Gespystem

Gespystem is a term used in linguistics and digital humanities to describe a framework for encoding spelling variation across languages and historical periods within a unified data model. It treats orthography as a layer separate from phonology and meaning, enabling systematic representation of multiple spellings for the same word and the mapping of those spellings to phonemic values and to semantic entries. The goal is to support corpus research, lexicography, and textual editing.

The core model comprises several layers: a graphemic layer mapping each visible character or digraph to a grapheme node; a phonological layer linking graphemes to one or more phonemes; a variant layer recording historical spellings and their contexts; and a metadata layer that tracks language, date range, source, and provenance. Relationships among layers are defined to support queries such as "find all spellings that map to a given phoneme in 15th-century texts."

Implementation typically leverages existing standards such as TEI for textual encoding, Unicode for characters, and RDF for interlinking concepts. Gespystem data can be stored in graph databases or structured XML/JSON, enabling interoperability and advanced search, variant rendering, and OCR post-correction workflows. The approach is used in digitization projects, historical lexicography, and cross-language comparison.

Limitations include increased data complexity, the need for detailed metadata, and uneven adoption across projects. Critics argue it may over-model spelling at the expense of readability for editors. Proponents contend that it clarifies orthographic variation and improves reproducibility in digitized corpora. See also: orthography, grapheme, TEI, OCR correction.