Gregorian melody, modality, and memory: Segmenting chant with Bayesian nonparametrics

2025-06-30

AI Summary

This study investigates whether Gregorian chant melodies are constructed from pre-existing melodic fragments (“collage theory”), examining the interplay among unsupervised segmentation, modal classification, and mnemonic efficiency. Method: We propose a Bayesian nonparametric model based on a nested hierarchical Pitman–Yor process that jointly models melodic sequences and drives adaptive segmentation; crucially, it explicitly incorporates monastic memory mechanisms into musical structure learning—a first in computational musicology. Contribution/Results: (1) Optimal segmentation does not support traditional collage theory but significantly improves modal classification accuracy—achieving state-of-the-art performance. (2) Initial and final segments exhibit high formulaicity, corroborating mnemonic compression strategies. (3) Segmentation boundaries strongly align with empirically observed performative phrase structures, revealing that melodic organization balances memorability and modal discriminability. These findings establish a novel paradigm for modeling medieval musical cognition and advance computational musicology.

Technology Category

• Machine Learning: Multimodal Learning
• Cognitive Modeling & Cognitive Systems: Computational Creativity
• Computer Vision: Segmentation

Application Category

• Web Mining and Content Analysis: Models for Web evolution
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs

Abstract

The idea that Gregorian melodies are constructed from some vocabulary of segments has long been a part of chant scholarship. This so-called "centonisation" theory has received much musicological criticism, but frequent re-use of certain melodic segments has been observed in chant melodies, and the intractable number of possible segmentations allowed the option that some undiscovered segmentation exists that will yet prove the value of centonisation, and recent empirical results have shown that segmentations can outperform music-theoretical features in mode classification. Inspired by the fact that Gregorian chant was memorised, we search for an optimal unsupervised segmentation of chant melody using nested hierarchical Pitman-Yor language models. The segmentation we find achieves state-of-the-art performance in mode classification. Modeling a monk memorising the melodies from one liturgical manuscript, we then find empirical evidence for the link between mode classification and memory efficiency, and observe more formulaic areas at the beginnings and ends of melodies corresponding to the practical role of modality in performance. However, the resulting segmentations themselves indicate that even such a memory-optimal segmentation is not what is understood as centonisation.

Problem

• Identify optimal segmentation of Gregorian chant melodies
• Explore link between mode classification and memory efficiency
• Evaluate centonisation theory using Bayesian nonparametric models

Innovation

• Bayesian nonparametrics for chant segmentation
• Pitman-Yor models optimize memory efficiency
• Segmentation enhances mode classification accuracy