Measuring what corpus analysis has overlooked: Three original indices for ideological density, restricted agency, and deontic load in pedagogical corpora

Corpus-assisted discourse studies lacks dedicated quantitative measures for three discourse phenomena that operate below the level of isolated keywords: lexical concentration, agentive restriction, and deontic regulation. This paper presents three original corpus-based indices calibrated on a 263,878-token multi-period corpus of official Ecuadorian school textbooks (2016–2026): the Ideological Lexical Density (DLI), the Restricted Agency Index (IAR), and the Deontic Load Index (ICD). A fourth measure, the Pair Ratio (Ratio_par), extends Kilgarriff's keyness framework to systematic lexical co-occurrence. Each index is formally defined, mathematically grounded in Systemic Functional Grammar and corpus statistics, calibrated against existing keyness measures (log-likelihood GZ, percentage difference %DIFF, and Ratio), and validated across three curricular sub-corpora. Results demonstrate that DLI, IAR, and ICD detect grammatical and distributional patterns invisible to frequency-based measures alone, providing reproducible, transferable instruments for discourse analysis across institutional, political, legal, media, and organisational corpora. The measures are implemented in BUNKER EGAS v15.0, an open corpus-assisted linguistic auditing system. Although calibrated on pedagogical Spanish, the formal procedures are designed to be portable and recalibrable to any corpus of institutionally produced language.