Bai Juyi's *Pipa Xing*, a Tang Dynasty poem, is renowned for its accessible style and portrayal of human experiences. Its numerous English translations provide a rich case study for exploring translation styles. This research employs Systemic Functional Linguistics (SFL) framework, focusing on text complexity as a key indicator of translation style. SFL views grammar as a means of construing experience, and text complexity is explored through lexical density (LD) and grammatical intricacy (GI). The study uses individuation theory within SFL to understand how translators' identities and affiliations influence their resource allocation and readership alignment, ultimately affecting translation styles. The research questions are: (1) What stylistic differences in text complexity do the translations exhibit? (2) How do translators' identities impact their meaning resource allocation and target readership affiliation, influencing translation styles?

The study draws upon SFL's three-dimensional model of language (phylogenesis, logogenesis, ontogenesis), corresponding to instantiation, realization, and individuation. Realization focuses on the relationship between semantic and lexicogrammatical strata. Instantiation connects language systems to instances of language use. Individuation explores how individual translators, constrained by cultural contexts, use allocated resources and establish affiliations with the target readership. Previous research on text complexity highlights its multidimensionality, measurable through LD and GI. Studies of translation acknowledge the translator's subjectivity and the influence of sociocultural context on translation choices. Individuation theory offers a framework to examine how translators' identities shape their interpretations and choices during the translation process.

The study uses Bai Juyi's *Pipa Xing* and nine English translations as its corpus, classifying translators into three groups based on nationality: Chinese, British/American, and Sino-Western collaborative. SysFan and SPSS 22.0 were used for quantitative and qualitative analysis. SysFan assisted in annotating taxis and logico-semantic relationships, automatically calculating LD and GI. SPSS 22.0 performed ANOVA tests to identify significant differences in LD and GI across translations. The poem was divided into six rhetorical stages for stage-specific analysis. LD was calculated as the ratio of lexical items to ranking clauses, while GI was calculated as the ratio of ranking clauses to sentences. The analysis also examined the distribution of taxis (parataxis and hypotaxis) and logico-semantic relations (elaboration, extension, enhancement) to understand the hierarchical and narrative characteristics of each translation. Cook's function word list was used to filter function words from the calculation of LD.

Statistical analysis revealed no significant differences in overall lexical density across the ten texts (original and nine translations). However, significant differences were found in grammatical intricacy. Translations with higher GI showed a preference for paratactic relationships (less hierarchical), while those with lower GI showed more hypotactic relationships (more hierarchical). Analysis of logico-semantic relationships revealed variations in emphasis on elaboration, extension, and enhancement. Some translations exhibited features of spoken language (high GI, paratactic relationships), others showed characteristics of written language (lower GI, hypotactic relationships), and some showed a mixture of both. The proportion of enhancement relationships indicated the strength of narrative features in each translation. The study identified three textual characteristics among the translations: spoken, written, and a mixed style. Some translations also showed stronger narrative features than others. The different translators' choices regarding lexical density, grammatical intricacy, taxis and logico-semantic relationships significantly influenced the translation styles.

The observed variations in translation styles are attributed to the translators' individuation processes. The study examines how translators' master identities (nationality), professional affiliations, and target readership affiliations shaped their lexicogrammatical choices. Chinese translators (Xu Yuanchong and Lin Yutang) produced translations with a spoken-written hybridity and relatively low grammatical intricacy, reflecting their cultural positioning and translation goals. British and American translators (Cranmer-Byng, Watson, Gaunt, Giles) showed varied styles, reflecting their diverse backgrounds and readership targets. Sino-Western collaborative translations (Bynner & Kiang, Yang & Yang, Zhang & Wilson) showed a balance of cultural considerations and target readership alignment. The translator's individuation process, encompassing resource allocation from the cultural reservoir and affiliation with the target readership, is a key factor in shaping translation styles.

The study demonstrates that translation styles are not arbitrary but shaped by the translators' individuation processes. Differences in text complexity, including lexical density and grammatical intricacy, reflect the translators' conscious and unconscious choices based on their identities and affiliations. Future research could expand this investigation to other literary works and languages, exploring the interplay between individuation, translation strategies, and the reception of translated texts. The study highlights the importance of considering the translator's role as a mediator between source and target cultures.

The study focuses on a specific corpus of *Pipa Xing* translations and may not generalize fully to all translations. The quantitative analysis relies on specific tools and methodologies, and alternative approaches may yield different results. The interpretation of stylistic variations relies on qualitative analysis, potentially involving subjective judgments.