This study presents a hybrid topic modelling framework for computational literary analysis that integrates Latent Dirichlet Allocation (LDA) with sparse Partial Least Squares Discriminant Analysis (sPLS-DA) to model thematic structure and longitudinal dynamics in narrative poetry. As a case study, we analyse Evgenij Onegin-Aleksandr S. Pushkin's novel in verse-using an Italian translation, testing whether unsupervised and supervised lexical structures converge in a small-corpus setting. The poetic text is segmented into thirty-five documents of lemmatised content words, from which five stable and interpretable topics emerge. To address small-corpus instability, a multi-seed consensus protocol is adopted. Using sPLS-DA as a supervised probe enhances interpretability by identifying lexical markers that refine each theme. Narrative hubs-groups of contiguous stanzas marking key episodes-extend the bag-of-words approach to the narrative level, revealing how thematic mixtures align with the poem's emotional and structural arc. Rather than replacing traditional literary interpretation, the proposed framework offers a computational form of close reading, illustrating how lightweight probabilistic models can yield reproducible thematic maps of complex poetic narratives, even when stylistic features such as metre, phonology, or native morphology are abstracted away. Despite relying on a single lemmatised translation, the approach provides a transparent methodological template applicable to other high-density literary texts in comparative studies.
Figures & tables
ID
Hub label
Start (ch., stanza)
End (ch., stanza)
H1
A typical day in Evgenij’s life
(1, XV)
(1, XXVIII)
H2
The friendship between Evgenij and the Poet
(1, XLV)
(1, L)
H3
Tat’jana’s introduction
(2, XXIV)
(2, XXIX)
H4
Tat’jana falls in love
(3, VII)
(3, XXI)
H5
Tat’jana’s letter
(3, XXXI)
(3, XXXI)
H6
Evgenij’s reply
(4, XI)
(4, XVII)
Table 1: Selected list of narrative hubs in Evgenij Onegin , with corresponding chapter (ch.) and stanza ranges.
Topic
Label
nk
γˉk
T 1
Natural & lyrical imagery
7
0.194
T 2
Individual expectations & aspirations
8
0.196
T 3
Lyrical–sentimental effusion
10
0.241
T 4
Social & mundane milieu
5
0.176
T 5
Feelings & social conventions
5
0.193
Table 2: Topics with their interpretive label, per-topic dominant-document count nk , and mean prevalence γˉk .
Figure 1: Top 20 terms per topic (prevalence-weighted). Panel area per topic is proportional to its mean prevalence; tile area per lemma is proportional to its within-topic weight. Italian lemmas (left); English glosses (right).
Figure 2: Narrative map across documents. Grey bands mark the narrative hubs described in Table 1 .
Topic
acc
bacc
acc0
bacc0
Natural & lyrical imagery
0.88
0.81
0.72
0.55
Individual expectations & aspirations
0.76
0.60
0.69
0.53
Lyrical–sentimental effusion
0.86
0.81
0.58
0.47
Social & mundane milieu
0.89
0.68
0.79
0.55
Feelings & social conventions
0.86
0.58
0.76
0.48
Table 3: Cross-validated performance of sPLS–DA OvR classifiers by topic. Columns: mean accuracy ( acc ), balanced accuracy ( bacc ), and stratified random baselines ( acc0 , bacc0 ).
Figure 3: Cross-over heatmap between sPLS–DA core-exclusive lemmas and LDA top lemmas. Diagonal values indicate within-theme overlap; off-diagonal values indicate cross-theme overlap.
Figure 4: Chord diagram showing interrelations between the five main themes (bottom) and their associated subthemes (top).
Theme / Subtheme
Lemmas (IT)
Lyrical–sentimental effusion
Art and culture
voce, pensiero
Characters
Evgenij
Countryside domestic life
casa, fuoco ‡
Dreams and ideals
anima, destino, giorno, mondo, sogno, speranza, tempo
Table 4: Tat’jana’s letter — lexical hub card. sPLS–DA lemmas are denoted by † or ‡ ; † , novel lemma (not in LDA top lists); ‡ , porous lemma (appearing among the top terms of another topic).
This study applies symbolic time series analysis and Markov modeling to explore the phonological structure of Evgenij Onegin-as captured through a graphemic vowel/consonant (V/C) encoding-and one contemporary Italian translation. Using a binary encoding inspired by Markov's original scheme, we construct minimalist probabilistic models that capture both local V/C dependencies and large-scale sequential patterns. A compact four-state Markov chain is shown to be descriptively accurate and generative, reproducing key features of the original sequences such as autocorrelation and memory depth. All findings are exploratory in nature and aim to highlight structural regularities while suggesting hypotheses about underlying narrative dynamics. The analysis reveals a marked asymmetry between the Russian and Italian texts: the original exhibits a gradual decline in memory depth, whereas the translation maintains a more uniform profile. To further investigate this divergence, we introduce phonological probes-short symbolic patterns that link surface structure to narrative-relevant cues. Tracked across the unfolding text, these probes reveal subtle connections between graphemic form and thematic development, particularly in the Russian original. By revisiting Markov's original proposal of applying symbolic analysis to a literary text and pairing it with contemporary tools from computational statistics and data science, this study shows that even minimalist Markov models can support exploratory analysis of complex poetic material. When complemented by a coarse layer of linguistic annotation, such models provide a general framework for comparative poetics and demonstrate that stylized structural patterns remain accessible through simple representations grounded in linguistic form.
Angelo Maria Sabatini
The BioRobotics Institute Scuola Superiore Sant’Anna Pisa, Italy
Thematic indexing -- the practice of assigning structured conceptual labels to sections of text -- is essential to scholarly access in large-scale literary and historical editions, yet it remains a largely manual, labour-intensive process. This paper explores the application of machine learning to automatic thematic indexing, using two substantial sub-corpora of the Complete Works of Voltaire as a test case: the Essai sur les mœurs et l'esprit des nations and the Questions sur l'Encyclopédie. The task is framed as a multi-label classification problem, in which a model must assign the set of index entries that a professional indexer would apply to a given page of text. We compare a range of approaches -- from encoder-based models with classification heads to generative large language models (LLMs) fine-tuned via Low-Rank Adaptation (LoRA) -- spanning model sizes from approximately 3 to 120 billion parameters. Our best-performing model, from the Mistral family in a 4-bit quantised configuration, achieves F1 scores of up to 0.67; we argue that these figures represent lower bounds, given the inherent subjectivity of professional indexing and the frequency with which model predictions prove semantically valid despite diverging from the print index. We further evaluate cross-corpus generalisation and conduct a detailed qualitative analysis of model behaviour on literary and rhetorical features of the source texts that prove particularly resistant to automated treatment. Our findings have implications for the broader challenge of providing structured thematic access to large-scale literary and historical corpora.
Topic modeling seeks to uncover latent semantic structure in text, with LDA providing a foundational probabilistic framework. While recent methods often incorporate external knowledge (e.g., pre-trained embeddings), such reliance limits applicability in emerging or underexplored domains. We introduce PRISM, a corpus-intrinsic method that derives a Dirichlet parameter from word co-occurrence statistics to initialize LDA without altering its generative process. Experiments on text and single cell RNA-seq data show that PRISM improves topic coherence and interpretability, rivaling models that rely on external knowledge. These results underscore the value of corpus-driven initialization for topic modeling in resource-constrained settings. Code is available at: https://github.com/shaham-lab/PRISM.
Tal Ishon, Yoav Goldberg, Uri Shaham
Department of Computer Science, Bar Ilan University