A corpus-based study of euphemising body parts in Arabic subtitles

The study examines how private body part-related taboo words (PBPRTW) from Hollywood films are rendered in Arabic subtitles, a context shaped by strong cultural, religious, and social norms and by Arabic diglossia. Euphemism enables discussion of sensitive topics indirectly and politely, and cultures vary in which taboos are acceptable and how they are expressed. Hollywood content, produced in a comparatively liberal context, poses particular challenges for subtitlers targeting Arabic-speaking audiences using Modern Standard Arabic (MSA), a formal register. The study aims to fill a research gap focused specifically on PBPRTW (rather than general profanity), exploring the extent of maintenance versus euphemisation, the strategies employed, and whether the function of taboo words is reflected in the subtitles. Research questions: (1) To what extent are instances of PBPRTW in Hollywood films maintained and euphemised when subtitled in Arabic? (2) What euphemism strategies are used by Arab subtitlers in dealing with PBPRTW? (3) Is the function of taboo words reflected in the way those words are subtitled?

Prior work on translating taboo words across languages has identified strategies such as literal translation, euphemism, neutralisation, mitigation, substitution, and omission, with omission and euphemism frequently dominant (e.g., Torres-Cuenca, 2016; Koponen, 2018; Khoshsaligheh et al., 2018; Abdelaal & Al Sarhani, 2021; Al-Zgoul & Al-Salman, 2022; Abu-Rayyash et al., 2023; Haider et al., 2023). Euphemism strategies have been theorized in monolingual contexts (Williams, 1975; Warren, 1992) and studied in Arabic spoken contexts (Rabab'ah & Al-Qarni, 2012; Al-Khasawneh, 2018), but less so in cross-cultural audiovisual translation using MSA. Existing subtitling studies often rely on small samples or single films/genres, lack specificity in classifying taboo categories and functions, and seldom analyze how functions affect strategy choice. The present study addresses these gaps with a large, balanced corpus focused specifically on PBPRTW, relating euphemism strategies to taboo functions within Arabic subtitling norms.

Design: Corpus-based, mixed methods (quantitative frequencies and percentages; qualitative analysis of subtitlers’ choices and motivations). Data: 75 Hollywood films (released 2000–2018) selected by six criteria: (1) major genres represented (action, adventure, comedy, drama, horror); (2) availability of Arabic subtitles via DVDs/streaming (Amazon Prime, Netflix, iTunes); (3) IMDB rating ≥ 5/10; (4) award-winning or nominated; (5) release date 2000–2018; (6) single entry per franchise (no sequels/prequels duplicates). Extraction and processing: English subtitles extracted to plain text with SubRip. Arabic subtitles required OCR; extraction was challenging due to segmentation and script-complexity issues (e.g., Adobe Acrobat OCR errors). English–Arabic subtitles were aligned, converted to spreadsheets, and uploaded to Sketch Engine for corpus analysis. Lemmatization and concordancing identified candidate PBPRTW lemmas and provided context to filter non-taboo homographs (e.g., “golf balls” removed). Classification: Taboo categories grounded in Allan & Burridge (2006). Functions of profanity based on Pinker (2007) and McEnery (2006), with four functions analyzed: descriptive, abusive, referential, idiomatic. Strategy framework combined cross-cultural/euphemism models (Williams, 1975; Warren, 1992; Davies, 2003; Farghal, 1995; Al-Adwan, 2015). A pilot on two films determined applicability and excluded monolingual-specific strategies (e.g., phonetic distortion). Inclusion threshold: Only PBPRTW occurring >150 total instances across the corpus were analyzed to enable manageable yet representative analysis (approx. ≥2 per film). Coding: For each occurrence, the English item, its function, the Arabic rendering, and the euphemism strategy were recorded. Inter-rater validation: two Arabic-speaking PhDs in translation studies randomly reviewed classifications for accuracy. Analysis: Quantitative counts of maintenance, neutralisation, toning down, omission, and strategy frequencies; qualitative “pair group” analysis (Toury, 2012) to identify translational patterns and rationale, including effects of audiovisual context (gestures, visuals) on subtitlers’ implicitness choices. Euphemism strategies examined: widening, metaphorical transfer, implication, metonymy, semantic misrepresentation, omission, and preservation.

Corpus totals: 1101 PBPRTW tokens meeting the frequency threshold. Lexeme distribution: ass 45% (N=492), dick 25% (N=274), asshole 18% (N=194), balls 13% (N=141). Functional distribution: descriptive 40% (N=437), abusive 30% (N=326), referential 16% (N=186), idiomatic 14% (N=152). Offensive load outcomes in Arabic subtitles: toned down 45% (N=495), neutralised 42% (N=463), omitted 8% (N=84), maintained 5% (approx.). Strategy frequencies (overall, N=1101): metaphorical transfer 383 (35%), widening 234 (21%), metonymy 215 (19.5%), omission 86 (8%), implication 82 (7%), preservation 58 (5%), semantic misrepresentation 43 (4%). Strategy–function associations: - Descriptive: widening most frequent, 199/437 (45%); frequent euphemisation of ass as مؤخرة and dick as عضو/جسد leading to generalisation and ambiguity; implication also used to link penis references to masculinity implicitly. - Abusive: metaphorical transfer dominant, 244/326 (≈75%); insults like dick/asshole rendered as culturally acceptable abuse terms (e.g., حقير، غبي، لعين) rather than body-part metaphors. - Referential: metonymy dominant, 132–133/186 (≈71–72%); ass referring to a person rendered with pronouns or neutral references (e.g., “Get his ass out” → “أخرجه”). - Idiomatic: metaphorical transfer dominant, 106/152 (≈70%); idioms with balls (courage) or ass (similes) rendered by sense (e.g., جرأة), losing taboo imagery and often humour. Preservation was rare (≈5%), observed mainly with balls → خصية (a formal medical term with limited synonyms). Overall, approximately 95% of PBPRTW were euphemised, neutralised, or omitted, and only about 5% were retained explicitly.

Findings demonstrate that Arabic subtitling of PBPRTW prioritises cultural acceptability and the formal MSA register over literal transfer, leading to extensive toning down, neutralisation, and selective omission. Diglossia (spoken vs. MSA) and differing cultural norms around profanity render literal body-part insults unidiomatic or unacceptable; thus subtitlers map taboo expressions to culturally resonant equivalents (e.g., generalised body terms, non-taboo insults, pronouns). Audiovisual context can support implicit strategies (e.g., gestures indicating size), allowing implication while avoiding explicit taboo. However, these choices often reduce the offensiveness load, dilute humour, and alter pragmatic force and characterisation. Clear patterns link functions to strategies: metonymy with referential uses, widening with descriptive uses, metaphorical transfer with idiomatic and abusive uses, and limited use of semantic misrepresentation and omission to manage severity. The heavy use of MSA further elevates formality, softening slang and intensities inherent in the originals, potentially affecting audience immersion and interpretation compared to source versions.

The study shows that in Arabic subtitles of 75 Hollywood films, only about 5% of PBPRTW were preserved explicitly; the remaining 95% were toned down, neutralised, or omitted via euphemism strategies. Function–strategy correlations were robust: metonymy with referential function (≈71.5%), metaphorical transfer with abusive (≈74.8%) and idiomatic (≈69.7%) functions, and widening with descriptive function (≈45.5%). While these practices meet cultural and register constraints, they can reduce humour and pragmatic intensity, affecting intended messaging and character portrayal. Practical implications include prioritising the retention of the function of profanity when equivalent offensiveness cannot be maintained, and raising subtitlers’ awareness of how profanity indexes character idiolects, sociolects, and emotional strength. Future work should explore strategies that better preserve functional load without violating norms, potentially leveraging audiovisual cues and creative, context-sensitive solutions.

Data were limited to official DVD/streaming Arabic subtitles; fansubbing was excluded. The focus was on subtitling (not dubbing) and on formal MSA, which may not reflect colloquial usage. Audience reception was not studied. Future research should include dubbing and fansubs, incorporate multimedia corpora, examine audience perception of euphemised content, and conduct diachronic analyses of offensive language in Arabic subtitling over time.