Visualising Overlooked Morphosyntactic Alternatives - A Morpheus-Based Extension for the Nestle 1904 Text-Fabric Dataset

This thesis augments the Nestle 1904 Greek New Testament Text-Fabric dataset (N1904-TF) by systematically incorporating every plausible morphological analysis for each word token, as generated by the Morpheus analyser for Ancient Greek. Rather than adjudicating between parses, the project makes ambiguity explicit, structured, and accessible through a reproducible processing pipeline built on a custom Python package (Morphkit). The extension (N1904addons) adds over one thousand new Text-Fabric features across three levels: word-level diagnostics, lemma-grouped variants, and detailed Morpheus parses. A similarity metric ranks alternatives against the original annotation, while entropy-based measures and phrase-level models capture how morphological ambiguity affects interpretation. Fully integrated into the Text-Fabric environment, this open-source extension enables efficient querying, visualization, and analysis of morphosyntactic alternatives. Demonstrated through empirical use cases, it advances transparent grammatical study and supports both theological and linguistic research by making ambiguity visible and measurable.