Audio description (AD) improves media access for visually impaired people and can therefore make an important contribution to social inclusion. What is essential in the development of effective descriptions is that they enable visually impaired audiences to create a coherent understanding of the story underlying the described audiovisual material. One important aspect that contributes to the success (or failure) of this are the ways in which characters, objects, actions and events that are part of the story are introduced and referred to in the descriptions. The UK's guidance on AD suggests in this respect that "[i]f something is identified by name or has already made an appearance, the definite article 'the' is used. If the subject or object is new, the indefinite 'a' is preferable" (OfCom 2006: 12). The reality of text production is, however, often more complex than that.
According to the taxonomy of referential identification suggested by Prince (1981) and refined by Brown and Yule (1983), treating a verbal referent as ‘new’ or as ‘known’ (through prior introduction) are only two of several options. There are also cases in which referents are presented as ‘given’ through activated world knowledge, ‘inferable’ from the context or ‘salient’ in the physical environment. Definite expressions are used in most of these cases, although the appropriateness of a particular expression (e.g., a pronoun) is subject to sensitive constraints and depends on the saliency of a referent in the discourse world (see, for example, Givón 1995).
The problem the audio describer faces is that visual communication has no direct counterpart for the verbal options of referential identification. In line with the general differences between ‘showing’ and ‘telling’ (Toolan 2001), the visual devices of indicating the status of a discourse entity seem to be more subtle and less rule-governed than the devices available in verbal communication (especially the distinction between indefinite and definite expressions and pronominalisation).
This has several implications for AD. On the one hand, an audio describer needs to make (subjective) decisions about the treatment of referents in the AD text, based on his/her judgement regarding a referent’s saliency and taking into account the known rules and constraints for the verbal identification of referents. On the other hand, the interplay between the AD text and the dialogue in a film or theatre play raises new questions – and may entail new rules – for referential identification in AD.
This paper will discuss some of the problems of identifying referents in AD. It will begin by outlining relevant frameworks of referential identification in verbal communication, then compare these to the situation in visual communication, and finally examine and exemplify problems in, and solutions for, AD.

References
Brown, G. & G. Yule (1983). Discourse Analysis. Cambridge: Cambridge University Press.
Givón, T. (1995): Coherence in text vs. coherence in mind. In Gernsbacher, M. & T. Givón (eds.), Coherence in spontaneous text. Amsterdam: Benjamins, 59-116.
Prince, E. (1981). Toward a taxonomy of given-new information. In P. Cole (ed.), Radical Pragmatics. New York: Academic Press. 223-56.
OFCOM (2006). ITC Guidance on Standards for Audiodescription. http://www.ofcom.org.uk/ tv/ifi/guidance/tv_access_serv/archive/audio_description_stnds/ (accessed 9 February 2009).
Toolan, M. (2001) Narrative. A critical linguistic introduction. London: Routledge.

Sabine BRAUN is senior lecturer in Translation Studies at the University of Surrey (UK). She holds an MA in Translation Studies from the University of Heidelberg and a PhD in Applied English Linguistics from the University of Tübingen (Germany). Her main interest is in the broad area of ‘communication and the new media’ and its pragmatic and discourse-based foundations. Her research areas include new forms of interpreting (videoconference/remote interpreting) and the use of video-based multimedia corpora in language and interpreter training. Her interest in audiovisual media has also led her to conduct research into audio description (AD). She is particularly interested in the discourse comprehension and production processes involved in AD as a complex form of intermodal mediation. She is the author of Kommunikation unter widrigen Umständen? Einsprachige und gedolmetschte Kommunikation in der Videokonferenz [Communication under adverse conditions: monolingual and interpreted videoconference communication] (Tübingen: Narr, 2004), co-editor of Corpus Technology and Language Pedagogy: New Resources, New Tools, New Methods (Frankfurt: Lang, 2006) and has developed a video-based corpus of spoken professional English (ELISA, http://www.corpora4learning.net/elisa). She was the organiser of the international research seminar "Audiodescription for visually impaired people: Towards an interdisciplinary research agenda" (Surrey, June 2008).