Making Characters Count. A Computational Approach to Scribal Profiling in 14th-Century Middle Dutch Manuscripts from the Carthusian Monastery of Herne

2025-08-26

AI Summary

This study addresses the challenge of distinguishing the scribal hands of 13 anonymous scribes in 14th-century manuscripts from the Herne Abbey scriptorium and critically evaluating traditional paleographic attributions. Methodologically, it constructs a diplomatic-text corpus derived from HTR (Handwritten Text Recognition) transcriptions and introduces “abbreviation density” as a novel quantitative metric, integrating bag-of-characters modeling with abbreviation-symbol features to build a character-level computational scribe profiling model. Its primary contribution lies in the first systematic quantification of abbreviation practices as a stylistic dimension, thereby overcoming key limitations of conventional paleographic and graphometric analysis. Results successfully differentiate individual scribal profiles, revise the traditional attribution of scribe α’s role in the Vienna manuscript, and uncover previously unrecognized cross-manuscript collaborative relationships—demonstrating that fine-grained character-level features effectively reconstruct manuscript production networks and transmission pathways.

Technology Category

• Natural Language Processing: Sentiment Analysis, Stylistic Analysis, and Argument Mining
• Data Mining & Knowledge Management: Knowledge Acquisition from the Web
• Application Domains: Humanities & Computational Social Science

Application Category

• Web Mining and Content Analysis: Models for Web evolution
• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs
• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics

Abstract

The Carthusian monastery of Herne was exceptionally prolific in producing high-quality manuscripts during the late 14th century. Although the scribes remain anonymous, previous research has distinguished thirteen different scribal hands based on paleography and codicology. In this study, we revisit this hypothesis through the lens of linguistic characteristics of the texts, using computational methods from the field of scribal profiling. Using a newly created corpus of diplomatic and HTR-based transcriptions, we analyze abbreviation practices across the Herne scribes and demonstrate that abbreviation density provides a distinctive metric for differentiating scribal hands. In combination with a stylometric bag-of-characters model with brevigraph features, this approach corroborates and refines earlier hypotheses about scribal attribution, including evidence that challenges the role of scribe $α$ in Vienna, ÖNB, SN 65. Our results highlight the value of combining computational stylometry with traditional codicology, showing how even the smallest elements of the written system -- characters and abbreviations -- can reveal patterns of scribal identity, collaboration, and manuscript transmission.

Problem

• Identifying anonymous scribes through computational linguistic analysis
• Differentiating scribal hands using abbreviation density metrics
• Refining scribal attribution hypotheses with computational stylometry methods

Innovation

• Computational scribal profiling using abbreviation density
• Bag-of-characters model with brevigraph features
• Combining computational stylometry with traditional codicology