"Well, uh no. I mean, you know". Discourse Markers in Movie Conversation.

Film translation is a challenging and creative process which requires the ability to decode the linguistic and cultural features of the source text and to transfer them appropriately and effectively into the target text. Differences between the Source Language/Culture and Target Language/Culture require translators to compensate for gaps in shared knowledge or linguistic features in order to achieve successful communication. Spoken discourse presents specific linguistic features (e.g. interjections, backchannels, attention signals, repetitions, reformulations, hesitators, discourse markers, vocatives, inter alia, cf. Halliday 1985, Biber et al. 1999, McCarthy 2003) which may not be readily transferred across languages. The present paper investigates whether the characteristics of spontaneous conversation are also to be found in movie language, an instance of non-spontaneous, prefabricated speech which is “written to be spoken as if it were not written” (cf. Gregory 1967 and Nencioni 1976) in order to sound authentic and if so, how they are translated in the dubbed version. In particular, the present paper focuses on the discourse markers you know and I mean. Their frequency of occurrence, semantics and pragmatics are investigated in a corpus of transcripts of dialogues from contemporary American movies directed from 2000 on and then compared to those found in the Italian dubbed versions of the same movies. The aim is to highlight the functional, rather than semantic, nature of discourse markers (cf. Bazzanella and Morra 2000) and to identify the strategies activated in the translation process to achieve an equivalent effect in the light of the constraints of the visual channel. A further aim is to verify whether the translation options provide evidence of universal features, such as simplification or normalization and leveling out, or whether they are influenced by cross-contact between the languages involved (cf. Baker 1998, Ulrych 1998). The analyses are corpus-driven (cf. Francis 1993, Tognini-Bonelli 2001) in that the theory is built up in the presence of the evidence found in the US spoken sub-corpus of the Bank of English and in the corpus of transcripts of American movies and their dubbed Italian versions.