The paper portrays a linguistic and textual analysis of a corpus of 18 museum audio descriptions (ADs) (a fairly neglected area of accessibility and audiovisual translation research) in order to pinpoint the salient features of this text type and relate them to current AD literature and guidelines. Results show that scripted and recorded museum AD texts comply with recommendations only partially. They guarantee vivid, imaginative and diverse language as well as substantial text informativity through the combination of high lexical diversity and the extensive use of descriptive adjectives as well as substantial lexical density. In spite of the use of short words, however, museum ADs seem more lexically and syntactically complex than expected, with their use of opaque technical terms, heavy adjectival phrases and long sentences. More systematic and contrastive research will help to (dis)confirm these results, whereas audience reception research will contribute to determine the real degree of usability of these new text types.
