sourceterm

Sourceterm is a term used in information science, linguistics, and localization to denote the original word, phrase, or identifier that identifies an item in its source language or source dataset. In translation and terminology workflows, the sourceterm represents the term as it appears in the source material, guiding the creation of equivalent expressions in target languages.

In lexicography and terminology management, the sourceterm often serves as the canonical form or headword from which variants, inflections, and translations are derived. In knowledge graphs and data provenance, sourceterms can function as primary labels or keys that link data items to their origins, supporting traceability and reproducibility.

Characteristics of sourceterms include that they are not universally standardized and can encompass single words, multiword phrases, inflected forms, or terms written in non-Latin scripts. They frequently require normalization when used across languages or systems. Ambiguities can arise when a sourceterm has multiple senses or when distinct domains share identical expressions, necessitating context or metadata for disambiguation.

Sourceterms are commonly contrasted with target terms (the translated or adapted equivalents), lemmas or canonical forms, and glosses or annotations. In practice, terminology databases capture sourceterms with language tags and source references to support multilingual indexing and retrieval. Examples illustrate how a sourceterm in English may map to different target terms in another language depending on sense, such as “bank” relating to a financial institution versus a riverbank.

See also: target term, lemma, gloss, terminology management, translation memory.