Digital Humanities has largely prioritised big-data methods such as distant reading and large-corpus analysis to identify macro-level literary patterns. While productive, these approaches often flatten affective nuance, marginal subjectivities, and context-specific meanings, particularly in translated, gendered, and regionally situated texts. This limitation is acute in narratives centred on lived experience, emotional labour, and everyday survival. Addressing this gap, the study advances a small-data, annotation-driven approach that foregrounds micro-level textual units to generate deeper interpretive insight.

This study is guided by three central research questions: (1) How can small-scale textual data enable richer interpretive insights into gender, urban precarity, and survival in literary texts? (2) What forms of meaning emerge when literary micro-units such as scenes, emotions, gestures, and character interactions are treated as data rather than narrative background?

This paper aims to demonstrate the methodological and theoretical potential of small data through an annotation-based close reading of English translations of selected Bengali novels by Bani Basu, with particular focus on MOM and The Fifth Man. The significance of the study lies in its focus on non-canonical, translated South Asian literature and its intervention in DH debates on scale, objectivity, and interpretation. By foregrounding gendered and urban experiences articulated through intimate narrative moments, the paper proposes an alternative DH praxis that privileges nuance, ambiguity, and situated meaning.

In this study, micro-narratives are understood as small, self-contained narrative units (MOM &The Fifth Man) such as brief scenes, character interactions, embodied gestures, emotional shifts, and spatial moments that carry disproportionate interpretive weight within the larger textual structure. Rather than functioning as mere descriptive details, these micro-narratives encode complex relations of gender, power, urban precarity, and survival. By focusing on domestic encounters, moments of fear, care, debt, and bodily vulnerability, the study treats such fragments as high-value small data that reveal historically and socially embedded meanings.

Methodologically, the study employs manual, sentence-level annotation of carefully selected narrative fragments rather than large-scale corpus analysis. The annotations focus on emotional states, gendered labour, domestic and urban spaces, bodily vulnerability, and survival strategies. Drawing on scholarship in small data studies (Manovich, 2012), feminist Digital Humanities (D’Ignazio & Klein, 2020), affect theory (Ahmed, 2004), and urban literary studies, the paper conceptualises annotation not as neutral metadata but as a form of scholarly interpretation. Annotation becomes a site where theory, textual evidence, and contextual knowledge intersect. To support interpretive transparency, the study incorporates visual outputs such as annotated text excerpts, emotion maps, and character–space interaction diagrams, which demonstrate how meaning clusters around micro-narrative units.

Thematic Coding: MOM (Bani Basu)

Thematic Coding: The Fifth Man (Bani Basu)

Interpretive Digital Annotation (Recogito) in The Fifth Man :

Interpretive Digital Annotation (Recogito) in MOM  :

The study employs Recogito, a digital annotation platform, as a core methodological tool to carry out close, interpretive annotation, situating the research firmly within Digital Humanities praxis. First chapters from the English translations of MOM and The Fifth Man were manually annotated at the sentence and micro-narrative level, treating annotation as an act of scholarly interpretation rather than neutral markup. For MOM, the annotation scheme foregrounds socio-gendered structures through tags such as Gendered Body, Patriarchy, and Economic Condition, enabling the systematic tracing of women’s labour, bodily vulnerability, dowry practices, and survival under economic precarity. In contrast, annotations in The Fifth Man focus on Subjectivity and Space, capturing fractured interiority, urban disorientation, and the city as an affective terrain.

The results indicate that small textual fragments of domestic scenes, fleeting encounters, moments of fear, care, debt, and endurance function as “high-value data points” that encode dense socio-economic, gendered, and postcolonial realities. In MOM, micro-narratives surrounding women’s bodies, dowry practices, caste hierarchies, servitude, and spectral justice reveal the moral economies of late-capitalist patriarchy and the commodification of female survival. In contrast, The Fifth Man foregrounds fractured subjectivities, urban anxiety, and postcolonial alienation through character-specific events and spatial disorientation, mapping the city as an affective and psychological terrain rather than a statistically measurable space.

One key challenge of this study is the labour-intensive nature of manual, sentence-level annotation, which prioritises interpretive depth over scale and limits the number of texts analysed. A second challenge lies in the subjectivity of interpretive annotation, as researcher positioning affects reliability and reproducibility, creating difficulties for standardisation and comparative analysis within Digital Humanities, despite being acknowledged as an epistemic strength.

I would like to elaborate that the corpus size has intentionally been kept comparatively small because the research is primarily based on the method of digital close reading of selected novels, where annotation functions as the central methodological tool. Since the process of annotation requires detailed textual engagement, interpretation, categorisation, and close examination of linguistic as well as thematic patterns, it becomes difficult to work with a very large corpus while maintaining the depth and accuracy of the analysis. Therefore, the focus of the study is more on qualitative depth rather than quantitative expansion of the corpus.

At the same time, I genuinely attempted to make the corpus more inclusive by searching for Dalit writings and region-specific Bengali literary texts. However, during the process of corpus selection, I encountered several limitations. The selection of texts was dependent upon multiple criteria, including digital availability of the texts, accessibility, suitability for annotation-based analysis, thematic relevance to the research objective, and the availability of reliable and complete versions of the novels. Although I searched extensively through different platforms and sources in order to locate region-specific and marginal literary works, I could not find sufficient materials that fulfilled all these criteria together.

As a result, the final corpus has inevitably become more dependent upon mainstream Bengali literary writings, which are largely dominated by upper-caste or ‘savarna’ Bengali authors. This was not an intentional exclusion of marginal voices; rather, it emerged from the practical and methodological limitations associated with corpus construction in digital humanities research. Nevertheless, throughout the process I have sincerely tried my best to include more diverse, regional, and underrepresented literary voices within the available scope and limitations of the study.

Future research can extend this study by developing collaborative and multi-layered annotation frameworks, involving multiple annotators with diverse theoretical perspectives. Such an approach would allow for comparative interpretation while addressing concerns of subjectivity through dialogic and feminist annotation practices.The study can also be extended by working with original-language texts alongside translations, enabling comparative annotation across linguistic versions. This would deepen insights into how affect, gender, and urban experience are reshaped through translation and would contribute to scholarship on multilingual Digital Humanities.

The discussion positions small data as historically embedded, affect-laden artefacts whose resistance to formal abstraction is an epistemic strength for literary and cultural analysis, as it foregrounds interpretive depth, contextual sensitivity, and ambiguity. It concludes that small-data thinking has significant implications for the ethical development of AI and NLP systems, especially for marginalised languages, translated texts, and region-specific literary traditions.

Key Words :

Small Data, Micro-Narratives, Annotation-Based Analysis, Digital Humanities, Gender, , Translated Literature,  Identity , Thematic Coding ,Recogito