Evidence from languages other than English has further proven that existing definitions of the turn need to be modified before being applied to cross- or inter-cultural data. In his observations in the Antiguan speech community, Reisman (1974) found that unlike the case with speakers of English overlapping speech is neither cut off nor 'repaired'. Rather, the current speaker continues speaking when another participant starts her utterance. In her study of conversations between New Yorkers and Californians, Tannen (1984) also identified cross-cultural misunderstandings regarding those overlaps that have usually been termed turn-taking signals. New Yorkers obviously rather use overlaps to support the speaking partner by uttering short questions, while Californians take them for turn-claiming signals. Kotthoff (1993) therefore states that the status of an overlap as either being a support or rather an interruption needs to be negotiated during the course of the interaction. This, again, implies that a static, technical definition focussing on an isolated turn must necessarily fail.
Szatrowski (1993), who analysed the structure of Japanese conversation, also encountered problems with the concept of the turn. She claims that, different from English conversations in which for the most part one participant talks one-sidedly (in turn), it is a destinctive feature of Japanese conversations that one participant offers a piece of information and the other participant cooperatively completes a unit she calls wadan (information unit). The concept of wadan resembles that of the turn in its functional definition. A wadan has been completed when a semantic contents, which carries the conversation further, has cooperatively been negotiated. Within one wadan, participants exchange sense units, which may be sentences, phrases or single words.