Recent studies have shown that thematic change in fiction is driven as much by individual authors' choices as by generational succession: differences between author cohorts explain half the variance in novel themes. This paper examines the role of cohort effects in poetry, extending the analysis beyond vocabulary to poetic features such as meter and rhyme. We use Jensen-Shannon divergence to compare the rates of change across feature levels, identify periods of rapid change, and find the features driving them. Then we apply decomposition methods to estimate how much the features' frequency trajectories are explained by either cohort effects or within-career individual change. Our preliminary results indicate that topics and rhyme features are relatively stable over time, with about half of their variance explained by cohort turnover. Metrical repertoire, however, is particularly volatile throughout time and mostly shaped by individual change within poets' careers (58.7%).

This short paper proposes Computational Environmental Humanities (CEH) as a nuanced methodological approach within Digital Environmental Humanities. While Environmental Humanities have traditionally relied on qualitative interpretation to analyze nature–culture entanglements, CEH emphasizes computational, data-driven, and exploratory methods to investigate environmental questions at scale. The short paper situates CEH within debates on distant viewing and the multimodal turn in Digital Humanities. As a case study, we analyze a large corpus of German-language children’s book illustrations from the long nineteenth century. Using visual embeddings and clustering techniques, we explore how environmental imaginaries of nature, rurality, and urbanity are constructed across thousands of images. Computational methods function here as exploratory tools that surface recurring visual patterns and support hypothesis generation, while interpretation remains central. Rather than replacing established approaches in Environmental Humanities, we argue that CEH rather complements them by enabling new forms of empirical access to large visual archives.

1. Brief summary of the research topic

In this talk, I present one possible approach to analysing chromatic and light-related spectra in the fictional worlds of the leading Czech Romantic writer Karel Hynek Mácha. Using quantitative and statistical methods, we will address the question of which structural principles organise the fictional worlds of Mácha’s prose with respect to their colour and light configurations. In literary prose, colours and lighting are not random features but manifestations of textual intention; together with other components of textual structure and the structure of fictional worlds, they contribute to the overall semantic profile of the work.

If we aim to identify the basic organisational principles of colourfulness and brightness in fictional worlds (in our case, Romantic prose), quantitative and statistical methods are an appropriate choice because they yield objectifiable evidence that can subsequently be interpreted within literary scholarship. The talk will employ both simple quantitative procedures, such as frequency analyses of the so-called basic colour terms, and statistical and corpus-based methods, including PCA, as well as context-oriented techniques such as concordances and collocation analyses. The results will be presented through a series of visualisations (bar charts, PCA plots) and structured tables (collocations, concordances) and will be compared with results produced by a RAG agent.

This paper reconstructs the global and interperipheral circulation of the Romanian novel (originals 1860-2000; translations 1860-2025) through quantitative and network-based modelling. Drawing on Smith’s concept of the “translationscape” (2017) and Baghiu’s semi-peripheral analysis (2021), the study operationalises translation as a relational and historically dynamic system embedded in institutional networks (Bourdieu 1993; Heilbron 1999; Sapiro 2014).

The dataset integrates approximately 8000 Romanian novels with 1400 documented translation editions (c. 250 titles) across more than 50 target languages. Translationscapes are modelled as time-sliced multilayer networks linking works, languages, publishers, and translators. Community detection (Louvain modularity) identifies cohesive clusters, enabling cross-period comparison of structural differentiation.

Preliminary results show that Romanian novel circulation is not reducible to a simple centre-periphery axis. Interwar translations display higher modular differentiation and balanced cluster distribution, often privileging Central and Eastern European languages before Western “core” markets. The socialist period produces denser but more centralised structures, while post-1989 circulation exhibits fragmentation alongside increased linguistic diversification. Re-edition depth further reveals differentiated integration patterns across language zones.

By transforming bibliographic corpora into comparable relational models, the paper demonstrates how world-literature theory can be operationalised through reproducible network analysis. It proposes a transferable framework for modelling peripheral literary mobility and invites discussion on how quantitative translationscapes reshape our understanding of global literary hierarchies.

Large language models represent word meanings as vectors in a multidimensional space. Building on this property, the present study offers a geometric perspective on parallelism in classical Chinese poetry, complementing traditional symbolic interpretations. In regulated verse, parallelism is a crafted pairing of two lines “side by side,” where words in the same positions mirror each other in grammar and meaning (often as matched or contrasted images). To automatically detect parallelism in poetic verse, the authors trained a BERT-based classifier on a dataset of over 140,000 regulated pentasyllabic poems (lüshi 律詩), achieving performance on par with generative models such as GPT-4.1 and DeepSeek R1. Unlike general-purpose models, the custom classifier yields unique insights into how poetic meaning is encoded geometrically: parallel lines exhibit alignment in the model’s attention patterns, with the "key" vectors of corresponding characters pointing in the same direction, while this alignment disappears in non-parallel couplets. This result extends Roman Jakobson's "poetic function” of language (equivalence projected from selection onto combination) by showing that poetic equivalence can be realized not only through sound and sequence but also through spatial relations in semantic vector space. The mechanism for this spatial equivalence can be understood through Peter Gärdenfors's theory of cognitive semantics, which posits that humans make sense of the world by organizing experience into distinct conceptual regions. From this perspective, parallelism functions as a bridging mechanism that temporarily unites such disparate regions of meaning, gesturing towards a deeper, geometric order that underlies language itself.

This short paper proposes a compact workflow for structuring and inspecting a historically situated micro-corpus and demonstrates it on ten dated poems by Arseny Nesmelov published in 1940 in Harbin (day/month/year and page metadata). The aim is exploratory and method-oriented: rather than claiming a finalized interpretation, we show how transparent preprocessing and inspectable outputs support data-grounded inquiry into émigré poetry and its social imaginaries (Hinrichs 1987).

Each poem is encoded with a lightweight schema (ISO date, page, title) and maintained in two text layers to preserve philological traceability while enabling reproducible processing. We extract interpretable features suited to small-scale data, focusing on (i) community indexing via first-person singular/plural forms (“I/We” and related collectives) and (ii) coercion-related imagery and vocabulary (judgement/violence lexica). Frequency profiles, collocation, and KWIC tables provide preliminary, inspectable results. Building on the idea that distributional evidence can complement close reading (Jockers 2013), we report initial findings from these outputs and outline how they will be refined before the conference.

Even at micro-corpus scale, the dated sequence enables cautious comparison across the 1940 timeline and helps identify where collective positioning and coercion imagery are foregrounded. For interpretive accountability, a subset is manually checked for preprocessing errors and false positives/negatives in seed lists, summarized via a brief error typology (polysemy, negation/irony, metaphorically marked lexemes). The contribution is a portable approach for turning émigré poetic materials into reusable data and tracing “I/We” community formation alongside coercion imagery across a dated sequence.