DOI: 10.18485/judig.2026.2.ch13 ISSN:

Hypothesis-Driven Corpus Querying at Scale: A Method and Its Application to - nje Nominalizations in BCMS

Predrag Kovačević, Marko Simonović, Marija Đokić Petrović
Show PDF Cite
This paper presents a reproducible method for hypothesis-driven corpus querying at scale and applies it to Bosnian/Croatian/Montenegrin/Serbian (BCMS) -nje nominalizations. Using pluralization and genitive complements as diagnostics of internal verbal structure, the study tests predictions concerning structural differences between perfective and imperfective deverbal nominals across four BCMS varieties. The paper introduces a fully automated workflow for large-scale corpus querying, enabling systematic testing of theoretically motivated grammatical hypotheses across 24,000 corpus queries. The results reveal stable asymmetries between perfective and imperfective nominalizations consistent with previous theoretical proposals, while also uncovering limited but systematic microvariation across BCMS varieties. Beyond its empirical contribution, the study provides a generalizable methodological framework for scalable and reproducible corpus-based grammatical research.

More from our Archive