speltthat

Speltthat is a proposed concept in digital humanities and historical linguistics that describes a systematic practice of annotating spelling variants within text corpora. The idea is to attach to each token a record of its historical spellings, dialect forms, and etymological notes, so researchers can study orthographic variation without losing the modern normalized form. The term combines "spelt" (the past tense of spell) and "that" as a shorthand for "that spelling history." The coinage is not tied to a single organization and has appeared in discussions of data models for diachronic corpora.

Core features include a standard metadata schema per token, fields such as standard form, variant spellings, time frame, geographic region, source documents, and confidence level; compatibility with TEI-encoded texts and JSON/XML representations; and support for queries that bridge spelling variants with meanings or lemmas.

Applications include historical text digitization, OCR post-correction workflows, diachronic search, and editorial projects that seek to preserve philological information in digital editions. Benefits include improved traceability of orthographic change and more accurate linguistic analyses across periods and dialects; drawbacks include the need for agreed standards and potential data bloat.

See also orthography, historical linguistics, digital humanities, TEI, diachronic corpus.