Text complexity and translation styles from the perspective of individuation: a case study of the English translations of *Pipa Xing*

Pipa Xing, a celebrated Tang dynasty narrative-lyric poem by Bai Juyi, has been widely translated into English by multiple translators who bring distinct perspectives and linguistic renderings. Drawing on Systemic Functional Linguistics (SFL), this study explores translation style via text complexity (lexicogrammatical complexity) and explains stylistic variation through the SFL concept of individuation, which links translators’ access to and mobilization of cultural-semiotic resources with their affiliations to target readerships. The study addresses two questions: (1) What styles and differences do the various translations exhibit in terms of text complexity? (2) From the perspective of individuation, how do translators, based on their identities, allocate meaning resources and align with target readerships, thereby influencing translation styles? The work situates grammar as meaning-making (lexicogrammar) and positions complexity at the text level as part of translation style, with translators’ subjectivity central in shaping re-instantiation of meaning from source to target texts.

The theoretical framework integrates SFL’s three-dimensional model of language (realization, instantiation, individuation) aligned with semogenesis (phylogenesis, logogenesis, ontogenesis). Realization concerns stratified linkage from semantics to lexicogrammar; translation is modeled as re-instantiation of meaning potential from source to target texts, while individuation accounts for how translators’ identities and affiliations shape interpretive and stylistic outcomes. Text complexity is treated as multidimensional, operationalized here via lexical density (LD: lexical items per ranking clause) and grammatical intricacy (GI: ranking clauses per sentence). Taxis (parataxis vs. hypotaxis) and logico-semantic relations (projection; expansion: elaboration, extension, enhancement) structure clause complexes and contribute to GI. Prior SFL work suggests spoken texts tend to higher GI, written texts to higher LD; narratives often feature enhancement (especially temporal). Individuation (allocation and affiliation) describes how meaning resources from the cultural reservoir are differentially available to individuals (master identities, communities) and how individuals align with readerships and communities, shaping stylistic choices in translation. The review situates translators’ subjectivity and sociocultural positioning (e.g., Venuti, Lefevere, Bourdieu) as key determinants of translational style.

Corpus: the Chinese source poem Pipa Xing and nine English translations by Xu (1994), Yang & Yang (1984), Lin (1960), Zhang & Wilson (1988), Bynner & Kiang (1929), Cranmer-Byng (1915), Watson (1984), Gaunt (1919), and Giles (1884). Translators were grouped by master identity (nationality): Chinese; British/American; Sino-Western collaborations. Tools: SysFan (for SFL-based annotation and computation) and SPSS 22.0 (for statistical testing). Measures: LD = number of lexical items / number of ranking clauses; GI = number of ranking clauses / number of sentences. Annotation: Clause boundaries marked with “||”, clause complex boundaries with “|||”. Taxis labeled numerically (parataxis: 1,2,...) and with Greek letters (hypotaxis: α, β...), and logico-semantic relations coded for projection (locution, idea) and expansion (elaboration, extension, enhancement). English LD computed excluding function words via Cook’s function word list (225 items). Chinese LD computed using a POS-based system distinguishing lexical vs. function words per Wang (2015). Procedure: (1) The ST was segmented into six rhetorical stages (S1–S6) aligned with narrative structure; all texts were aligned at text, stage, sentence, and clause levels. (2) SysFan provided counts for words, clauses, sentences, LD, GI, and distributions of taxis and logico-semantic relations. (3) Comparative analyses were performed at overall and stage levels; ANOVAs tested differences among texts (Shapiro–Wilk for normality; homogeneity tests; Welch’s ANOVA where required). (4) Qualitative exemplars examined hierarchical complexity (taxis layering) and narrativity (enhancement vs. extension/elaboration) to interpret stylistic profiles. (5) Findings were synthesized and interpreted through the individuation framework to relate translators’ identities and affiliations to stylistic outcomes.

- Lexical density (LD): Across the ten texts (ST plus nine translations), LD differences were not statistically significant (one-way ANOVA p = 0.159). Translations thus show broadly comparable information density at the lexical level. - Grammatical intricacy (GI): Differences among translations were statistically significant (Welch’s ANOVA p = 0.005). Higher-GI translations: T3 (Yang & Yang; GI ≈ 5.16), T8 (Watson; 3.71), T7 (Cranmer-Byng; 3.35), T9 (Gaunt; 3.16), T6 (Bynner & Kiang; 3.02). Lower-GI translations: T10 (Giles; 2.82), T5 (Zhang & Wilson; 2.77), T2 (Xu; 2.65), T4 (Lin; 2.40). - Taxis and hierarchy: In the ST, parataxis predominates (≈72.9%). Among higher-GI translations, parataxis generally predominates, especially in T3, where high GI stems mainly from many ranking clauses per sentence rather than deeper hypotactic layering. In contrast, some low-GI translations (notably T5 and T10) show higher proportions of hypotaxis than parataxis, indicating more hierarchical clause embedding despite fewer clauses per sentence overall. - Logico-semantic relations and narrativity: All texts favor extension and enhancement over elaboration, but emphases differ. T3, T4, T6, T8, and T9 emphasize extension, while the ST, T2, T5, T7, and T10 emphasize enhancement. Given enhancement’s association with narrative sequencing (e.g., temporal relations), T2, T5, T7, and T10 display stronger narrativity. - Textual style profiles: Based on GI and taxis, five translations (T3, T8, T7, T9, T6) exhibit spoken-text features (higher GI, paratactic dominance). T10 and T5 exhibit written-text features (lower GI, hypotactic prevalence). T2 and T4 occupy a mixed spoken–written profile (low GI with paratactic dominance). These stylistic clusters arise despite similar LD across translations. - Exemplars confirmed that similar LD values can mask structural differences: e.g., segments in T5 vs. T9 show matched LD but differing clause complex structures and hierarchical depth, driving GI variation.

The findings address the research questions by showing that stylistic divergence among translations of the same ST primarily resides in grammatical organization (GI, taxis, and logico-semantic configurations) rather than lexical density. Mapping these differences to spoken vs. written stylistic profiles and degrees of narrativity reveals consistent clusters. From an individuation perspective, translators’ master identities, community affiliations, and target readership alignments help explain why they mobilize distinct lexicogrammatical resources in re-instantiating the ST. Chinese translators Xu and Lin, oriented toward accessibility and cross-cultural communication (but with differing missions and readerships), produced low-GI renderings with spoken–written hybridity. British/American translators display varied trajectories: Giles’ sinological orientation aligns with a written, narratively rich style; Gaunt, Cranmer-Byng, and Watson target broader publics or students with more colloquial, spoken-like renderings and reduced narrative enhancement in some cases. Sino-Western collaborations reflect negotiated identities and readerships: Bynner & Kiang and Yang & Yang favor spoken styles and accessibility for general readers and poets, whereas Zhang & Wilson maintain a more written style and stronger narrativity for academic and comparative literature audiences in both cultures. Individuation clarifies how allocated cultural resources and affiliation choices shape translators’ lexicogrammatical decisions, producing observable stylistic diversity across TTs.

Using SFL-based text complexity metrics and individuation theory, this study shows that nine English translations of Pipa Xing have similar lexical density but differ significantly in grammatical intricacy, taxis hierarchy, and logico-semantic patterning, yielding distinct spoken, written, and hybrid stylistic profiles and varying narrativity. The individuation framework elucidates how translators’ sociocultural positioning and affiliations with target readerships guide the mobilization of lexicogrammatical resources, resulting in systematic stylistic variation. The study contributes a model of translator individuation that visually and conceptually links cultural reservoirs, identities, affiliations, and individual repertoires to stylistic outcomes in translation, offering a useful lens for analyzing translator subjectivity and sociocultural mediation in literary translation.