1. Introduction

Literary translation is fundamentally an ongoing negotiation between formal fidelity and readerly acceptance—a tension that becomes particularly acute in the translingual transfer of stream-of-consciousness narrative. The "neurotic rhythm" and psychological authenticity of such narratives rely heavily on language-specific formal markers, such as embedded complex sentences, emphatic punctuation (e.g., dashes and ellipses), and culturally laden psychological vocabulary. When crossing typological boundaries, these features pose significant challenges to translatability.

Fyodor Dostoevsky’s Notes from Underground, a pioneering work of modernist psychological fiction, offers an ideal test case for examining these challenges. Its major translations reflect a clear diachronic spectrum of translational ethics: from Constance Garnett’s domesticated version (1918), to Richard Pevear and Larissa Volokhonsky’s foreignizing approach (2004), to Zhonglun Zang’s creatively compensated Chinese translation (2019). The fundamental divergence in strategy among these versions provides a rich comparative basis for investigating how stream-of-consciousness features are transformed, attenuated, or recreated across languages.

This study aims to address a gap in the literature by developing an integrated framework combining Bakhtinian dialogic theory with empirical translation analysis. It systematically examines the mechanisms of cross-linguistic transfer—from Russian to English and Chinese—in the stream-of-consciousness passages of Notes from Underground. Through a mixed-method approach incorporating both quantitative and qualitative analysis, the study seeks to reveal patterns and constraints in translating psychological narrative across disparate linguistic systems.

The paper is structured as follows: Section 2 reviews relevant literature; Section 3 outlines the methodology; Sections 4 and 5 present and discuss the findings; and Section 6 concludes with implications and reflections.

2. Research background

2.1. The translation of stream-of-consciousness features

Early research predominantly focused on stream-of-consciousness writers from Western Europe, such as Virginia Woolf and James Joyce. The core of translation studies for stream-of-consciousness narrative has consistently revolved around the tension between the idiosyncrasies of literary form and the universal principles of translation. On one hand, as argued by Brian Fitch, "the signifiers of two different linguistic systems are always and necessarily incompatible with one another" [1]. This fundamental linguistic disparity is drastically amplified by the radical distortion of linguistic conventions in stream-of-consciousness writing—exemplified by elongated compound sentences, fragmented punctuation, and highly private associations—thereby provoking philosophical debates regarding its potential "untranslatability". On the other hand, the cultural approach to translation advocated by Susan Bassnett offers a counterargument grounded in optimism. She contends that texts are indeed translatable, that commonalities exist amidst cultural differences, and that translators can employ creative strategies to achieve functional equivalence [2]. Nevertheless, a systematic analysis of Russian stream-of-consciousness, particularly Dostoevsky's "neurotic" style, remains notably underexplored in scholarly inquiry.

2.2. Bakhtin’s theory of polyphony and its implications for translation

Mikhail Bakhtin’s theory of polyphony [3] is central to understanding Notes from the Underground. He proposed that Dostoevsky created a world of independent and unmerged consciousnesses, in which the character’s voice engages in dialogue with the author’s on equal footing... This polyphonic structure is profoundly manifested through the technique of stream of consciousness, which serves as one of the unique features of psychological novels. The term "stream of consciousness writing" refers to a narrative mode that expresses the uninterrupted flow of thoughts and feelings of a narrator or character, allowing readers to trace shifts in emotional states and perceptual processes [4]. It essentially functions as an interior monologue that aims to capture the natural, often associative, progression of a character’s thinking.

Thus, the challenge of translation lies in preserving this "dialogic discourse"—embedded within the subjective and fluid stream of consciousness—along with the inherent contradictoriness and unfinalizability of the character’s consciousness. Failure to do so would reduce the text to a monologic narrative, thereby flattening the very polyphony that defines Dostoevsky’s art.

2.3. The critical genealogy of translations of Notes from the Underground: from domestication and foreignization to cross-linguistic transfer

The polyphonic narrative and stream-of-consciousness style of Notes from the Underground pose significant challenges for its translators, giving rise to distinctly varied translations that have sparked prolonged scholarly debate. These versions clearly reflect the diachronic evolution of translation ethics.

Firstly, Constance Garnett’s translation (1918) is regarded as an exemplary early model of domestication strategy [5], though its literary fidelity remains contentious. Known for her rapid and prolific output, Garnett prioritized fluency and readability in her translations. However, subsequent studies have widely criticized her approach for sacrificing the complexity and heteroglossia of the original text to accommodate the aesthetic preferences of English-language readers. The most representative criticism comes from Vladimir Nabokov, who sharply described Garnett’s translations as "dry", "flat", "always unbearably demure". Furthermore, Edmund Wilson noted that Garnett’s work was "full of errors" [6]. By simplifying long and complex sentences, replacing culture-specific terms, and toning down the idiosyncrasies of characters’ voices, Garnett’s approach significantly diminished the polyphonic tension and claustrophobic intensity emblematic of the “underground” spirit. Her translation exemplifies a typical domestication strategy, mirroring the translational norms of the early twentieth century.

In contrast, the collaborative translation by Pevear and Volokhonsky (1994) is widely recognized as a representative example of contemporary foreignization strategy, aiming to restore Dostoevsky’s "original voice". Their cooperative model—whereby Volokhonsky produced a literal draft which Pevear then refined stylistically—itself constitutes a methodological innovation. Their translation deliberately preserves Russian syntactic structures, punctuation, and cultural imagery, striving to bring the reader to the author. This approach has "almost unanimously praised by reviewers and Slavic scholars", being "lauded for restoring the idiosyncrasies of the originals—the page-long sentences and repetitions of Tolstoy, the cacophonous competing voices of Dostoevsky". However, this foreignization strategy also introduces issues of "readability". Some critical reviews note that the translation is "too literal or prone to unidiomatic turns of phrase" [7].

Finally, the Chinese translation by Zang Zhonglun (2019) exemplifies the unique challenges and strategies involved in cross-linguistic translation between Chinese and Russian. Due to the significant typological distance between the two languages, pure syntactic imitation is nearly impossible. As a result, Zang’s "literal translation" is, in essence, a form of creative literalism. While striving to preserve the sentence structures and narrative rhythm of the original text, his translation necessarily employs language-specific compensatory strategies—such as using four-character idioms to enhance expressiveness, adopting modal particles to convey emotional intensity, and incorporating intratextual annotations to handle culture-specific terms. However, his frequent use of dialect and colloquialisms in place of neutral expressions in the original, though enhancing the vivacity of the text, may also introduce cultural dissonance.

In summary, the scholarly critique of the three aforementioned translations of Notes from the Underground elucidates a clear trajectory: from a reader-friendly domestication, to a text-centric foreignization, and finally to a creative literalism that directly confronts the linguistic divide. This evolution provides a robust theoretical foundation for the subsequent micro-level textual analysis undertaken in this study.

2.4. Identification of the research gap

Although general evaluations of these translators do exist, the current body of scholarship remains predominantly focused on stream-of-consciousness works by Western European authors such as Woolf, Joyce, and Proust—especially in the context of English-Chinese translation. By contrast, Russian stream-of-consciousness literature, particularly exemplified by Dostoevsky’s Notes from Underground, has been relatively underexplored. More critically, while existing contributions provide valuable insights, they often rely on theoretical discussion or qualitative commentary rather than rigorous empirical inquiry. There is a discernible lack of quantitatively oriented studies that employ parallel corpora and manual annotation techniques to systematically compare the transference of stream-of-consciousness features across distinct translational paradigms. Furthermore, substantial empirical evidence—particularly from reader-response experiments—is notably absent regarding how specific translation choices affect reader perceptions, such as the inherent trade-off between reading fluency and psychological authenticity. It is this specific gap in the literature that the present study aims to address.

3. Methodology

3.1. Corpus construction and text selection

A tri-lingual parallel corpus was constructed comprising the following components:

Source Text (ST): The Russian text was sourced from the authoritative edition of Dostoevsky's collected works.

Target Texts (TT):

TT1 (EN): Pevear & Volokhonsky’s translation (2004). 

TT2 (EN): Constance Garnett’s translation (1918). 

TT3 (ZH): Zang Zhonglun’s translation (2019).

Three key stream-of-consciousness monologue segments were selected, with a combined total of approximately 4,600 Russian words. These excerpts were drawn from the following narrative situations: the opening monologue, the psychological aftermath of the conflict with the officer, and the self-revelatory soliloquy on the snowy night. This selection ensures thematic and stylistic diversity within the analyzed corpus, covering both introspective and emotionally charged passages essential to the novel’s narrative strategy.

3.2. Analytical framework

The analysis was conducted across three distinct levels. At the syntactic level, manual annotation was employed to examine syntactic transformation strategies—namely splitting, one-to-one correspondence, and blending—across all three translated versions, along with an assessment of punctuation fidelity, using the em dash as a representative marker. On the lexical and narrative levels, a parallel comparative approach was adopted for the CG and PV English translations. Through a controlled participant experiment, differences and correlations between Psychological Authenticity and Reading Fluency were empirically investigated. Additionally, the Chinese translation was analyzed separately to identify and describe compensatory mechanisms specific to the target language. In this phase of the research, a qualitative and example-based analytical approach was adopted.

4. Experimental design and data analysis

4.1. Syntactic-level analysis

4.1.1. Definition of the syntactic unit and analytical methodology

Given the syntactic idiosyncrasies inherent in stream-of-consciousness literature, this study adopts an operational definition of the "sentence" centered on the "stream-of-consciousness sense unit". In determining boundaries, periods, question marks, and exclamation points are treated as primary boundaries, warranting segmentation. For semicolons and em-dashes, contextual judgment is applied: if the preceding and subsequent content belongs to the same intensive, coherent unit of psychological activity or narrative thought, it is not segmented and is treated as a single complex sentence; if the content following the mark introduces a clear shift in thought or a new narrative direction, it is segmented. This approach aims to preserve the abruptness and dialogic nature of the original thought processes, aligning with the analytical principles of Bakhtin’s theory of polyphony.

To quantitatively analyze translators’ strategies in handling complex syntax, all Russian source sentences with a word count ≥25 are defined as "long sentences" and serve as the unit of analysis (N=61). Through manual comparison of the source and target texts, each sentence was annotated for its type of syntactic transformation, classified as follows:

Split: A Russian long sentence translated into two or more sentences in the target language.

One-to-One: A Russian long sentence translated into one long sentence in the target language.

Fusion: Two Russian sentences (one of which may be long) merged into a single sentence in the target language.

4.1.2. Results and analysis

The distribution of syntactic operations is presented in Table 1; the analysis of dash handling strategies is shown in Table 2.

Analysis of the data reveals three distinct translation philosophies:

P&V’s Absolute Foreignization: As shown in Table 1, P&V’s data (0% split, 0% fused, 100% one-to-one) is remarkably consistent. This indicates a core principle of maximally preserving the original syntactic form. This pursuit of formal equivalence is further evidenced in punctuation (Table 2), where they retained 85.7% of the original dashes to maintain the rhythm and abruptness of the source text's thought processes. Their extremely low omission and addition rates confirm a strategy of radical foreignization, aiming to immerse the reader in Dostoevsky’s distinctive, often discordant narrative rhythm—even at the cost of challenging target-language readability.

Garnett’ Selective Domestication: In contrast, Garnett’s strategy shows strong interventionist tendencies across all dimensions. She was the only translator to employ both splitting (11.5%) and fusion (1.6%) (Table 1), actively restructuring sentences for fluency. Correspondingly, in punctuation (Table 2), she omitted 50% of dashes and freely added new ones (17.9%). This consistent pattern reflects a domestication-oriented approach: she intervenes where she deems necessary to optimize clarity and align the text with the expectations of the English literary tradition, prioritizing a smooth and readable translation.

Zang’s Asymmetric Fidelity: Zang’s strategy demonstrates a unique form of principled literalness. Syntactically (Table 1), he was highly conservative (0% split), adhering to original sentence boundaries at a rate of 96.7% one-to-one. In punctuation (Table 2), however, he omitted a significant number of dashes (28.6%) but scrupulously avoided adding any (0%). This pattern reveals a core principle: maximal fidelity to source form where possible, coupled with necessary adaptations driven by an awareness of typological differences between Russian and Chinese. His is a literalist approach defined by caution and balance.

4.2. Lexico-narrative level analysis

4.2.1. Research design

This study used a psycholinguistic experimental method to quantitatively compare the English translations of Fyodor Dostoeveyky’s Notes from Underground by Constance Garnett (CG) and Richard Pevear & Larissa Volokhonsky (P&V). The analysis was carried out on two complementary levels: (1) the lexical level, which examined differences in emotional intensity of key psychological terms, and (2) the narrative level, which assessed overall reading fluency and psychological authenticity of extended passages. This combined approach offers a more complete understanding of how domestication and foreignization strategies operate in literary translation.

4.2.2. Participants

Ten native English speakers participated in the study, comprising five females and five males aged between 18 and 35. All participants held at least a bachelor’s degree, with educational or strong personal interests in English literature, ensuring they possessed the necessary literacy skills and sensitivity to nuanced literary language.

4.2.3. Lexical-level experiment: emotional intensity assessment

Eight pairs of translated words (C.G vs. P&V) representing key psychological states were selected for the lexical-level experiment. Each target word was embedded into a short English sentence derived from its original context to maintain narrative coherence. The stimuli were integrated into an online questionnaire, with all 16 items presented in fully randomized order. Translator attribution was concealed throughout the study to mitigate bias. Participants were instructed to read each sentence and rate the emotional intensity of a bolded target word using a 5-point Likert scale, ranging from 1 (very weak) to 5 (very strong). A paired-sample t-test was conducted to examine whether average intensity ratings differed significantly between the two translations across the eight word pairs, with a significance threshold set at α = .05.

4.2.4. Narrative-level experiment: reading fluency assessment

Extended passages of interior monologue from the novel, presented in both C.G and P&V translations, served as the stimuli for the narrative-level experiment. The passages were presented in random order. After reading each passage, participants completed two separate 5-point Likert scales. The Reading Fluency Scale assessed ease of reading through four items: Cognitive Effort (from 1 = very effortful to 5 = effortless), Linguistic Naturalness (1 = very unnatural to 5 = very natural), Rhythmic Comfort (1 = very jarring to 5 = very smooth), and Overall Readability (1 = very poor to 5 = very good). Mean scores for each scale were computed for both translations.

4.2.5. Results and analysis

The emotional intensity assessment is presented in Table 3, and the reading fluency assessment is presented in table 4.

The results of the paired-sample t-test (table 1) revealed a statistically significant difference in the perceived emotional intensity between the two translations (p < .001). This highly significant finding provides strong quantitative evidence that Garnett's (C.G.) and Pevear & Volokhonsky's (P&V) translations systematically differ in their rendering of key psychological state words.

Descriptive analysis of participant ratings revealed a clear differential profile between the two translations across the two dimensions. Constance Garnett’s domesticated version received notably higher ratings on reading fluency (M = 4.35), though it scored relatively lower on psychological authenticity (M = 3.31). In contrast, the foreignized translation by Pevear and Volokhonsky achieved the highest score in psychological authenticity (M = 4.24), while receiving the lowest rating in reading fluency (M = 3.43).

The findings of this study provide robust empirical support for the classic theoretical trade-off between domestication and foreignization in translation.

The high mean reading fluency score of 4.35 for the CG translation indicates that idiomatic lexical choices, syntactic structures aligned with English conventions, and fluent narrative pacing successfully optimized the target-language reader’s experience, significantly reducing cognitive load. However, this optimization came at the cost of reduced psychological authenticity (M = 3.31). While the translation may be more "readable", it effectively flattens the abrupt, self-contradictory, and neurotic rhythm that characterizes the underground man’s unique voice in Dostoevsky’s original.

Conversely, the elevated psychological authenticity score of 4.24 for the P&V translation demonstrates that by deliberately preserving source-language syntactic structures, lexical peculiarities, and narrative rhythm, their approach successfully transplants the tortured, self-lacerating psychological state of the original into English, offering a higher-fidelity rendition of its psychological realism. Nevertheless, this foreignizing estrangement inevitably increased reading difficulty for English-speaking participants, resulting in a significantly lower fluency score (M = 3.43).

In summary, these results suggest that when translating psychologically complex texts such as Notes from Underground, it is particularly challenging to achieve optimal performance in both fluency and authenticity. The choice of translation strategy is ultimately a value-based decision: whether to prioritize readerly comfort or interpretive fidelity to the author’s psychological portrayal.

5.  Compensatory mechanisms in Zang Zhonglun’s Chinese translation

Moving beyond the comparative framework of the two English versions, this study separately examines the Chinese translation by Zang Zhonglun to elucidate how he employed target-language-oriented compensatory strategies to artistically counteract potential losses inherent in translating from Russian to Chinese. The analysis reveals that Zang’s compensatory mechanisms operate primarily across two dimensions.

5.1. Compensation through rhythm, lexical naturalization, and cultural imagery

The translator actively leverages language-specific resources of Chinese—including its rhythmic patterns, fixed expressions, and cultural lexicon—to enhance the literary quality and cultural resonance of the translation while maintaining faithfulness to the source text.

5.1.1. Creative use of four-character phrases and idioms (成语, chengyu)

Zang frequently utilizes four-character phrases and classical idioms to transform straightforward Russian statements into expressions rich in meaning and aligned with Chinese aesthetic preferences. The effectiveness of this strategy derives from the inherent advantages of Chinese idioms: they are often symmetrical in structure, possessing beautiful syllables and a coordinated rhythm that evokes a strong sense of aesthetic pleasure. Furthermore, they are generally concise in form yet vivid and expressive in effect. This approach is deeply rooted in a long-standing Chinese tradition of applying domestication in idiom translation, which emphasizes readability and cultural resonance for the target audience [8].

Example 1: The simple Russian clause "Я прежде служил, а теперь не служу" ("I used to serve, but now I do not") is rendered as “我以前在官署供职,可是现在已挂冠归隐.”.

Analysis: The idiom “挂冠归隐” (guà guān guī yǐn—literally, “hang up one’s official hat and live in seclusion”) does far more than convey the basic meaning of leaving employment. It compensates by importing a profound cultural subtext, evoking a classical Chinese trope of scholarly withdrawal that carries connotations of integrity and finality, thus adding significant literary and historical weight absent from the original.

Example 2: The repetitive structure "Я злой человек. Непривлекательный я человек" ("I am a wicked man. I am an unattractive man.") is translated as “我是一个心怀歹毒的人.我是一个其貌不扬的人.”.

Analysis: The parallel four-character phrases “心怀歹毒” (xīn huái dǎidú—harboring malice) and “其貌不扬” (qí mào bù yáng—of undistinguished appearance) compensate for the original repetition by creating a rhetorically sophisticated, rhythmically balanced couplet in Chinese, elevating the stylistic register.

Example 3: The emotionally charged repetition "Мы бы так зажили! так зажили!" ("How we would have lived! lived!") is translated as “我们将会握手言欢!成为莫逆之交!”.

Analysis: The consecutive idioms “握手言欢” (wòshǒu yánhuān—to hold hands and converse cheerfully, denoting reconciliation) and “莫逆之交” (mònì zhī jiāo—friendship with no discord, bosom friends) powerfully compensate for the emotional intensity of the original. This translation vividly captures the speaker’s fervent excitement, a effect far surpassing a literal translation.

5.1.2. Compensation and equivalence of cultural imagery

Zang adeptly uses culturally specific Chinese concepts to create pragmatic equivalence for Russian expressions.

Example 4: The phrase "Он не выдержал и продолжал далее" ("He could not bear it and continued further") is translated in the afterword as “他忍不住继续秉笔直书”.

Analysis: “秉笔直书” (bǐng bǐ zhí shū) is a core concept from Chinese historiographical tradition, describing the solemn act of a court historian writing truthfully without fear or favor. Its use here compensates for the neutrality of “continued writing,” imbuing the narrator’s action with a sense of gravity and literary solemnity—as if he is composing an honest record of his inner world—thus exemplifying high-level cultural compensation. Furthermore, for specifically Russian concepts, Zang’s translation consistently employs a strategy of literal translation coupled with explanatory notes, ensuring comprehensibility for the Chinese reader while preserving foreign cultural elements.

5.2. Compensation through mood and affect

Explicitation of Modal Particles (语气助词, yǔqì zhùcí) is widely used to meticulously reconstruct the underground man’s complex and shifting vocal tones. A quantitative analysis of a ~10,000-character sample identified 213 added modal particles (e.g., “呢” ne, “吧” ba, “啊” a, “吗” ma). These additions are highly deliberate and serve to compensate for subtle modalities—hesitation, rhetorical questioning, emphasis, exclamation—that are only implied in the Russian original.

These pervasive particles effectively soften the potential stiffness of a literal translation. They inject a flowing, colloquial breath into the narrator’s monologue, making his processes of self-analysis and self-contradiction sound more authentically neurotic and immediate. This strategy significantly enhances the emotional impact and immersive quality of the psychological narrative.

In summary, Zang Zhonglun’s strategy transcends literal translation; it constitutes a deep, creative, and target-language-oriented compensation. By masterfully mobilizing the unique resources of Chinese—its rhythm, idioms, cultural imagery, and modal particles—Zang successfully recreates the voice of the underground man, both frantic and profound, within a Chinese linguistic and cultural context. His translation offers an inspiring Sinophone solution to the problem of translatability for stream-of-consciousness literature, demonstrating that his work is not merely a linguistic transfer but a veritable recreation of literariness in a new cultural soil.

6.  Conclusion

This study has established a systematic and multi-dimensional framework for analyzing the cross-linguistic transfer of stream-of-consciousness narrative, a genre long considered among the most challenging to translate due to its syntactic and psychological complexity. By integrating fine-grained textual analysis with empirical reader-response data, the research moves beyond speculative theoretical discourse to offer an empirically grounded demonstration of how stylistic and psychological features are negotiated across languages.

Key findings underscore the profound impact of translation strategy on the reproduction of stream-of-consciousness features: foreignization, as exemplified by Pevear and Volokhonsky, proved effective in preserving the “neurotic rhythm” and syntactic fragmentation characteristic of Dostoevsky’s original, albeit at the cost of readability. In contrast, Constance Garnett’s domesticating approach significantly enhanced fluency while reducing psychological authenticity. Furthermore, the study revealed that typological distance between source and target languages necessitates the use of creative compensatory strategies, as illustrated by Zang Zhonglun’s translation, which relied heavily on idiomatic and rhythmic adaptations to approximate the effects of the original.

These outcomes carry important implications for both translation theory and practice. They provide measurable evidence of the inevitable trade-offs between fidelity and acceptability, particularly in linguistically and culturally distant contexts. Moreover, the methodological approach developed here—combining parallel corpora analysis with reader-centred experiments—offers a replicable model for future research on literary translation.

That said, a limitation of this study is its focus on a single author and text, which restricts the generalisability of the findings. Future studies could productively apply this framework to other stream-of-consciousness writers across a wider range of language pairs. Additional methodologies, such as eye-tracking, could also be incorporated to provide deeper insights into the cognitive load associated with different translation styles. Ultimately, this research contributes not only to Dostoevsky translation studies but also to the broader interdisciplinary dialogue between literary linguistics and empirical translation research.