Daejeon, July 27–31
This paper examines how Soviet drama for adult audiences in the final years of Stalinism (1945–1953) constructs the relationship between the private family and the ideological “Great Family” of the socialist collective—a symbolic model of social order in which the political leader is positioned as a paternal figure and citizens are imagined as metaphorical “children,” bound by affective ties of duty and obedience rather than by blood. Focusing on linguistic and thematic patterns, it analyzes how postwar theatre mobilizes gender, labor, and kinship to shape subjectivity and regulate affect in accordance with Stalinist cultural norms.
The study is based on a curated corpus of 46 plays written and staged in the immediate postwar period. It combines TEI-based text encoding with TF-IDF keyword analysis, topic modeling (LDA), and statistical comparison of gendered speech (Z-analysis). I argue that late Stalinist drama subordinates private emotional life to the imperatives of labor, vigilance, and collective mobilization, while selectively deploying kinship as a marker of ideological alignment.
The project builds on scholarship emphasizing symbolic kinship and the “mythology of the great Soviet family” in Stalinist culture (Clark 2000), as well as the performative nature of loyalty and vigilance (Dobrenko 2020). Methodologically, it follows distant reading (Moretti 2013) and computational literary analysis (Jockers 2013; Underwood 2019), while engaging with debates on small-data analysis, annotation as interpretation, and the tension between quantitative pattern detection and close reading (Bode 2018; Piper 2020). It demonstrates how DH approaches illuminate patterns difficult to trace through close reading alone.
The corpus comprises 46 Soviet plays for adult audiences written between 1945 and 1953, averaging approximately 25,000 tokens per play. It includes prize-winning works published in Sovetskaia dramaturgiia—associated with the Stalin Prize—as well as widely circulated plays from the journal Teatr. All texts were digitized using OCR, manually corrected, and encoded in TEI (Burnard 2014; Ide / Véronis 1995; Romary 2008; Sperberg-McQueen / Burnard 1995), producing a small, homogeneous dataset suitable for computational analysis.
During preprocessing, stage directions and other non-spoken elements were removed to isolate character utterances. Frequent geographical references were filtered to reduce noise in keyword extraction. All texts were lemmatized using the Python library pymystem3, designed for Russian morphological analysis; given the high degree of inflection, this was essential for reducing sparsity. Remaining errors were corrected through manual normalization.
To identify characteristic lexemes, I computed TF-IDF scores (Manning et al. 2008; Salton / Buckley 1988) using external reference corpora (children’s plays of the Stalinist period and Thaw-era drama). High-TF-IDF words were grouped into four semantic categories: labor and production; the glorified national past; external threats and defense of the Motherland; and family. Only around 8% of the distinctive vocabulary relates to family, while nearly 44% concerns labor and state-oriented activity, reflecting the dominance of work and vigilance over private relations.
To examine how kinship operates within this imbalance, I introduced a character-based analytical layer. Thirteen character types were identified through close reading and existing typologies of socialist realist culture (Dobrenko 2020; Dunham 1990), including Misguided Managers, Progressive Workers, Military Figures, False Heroes, and other ideological roles. These types were manually annotated using a rule-based schema specifying occupational role, ideological alignment, and narrative function. Analysis was conducted in Python (NLTK, scikit-learn, Gensim), integrating qualitative annotation with quantitative modeling.
TF-IDF analysis by character type shows that kinship is unevenly distributed and functions as a marker of ideological positioning. Progressive Workers and Model Kolkhoz Members consistently invoke family ties, whereas other groups deploy kinship vocabulary differently. The Misguided Manager is associated with quasi-kinship terms, pointing to informal networks of patronage (cf. Ledeneva, 1998). The False Hero exhibits clusters of domestic vocabulary embedded in narratives of deception and fictive family incorporation, while the Progressive Worker mobilizes extended kinship aligned with the symbolic “Great Family” (Clark 2000). Kinship thus emerges as a structured ideological resource rather than a stable thematic domain.
Gendered speech patterns were examined using Z-score analysis (Evert 2008), focusing on fictional male and female characters as categories of representation (Rybicki 2016; Underwood 2019; Weidman / O’Sullivan 2017). Female characters refer to family and domestic relations significantly more often, whereas male characters predominantly use vocabulary associated with the state and institutional power. This distribution aligns with Stalinist gender ideology, which positioned women as emotional stabilizers of the socialist household and men as embodiments of labor and political authority (cf. Fürst 2010; Fitzpatrick 1999; Attwood 1999; Buckley 1989; Zdravomyslova / Temkina 2003).
Finally, topic modeling (LDA) reveals four stable thematic clusters (Blei et al. 2003; Blei 2012): kolkhoz and family; industrial production; the glorified national past; and wartime discourse. Crucially, kinship does not emerge as an independent topic but is embedded within the language of labor and institutional organization, indicating that family discourse is structurally integrated into narratives of work and duty.
These findings demonstrate that family in late Stalinist drama is not merely a thematic residue of private life but a linguistic and ideological mechanism. Kinship markers are selectively deployed to differentiate morally and politically “correct” subjects from deviant or suspect figures, while gendered speech patterns reinforce normative models of citizenship.
By combining lexical analysis, character typology, gendered comparison, and topic modeling, the paper shows how computational methods reveal structural regularities difficult to detect through close reading alone. Kinship emerges not as a separate domain but as a flexible semiotic resource embedded within the language of labor, authority, and collective belonging. The central claim is that in late Stalinist drama, kinship is not simply represented but instrumentalized. It serves as an idiom through which labor is moralized, authority is legitimized, and the relationship between the individual and the state is rendered in affective, familial terms.
The paper contributes to digital humanities by demonstrating the analytical potential of small, annotated corpora and by extending computational literary analysis to a non-Anglophone dramatic tradition. It offers a new perspective on late Stalinist culture by showing that the boundary between the private and the political is systematically reorganized through the controlled deployment of kinship as a symbolic resource.