This article presents the Polish contribution to the PressMint project, a CLARIN initiative aimed at creating a pan-European, multilingual corpus of historical newspapers. The Polish dataset consists of three subcorpora spanning 110 years (1830–1939). The first two components are drawn from the Microcorpus of Nineteenth-Century Polish (its short press texts and journalistic texts subcorpora), each containing 200 samples of brief news items and journalistic articles from diverse periodicals. The third component, the InterWar Corpus, covers the period 1918–1939 and comprises approximately 6.5 million words from complete newspaper issues, representing the territory of the interwar Republic of Poland. The authors argue for the scholarly value of historical press, highlighting its precise chronological dating as a key advantage for diachronic research despite challenges such as heterogeneous content and anonymous authorship. The conversion pipeline maps source metadata to a standardized TEI format and enriches texts with linguistic annotation using the Hydra NLP tool, providing lemmatization, part-of-speech tagging (mapped to Universal Dependencies), dependency parsing, and named entity recognition. The resulting openly accessible dataset enables cross-linguistic comparison and distant reading of historical press materials on a European scale.

The article focuses on a stylometric analysis of authentic literary texts and thematically related texts generated by large language models. The texts under study represent a fairly broad cross-section of twentieth-century French literature. Five models were used to generate the texts (ChatGPT 4-o, GPT 4-o mini, DeepSeek v.3, c4ai-command-r-plus, and c4ai-command-a). The original human-written stories of approximately 20,000 characters were summarized, and new narratives were then generated on the basis of these abstracts. In terms of plot and style, they were intended to resemble the originals. The research carried out with TF-IDF of the most frequent words showed that texts generated by specific LLMs and written by humans cluster relatively well as distinct groups. The experiments also showed that the “authorial” specificity of machine-generated texts partly matches the original clustering of human-written source texts.

The paper focuses on the integration of Natural Language Processing (NLP) techniques to analyze extensive chronological text corpora. This research underscores the synergy between humanistic inquiry and computational methods, especially in the processing and analysis of sequential textual data known as lexical series. A reference workflow for chronological corpus analysis is introduced, outlining the methodologies applicable to the ChronoPress corpus, a data set that encompasses 22 years of Polish press from 1945 to 1966. The study showcases the potential of this approach in uncovering cultural and historical patterns through the analysis of lexical series. The findings highlight both the challenges and opportunities present in leveraging lexical series analysis within Digital Humanities, emphasizing the necessity for advanced data filtering and anomaly detection algorithms to effectively manage the vast and intricate datasets characteristic of this field.

The subject of this article is the application of NLP and text-mining methods to the analysis of two large bibliographies: Polish one, based on the catalogs of the National Library in Warsaw, and the other German one, created by Deutsche Nationalbibliothek. The data in both collections are stored in MARC 21 format, allowing the selection of relevant fields that are used for further processing (basically author, title, and date). The volume of the Polish corpus (after filtering out non-relevant or incomplete items) includes 1.4 mln of records, and that of the German corpus 7.5 mln records. The time span of both bibliographies extends from 1801 to 2021. The aim of the study is to compare the gender distribution of book authors in Polish and German databases over more than two centuries. The proportions of male and female authors since 1801 were calculated automatically, and NLP methods such as document vector embedding based on deep BERT networks were used to extract topics from titles. The gender of the Polish authors was recognized based on the morphology of the first names, and that of the German authors based on a predefined list. The study found that the proportion of female authors has been steadily increasing both in Poland and in German countries (currently around 43%). However, the topics of women’s and men’s writings invariably remain different since 1801.