Realwords

Realwords is a term used in linguistics and language technology to denote strings that have established lexical status in a language’s standard resources, such as dictionaries and large corpora. A Realword is typically a word with a defined meaning, conventional pronunciation, and a recognized part of speech within the language community. The concept covers base forms as well as inflected or derived forms that are attested in usage, such as run, running, and runner.

Realwords are contrasted with non-words or pseudo-words, which are pronounceable strings lacking recognized meaning or lexical status. In research and applications, Realwords are often distinguished from nonce words or invented terms used in experiments or training data. In natural language processing, identifying Realwords aids tasks such as spell checking, lexical disambiguation, and the evaluation of language models, where authentic vocabulary is expected to appear. The status of a Realword can influence how a system handles morphology, syntax, and repair in text generation or analysis.

Definitions of what counts as a Realword can vary by language, region, and dictionary edition. Some terms may be dialectal, archaic, or technical, and their status can change over time as language evolves. Brand names and certain proper nouns may receive lexical status in dictionaries but are treated differently in some computational tasks, depending on context and application.

See also:

- Lexicography

- Lexicon

- Pseudo-word

- Neologism

- Language model

- Corpus linguistics