Since the rise of distant reading, quantitative and data-driven approaches have become central to digital humanities research on literature. Among them is the literary mapping, where maps function as tools for abstraction and comparison (Moretti 1998; 2005). By projecting narrative texts onto geographic space, mapping methods seek to reveal spatial distributions and large-scale patterns that remain difficult to perceive through close reading alone (Moretti, 2005, pp.53-54). However, for many literary objects, especially the work of a single author, the available textual corpus is limited in scale. Mapping-based analysis often produces confirmatory rather than revelatory results, merely graphic confirmation of what was never really in doubt (Davis, 2011, p.16). This limitation raises a fundamental question for literary mapping: should mapping always aim at the production of patterns, and how can mapped data remain open to interpretation?

Addressing this problem requires a reconsideration of how measurement relates to literary concepts and interpretation. Drawing on Thomas Kuhn, Moretti argued that measurement alone cannot generate new conceptual knowledge (Yang, 2017). Moretti proposed operationalizing as a methodological leap from measurement to reconceptualization, describing a process in which theoretical concepts are translated into operations that enable measurement, and then returned to theory for critical revision (Moretti, 2013). This study adopts operationalizing as a methodological stance that keeps literary interpretation central to quantitative work.

The research focuses on four novels by Roberto Bolaño—2666, Los sinsabores del verdadero policía, Los detectives salvajes, and Sepulcros de vaqueros. Bolaño has frequently been discussed in terms of postnationalism, deterritorialization, and spatial anxiety (Medina, 2009; López-Vicuña, 2012; Figueroa, 2013). Scholarship has emphasized Santa Teresa as both a symbolic space of globalization and a narrative node that concentrates violence, academic mobility, and literary pursuit (González-González & Candia-Cáceres, 2017; Yan, 2017). Given that Bolaño’s novels articulate postnational identity through spatial anxiety and deterritorialized movement, literary cartography (Jameson, 1992; Tally, 2011) offers a theoretical framework capable of linking narrative space, symbolic mapping, and questions of belonging.

Methodologically, the study operationalizes the literary concept of mobility in order to examine how spatial anxiety and postnational identity are narrated across different spaces and social groups in Bolaño’s fiction. Mobility is approached as a narrative and social practice that reflects unequal access to movement, forms of belonging, and narrative function. While Wilkens et al. (2024) conceptualize mobility through spatial metrics such as geographic diversity and travel distance, this study shifts the focus toward a linguistic dimension. It is measured using a mobility density indicator, defined as the frequency of movement-related verbs per thousand words. A movement verb list was constructed through a combination of word embedding expansion and manual curation. In addition, transportation modes are identified and classified to examine mobility media, following sociological approaches that conceptualize mobility as a socially differentiated resource rather than a neutral act of movement (Adey, 2017; Massey, 2013; Cresswell, 2010).

Preliminary results indicate that literary critics in the novels exhibit higher mobility density in European cities and Mexico City, while episodes set in Santa Teresa display lower mobility, particularly in sections involving academic exile and gendered violence. Transportation choices further differentiate social positions, with private vehicles associated with mobile literary elites. These findings suggest that mobility in Bolaño’s fiction is unevenly distributed and socially structured.

To capture narrative movement at a finer scale, the study initially attempted to extract character trajectories automatically using a large language model configured for structured information extraction. Due to low accuracy and unstable outputs, this approach was abandoned. All character movements are therefore annotated manually, highlighting the limitations of current language models for small, ambiguous literary corpora.

In the next stage, both geographic maps and abstract maps will be produced to avoid conflating literary space with physical geography. Geographic maps anchor narrative movements to verifiable locations, while abstract maps emphasize discontinuities, gaps, and character-centered trajectories that resist realistic reconstruction (Piatti et al., 2009). In line with literary cartography theory, mapping functions not as an end in itself, but as a means to generate new questions about how narrative space, identity, and mobility are conceptualized in Bolaño’s work.