an epistemic state upgrader and an intensifying adjective modifying the negatively charged noun embarrassment. A more subtle pragmatic shift consists in changing the syntax by replacing the noun phrase her appointment with a verbal construction, which results in more agency being attributed to the decision-makers.

6.2.2Quantitative analysis

Throughout the whole corpus, 293 impoliteness events have been revealed, amounting to the average of 1.34 per speech and one event per every 64 seconds of plenary talk. However, the distribution across speakers and contributions is very uneven, as 42.2% (92) of speeches in the corpus, although clearly face-threatening, do not contain any impoliteness events. 126 ‘impolite’ speeches typically contain from 1 to 5 impoliteness events, with the record score of 11 for a fairly long speech by Nigel Farage.

For 22.87% of impoliteness events in my corpus, the pragmatic effect was preserved by the Polish interpreters. Mitigation becomes manifest in elimination accounting for 10.24% and in attenuation – for 62.46% of the interpreting solutions in the corpus. Strengthening accounts for only 4.44% of the interpreting solutions, and no impoliteness events were created by the interpreters. Therefore, impoliteness events decisively tend to be mitigated in the Polish interpretations (in 72.7% of cases), but they are about six times more likely to get attenuated than eliminated completely.

Furthermore, I have proposed a simple (although, admittedly, not very fine-grained) scoring system for facework strategies. Preservation of the pragmatic effect scores 0 as it does not change the baseline level of impoliteness, attenuation scores −1, elimination −2, and strengthening +1. Using the suggested system, an impoliteness event in the corpus, on average, scores −0.78. This system also allows me to consider impoliteness at the level of a single impolite contribution,44.But note that the number of impoliteness events in the original speech is a variable that may strongly influence the facework score of its interpretation. and the average facework score for a Polish interpretation amounts to −1.83. Only one interpretation has a positive score of +1 (which suggests that strengthening may be employed mainly as a compensation for face-threat that was mitigated elsewhere in the same contribution), 14 score 0, and as many as 111 (88.1%) have a negative score, i.e. their overall impoliteness level appears lower than in the original due to interpreters’ mitigating moves. However, the negative facework scores tend to be rather low (−1 for 54 interpretations) and fall within the range of −8 (with scores of −6, −7 and −8 for one interpretation each). Interestingly, the interpreters’ attitude towards impoliteness, if considered for the whole Polish booth (seen as a community of practice that may have developed certain common habits and solutions to typical problems) seems rather inconsistent, as the highest negative facework scores were not achieved, as could perhaps be expected, by the speeches with the highest offensive potential (i.e. the ones that abounded the most in impoliteness events). The speeches that featured the greatest degree of interpreter mitigation, i.e. the ones that scored −8, −7 and −6,55.Each of the three interpretations was provided by a different individual. contained 4, 6 and 6 impoliteness events, respectively. The most impolite speech, in turn, that contained 11 impoliteness events, scored only −4. This may lead to the hypothesis that individual interpreting styles may play an important role here, and whether these interpreting styles correlate with the sex of the interpreter remains to be seen.

6.3Results considering the interpreters’ gender

As the contributions in the corpus are short and each is handled by a single interpreter66.Except for one speech during which another interpreter takes over from his colleague. However, both the interpreters are male, so the speech as a whole can be assigned to the same subcorpus. whose sex can be determined on the basis of the voice in the recording, the above results give us a good basis for a comparison between the male and the female interpreters. If, indeed, women tend to mitigate impoliteness to a greater degree than men, the distribution of facework strategies for each sex should differ, showing the preference of women for elimination and attenuation and the preference of men for preservation of the pragmatic effect. Consequently, female interpreters should achieve considerably higher average negative facework scores both at the level of a single impoliteness event and a single impolite speech. Moreover, it could be hypothesized that the interpretations with the highest negative scores should be delivered predominantly by women, whereas interpretations by the men should remain closer to 0.

Out of the 126 speeches containing impoliteness events, 80 (63.5%) were interpreted by women and 46 (36.5%) were interpreted by men. The women were exposed to 197 (67.23%), and the men to 96 (32.77%) impoliteness events, which means that the speeches interpreted by the women were, on average, slightly more impolite. The overall proportion of female to male renditions of impoliteness is close to 2:1. When we consider the facework strategies employed by the male and the female interpreters to render the impoliteness events they were faced with, their distribution appears alike (see Figure 2).

Figure 2.Impoliteness events in the UKIP corpus as interpreted from English into Polish by the female and the male interpreters

The observed facework strategies considered in relation to the interpreter gender account for the chi-squared value of 6.0 (df = 3), which means that the difference between the men and the women is not statistically significant for p < 0.05.

If we ascribe numerical values to the facework strategies as explained in 6.2.2, the average score at the level of a single impoliteness event is −0.73 for the women and −0.89 for the men. At the level of a single speech, the average facework scores are −1.81 for the women and −1.87 for the men. The distribution of facework scores across individual speeches is shown in Table 2.

Table 2.Facework scores at the level of a whole speech achieved by the female and the male interpreters

	1	0	−1	−2	−3	−4	−5	−6	−7	−8	Total
Women	1	7	36	13	12	8	2	1	0	0	80
Men	0	7	18	10	4	4	1	0	1	1	46

T-test for independent samples shows that these score distributions cannot be regarded as significantly different (t = 0.2, p < 0.05).

7.Discussion and conclusions

The obtained results disprove the hypothesis that the Polish female interpreters in the EP carry out more facework related to impoliteness expressed by original speakers than their male peers. They also fail to corroborate the gender bias manifest in the male interpreters’ tendency to tone down face-threat revealed by Magnifico and Defrancq (2016)Magnifico, Cédric and Bart Defrancq 2016 “Impoliteness in Interpreting: A Question of Gender?” Translation & Interpreting 8 (2): 26–45.. One of possible reasons for this discrepancy between the studies carried out in the same setting and applying similar methodology is that in each case the authors analyzed different language combinations and varying levels of face threat. The results of both the studies combined, however, should probably suffice to reject the hypothesis that in the EP setting, it is mainly female interpreters who engage in facework to save the face of addressees from grave FTAs made by their political opponents.

Where both the studies clearly converge is the general conclusion that mitigation is frequent and it should not be attributed exclusively to cognitive overload. Why should interpreters feel inclined to mitigate face-threat present in the original utterances? I will try to briefly present three explanatory hypotheses that allow us to look at mitigation from different angles. Note that they do not rule each other out and that the list is not exhaustive, as other factors could also be considered as possible reasons for mitigation.

7.1Mitigation as a norm in conference interpreting

Magnifico and Defrancq link faithful transfer of impoliteness with adherence to governing interpreting norms: “Female interpreters render most unmitigated FTAs straightforwardly, which could be the result of a desire to prioritize the professional norms” (2016Magnifico, Cédric and Bart Defrancq 2016 “Impoliteness in Interpreting: A Question of Gender?” Translation & Interpreting 8 (2): 26–45., 42). I beg to differ on this point. The problem is that, in the context of conference interpreting, the existence of any norms that would prescribe faithful transfer of speakers’ impoliteness is very dubious, to say the least.

For community interpreting, and court interpreting in particular, the relevant norm is clear and consistent across various codes of conduct and handbooks. It can be neatly summarized as follows: “The interpreter does not act as a censor. It is the responsibility of the other parties to choose to put things in a particular way and, if they make unfortunate or inappropriate choices, it is they who must be held responsible for any consequences of communicative breakdown” (Gentile et al. 1996Gentile, Adolfo, Uldis Ozolins, and Mary Vasilakakos, M. 1996 Liason Interpreting – A Handbook. Melbourne: Melbourne University Press., 49). For conference interpreting, the situation is different. This is not to say that norms do not operate there, but, rather, that their nature is often more implicit. Generally, interpreters appear to rely on deeply internalized performance norms. This is confirmed by the surveys discussed by Zwischenberger, in which interpreters report that the main “reason for both satisfaction and dissatisfaction [with the job] was the fulfilling or non-fulfilling one’s own standards” (2017Zwischenberger, Cornelia 2017 “Professional Self-perception of the Social Role of Conference Interpreters. In The Changing Role of the Interpreter. Contextualising Norms, Ethics and Quality Standards, ed. by Marta Biagini, Michael S. Boyd, and Claudia Monacelli, 52–73. New Yok and London: Routledge. , 70). However, individuals’ standards cannot have emerged in a vacuum, they have taken shape in a social reality, during encounters with trainers, colleagues and clients who have voiced particular expectations. As rightly pointed out by Marzocchi, “norms are established over time and through a complex mechanism involving a variety of agents inside and outside the specific institution” (2017Marzocchi, Carlo 2017 “Norms Revisited”. In The Changing Role of the Interpreter. Contextualising Norms, Ethics and Quality Standards, ed. by Marta Biagini, Michael S. Boyd, and Claudia Monacelli, 219–227. New Yok and London: Routledge. , 221).

EU interpreters do not have any institution-specific code of professional ethics. Many of them may feel some allegiance to the code of AIIC, the International Association of Conference Interpreters, however, it is much too general to clarify the issue of speaker-generated face-threat. As rightly pointed out by Diriker, this code “foregrounds secrecy, confidentiality, collegiality and integrity” but “does not specify what constitutes an ethical interpreting performance” (2004Diriker, Ebru 2004 De-/re-contextualizing Conference Interpreting. Interpreters in the Ivory Tower? Amsterdam: John Benjamins. , 30).

According to Marzocchi, the lack of explicitly formulated codified norms is “a missing link, a grey area in the way the conference interpreting profession depicts itself” (2005Marzocchi, Carlo 2005 “On Norms and Ethics in the Discourse on Interpreting.” The Interpreter’s Newsletter 13: 87–107., 100). Furthermore, he believes that prescriptive literature written by the ‘founding fathers’ of conference interpreting has played a pivotal role in shaping the professional ethics. A famous handbook by Jean Herbert (1952/1956Herbert, Jean 1952 English Translation 1956. The Interpreter’s Handbook: How to Become a Conference Interpreter. Geneva: Librairie de l’Université.) contains the following recommendation on impoliteness: “Certain offensive phrases which may go further than the speaker intended or realised should preferably be attenuated. An interpreter who fails to do so does not fulfill his real mission” (1956Herbert, Jean 1952 English Translation 1956. The Interpreter’s Handbook: How to Become a Conference Interpreter. Geneva: Librairie de l’Université., 52). However, Herbert fails to account for the fact that the offensiveness may be fully intended by the speaker, which may well create the impression that any impoliteness is out of place in the elevated context of high-level international politics. According to Kahane (2007)Kahane, Eduardo 2007 “Interpreters in Conflict Zones: The Limits of Neutrality.”, http://​aiic​.net​/page​/2691​/interpreters​-in​-conflict​-zones​-the​-limits​-of​-neutrality​/lang​/1 (accessed January 29, 2016)., conference interpreters have adopted the “sublime role […] as facilitators of dialogue or even messengers of peace”. This attitude is hardly confined to the mid-20th century, although the categorical tone probably belongs to the past. In a manual by a renowned EU interpreter that enjoys great popularity nowadays we read that “there are occasions when an interpreter may tone down comments to take the sting out of a meeting: repeating tactless or rude comments may in some cases be in the interest neither of the speaker, nor of the addressee, nor of the proceedings in general” (Jones 1998Jones, Roderick 1998 Conference Interpreting Explained. Manchester: St. Jerome., 21). The same author also discusses examples of impoliteness he would recommend to mitigate and transfer as it is.

In conclusion, mitigation of impoliteness is, to a large extent, left to the interpreter’s judgment in any particular case. This allows the interpreter much leeway, but, on the other hand, may also result in doubts and confusion. It is easy to find an external justification in the existing professional discourse for the decision to tone down the speaker’s impoliteness.

7.2Mitigation as self-censorship

Censorship and self-censorship are issues that often emerge in Translation Studies in relation to literary translation, but, perhaps unjustly, hardly ever get mentioned in the context of interpreting. According to Santaemilia, translators often “censor themselves – either voluntarily or involuntarily – in order to produce rewritings that are ‘acceptable’ from both social and personal perspectives” (2008Santaemilia, José 2008 “The Translation of Sex-related Language: The Danger(s) of Self Censorship(s).” TTR 21 (2): 221–252. , 221–222). An objectionable source text may undergo “all the imaginable forms of elimination, distortion, downgrading, misadjustment, infidelity and so on” (ibid: 224), a description that could refer to many shifts performed by interpreters on extracts from the UKIP corpus discussed here. As rightly pointed out by Ben-Ari, self-censorship is “so deeply rooted a mechanism that it has become a term in psychology, meaning the agent in the unconscious that is responsible for censorship” (2010Ben-Ari, Nitsa 2010 “When Literary Censorship is not Strictly Enforced, Self-censorship Rushes in.” TTR 23 (2): 133–166. , 135). Tymoczko deplores “the tendency of translators to buy into dominant views and to stop themselves from textual production suggesting difference or dissent” (2007Tymoczko, Maria 2007 Enlarging Translation, Empowering Translators. Manchester: St. Jerome., 257). Although she admits that some self-censorship is unavoidable, she advocates self-reflexivity as the main tool to counteract it. By contrast to interpreters, translators do not work under severe time pressure and they are not so deeply immersed in the communicative process, therefore, they may be able to heed this advice.

Impoliteness may well be described as a taboo that requires some effort to break. Allan and Burridge note that “[b]y default we are polite, euphemistic, orthophemistic and inoffensive; and we censor our language use to eschew tabooed topics in pursuit of well-being for ourselves and for others” (2006Allan, Keith, and Kate Burridge 2006 Forbidden Words. Taboo and the Censoring of Language. Cambridge: Cambridge University Press. , 2). It may be quite difficult to reject this attitude, on the spot, while speaking on someone else’s behalf as interpreters do. Therefore, mitigation may stem from the interpreter’s desire (often unconscious) to bring the message closer to his/her own politeness standards as well as the standards that generally apply to parliamentary discourse. The fact that MEPs are sometimes formally punished by the President of the EP for excessive impoliteness (which happened both to Nigel Farage and Godfrey Bloom) is likely to reinforce the interpreters’ internal censorship mechanisms.

7.3Mitigation as the interpreter’s intervention

Unlike the previous hypothesis, this one highlights the interpreter’s involvement as an active agent, consciously making decisions to introduce some changes to the source text for the sake of his/her own agenda. It is typically discussed in the context of community interpreting, which often involves great power imbalances between the interlocutors that the interpreter may want to redress. Obviously, intervention goes blatantly against the notion of interpreter impartiality and the conduit model of interpreting. Still, there are some academics ready to endorse it, among them Mona Baker, who believes that the interpreter has the right to challenge the speaker’s ideas or at least distance him/herself from them depending on “what kind of ‘narrative’ a source text elaborates” and whether “you agree with what the speaker or text says”, which also includes withholding content that “would cause unnecessary hurt and offence” (Baker and Chesterman 2008Baker, Mona, and Andrew Chesterman 2008 “Ethics of Renarration. Mona Baker is Interviewed by Andrew Chesterman.” Cultus 1: 10–33., 15).

Katan (2011)Katan, David 2011 “Interpreting as Intervention: Norms, Beliefs and Strategies.” In Interpretazione e mediazone: Un’opposizione inconcilabile?, ed. by José Francisco Medina Montero, and Sarah Tripepi Winteringham, 33–66. Rome: ARACNE. comprehensively discusses the notion of interpreter intervention, describing several levels at which it may appear. In accordance with his classification, the shifts revealed here might be seen as either pragmatic or ideological, depending on whether we see the impoliteness present in the original as confined purely to the personal dimension or endow it with political meaning, as a strategy adopted by a minority group (Eurosceptics) to oppose the dominant ideology in the EP. In the former case, the interpreter would struggle to reduce face-threat to all the parties concerned (including him/herself) in the name of general good rapport. In the latter, the interpreter would side with the powerful, dominant discourse by undertaking interventions to save the face of individuals (and institutions) attacked by a less powerful speaker. Considered from this perspective, my findings might be seen as corroborating those of Beaton (2007)Beaton, Morven 2007 Intertextuality and Ideology in Interpreter-mediated Communication: The Case of the European Parliament. PhD diss., Heriot-Watt University., who repeatedly talks about EP interpreters strengthening “EU institutional hegemony”.

7.4Final conclusions

As gender-specific patterns in handling impoliteness in UKIP Eurosceptic discourse failed to be established, the next step to take might be to explore possible differences among the individual interpreters to see how consistent they are in their approaches to impoliteness. Furthermore, the above explanatory hypotheses could be verified by means of ethnographic methods such as interviews and focus groups. However, such research requires extensive cooperation from members of the relevant community of practice, i.e. EP interpreters. In particular, it would be interesting to gauge their perception of professional norms. This empirical study could provide a sufficient number of examples of various authentic mitigating and non-mitigating solutions to be presented to interpreters for assessment. Another question that emerges is to what extent the interpreter’s own private and professional face is at stake when s/he is required to transfer impoliteness.

As performance norms in conference interpreting are fuzzy, maybe what we need is an honest debate among the professional community (with the possible participation of interpreting researchers and pragmaticians) to work out commonly accepted standards of ethical performance as regards transfer of pragmatic meaning that would be known and clear both to interpreters and to users of their services? It is my hope that this paper may contribute to trigger such a discussion.