Visualising Overlooked Syntactic Alternatives in the Greek New Testament

Paper presented at the IOSOT 2025 conference in the section Digital Humanities and Computational Approaches to the Bible. This paper explores how overlooked syntactic alternatives in the Greek New Testament can be visualized by extending the N1904-TF dataset. By integrating Morpheus-derived morphological analyses through a custom pipeline, multiple possible parses for each word become visible, showing how ambiguity at the morphological level can shift phrase functions and reveal alternative readings often hidden in disambiguated datasets. While this method enriches grammatical exploration, it also exposes inherent tensions: N1904-TF assigns one disambiguated morphosyntactic tag per token, whereas the Morpheus data records all legitimate possible renderings, including those resulting from underspecification. The principles of contextual adjudication versus recording underspecification therefore result in incompatibility when performing comparisons. The proposed way forward is to codify underspecification (distinguishing all possible values from the chosen contextual reading) and to expand the training data with corpora like the Septuagint, paving the way for more precise and transparent computational study of the New Testament