Modern linguistics embraces the idea that human sense making of words and other commu nicative resources is a social, dynamic activity. Inspired by the literary scholar and philos opher Michail Bachtin ( 1986), among others, language and communication researchers are increasingly applying a dialogical view of language and mind ( see, e.g. Linell 2009). Dialogic theory is also the basis for several studies of interpreter- mediated interaction. Applying a dialogical view of language and mind means, among other things, that interpreters are per ceived of as active participants, with their own agency, rather than as passive instruments. Also, the everyday perception of interpreting of spontaneous talk in interaction as a simple transfer of clear- cut messages in one language to equally unambiguous messages in an other language must be dismissed. A significant number of studies show that spontaneous interpreter- mediated interaction, spoken and/ or signed, are complex events, both linguisti cally and socially. For instance, participants may produce talk not necessarily to add content, but also to organize the very talking. Nevertheless, in a certain sense, interpreting can be described as a ‘ monologising practice in a dialogically organized world’ ( Wadensjö 2004) in that, when interpreters perform consecutively, they tend to treat participants’ talk chunk by chunk, regardless of whether they have been allocated the turn or take it on their own initia tive. When an utterance is rendered in another language, if not checked for ambiguities, it is thus treated as a more or less unambiguous chunk of talk. Doing so, interpreters are expected to be observant of linguistic details and of how talk fits in with the on- going exchange, an expectation which involves interpreters’ familiarity with the larger, institutional and cultural contexts. Within interpreting studies, indeed all authors in this volume, perceive of interpreters as active participants with their own agency. Yet, it can be argued, in line with, for example, Inghilleri ( 2012) and Määttä ( 2015), that also researchers who do not see interpreters as translation machines, but assume that they are active, sense- making participants with their own agency, paradoxically still tend to cement the simple conduit metaphor in the view of both communication and interpreters’ work, by their exclusive focus on interpreters’ choice of words and expressions in the other language, that is, looking at ( and evaluating) interpret ing exclusively on the basis of the ‘ monologizing’ component of their activity, as one would look at source texts in relation to target texts. What is ‘ the conduit metaphor’? Interpreting, particularly in public service, has not been of interest exclusively to linguists and interpreting/ translation scholars. In articles in Medicine and Law journals, one some times finds ‘ the ( generic) interpreter’ described in terms of a tool that, if only maneuvered correctly by the medical or legal expert, will work so as to not affect the case in question in any unpredictable way. This description may be motivated by a collegial request, that medical and legal professionals take responsibility for handling encounters with patients or clients with whom they do not share a language, rather than handing over what might seem in their view ‘ too much’ responsibility to the interpreter. Nevertheless, the tool metaphor for the interpreter is based on a simple but unrealistic idea that interpreters are, or can be, some kind of non- subjects, or translation machines, programmed to switch between languages in some automatized way. Moreover, this view of interpreters is based on a simplistic view of language and human communication, what linguists usually refer to as the conduit metaphor ( Reddy 1979). The conduit metaphor becomes visible in speech and writing in phrases such as ‘ his words expressed strong feelings’, and ‘ I took my ideas from her text’. In meta- language about language and communication, words and phrases appear as a kind of container, into which the speaker or the writer puts thoughts, feelings and meanings, exactly the same thoughts, feelings and meanings that listeners and readers are expected to eventually be able to re trieve. The conduit metaphor ignores that all individuals are interpreting subjects, who cre ate meaning of words and expressions based on their respective knowledge, experiences and expectations; in other words, that human communication is seldom completely predictable. At the same time, the conduit metaphor is firmly established in our everyday thinking and talking about words. The idea of words as containers of meaning is perhaps so familiar that we, in everyday life, even have a hard time to think about language and communication in a different way. But applied in the area of interpreting, the conduit metaphor, in Reddy’s ( 1979) sense, can be quite misleading, for various reasons. In studies of interpreting, the conduit metaphor has sometimes been used also to depict interpreters, based again on a perception of words as containers of meaning, which inter preters allegedly have the task to retrieve and then load onto other containers, that is, words in another language. Again, interpreting is depicted as something individuals do in relation only to words that ‘ have’ a meaning glued onto them, as it were, rather than in relation to other human beings in a social context. In her groundbreaking book on sign language inter preting, Roy ( 2000: 101– 4), also drawing on Reddy ( 1979), notes that the conduit metaphor implies a number of basic assumptions about language, and simultaneously demonstrates that, in the professionalization process of sign language interpreters, the conduit metaphor in certain ways responded to the interpreters’ need to identify less as helpers and more as just this – professionals. Strongly related to the conduit metaphor in meta- language about language and com munication is the everyday perception of ( good) interpreters as performing word- by- word translation. In their review of interpreting practices within the Australian legal system, Laster and Taylor ( 1994) shed light on what appear to be conflicting attitudes between law yers and interpreters as to how interpreting can and should take place. They believe that the professions’ different perceptions and knowledge of word- for- word interpreting in them selves constitute a breeding ground for mistrust and believe that this mistrust would rather increase if interpreters were to sign an unrealistic demand for word- for- word interpreting. Instead of thinking of the interpreter in terms of a channel or a machine, they suggest that the court should consider the interpreter as a communication promoter and point to the need to build mutual respect for how representatives of each profession exercise their professional discretion, that is, how the actors take decisions founded in well- established professional knowledge and professional ethics ( Laster and Taylor 1994: 126). This argument, published already in the 1990s, is no less valid today. In this volume, interpreters’ mandate and obliga tions in terms of professional discretion are discussed by Norwegian scholar Hanne Skaaden ( Chapter 16) Definitions of ‘ interpreting’ Research on interpreting can roughly be divided into studies focusing on the interpreting individual and/ or this individual’s cognitive processes on the one hand, and, on the other hand, on interpreting as linguistic, communicative and social interaction. These two cat egories of studies are characterized by their different ways of delimiting interpreting as a research object. In the former case, a definition of interpreting similar to that proposed by Franz Pöchhacker is applied: Interpreting is a form of Translation in which a first and final rendition in another language is produced on the basis of a one- time presentation of an utterance in a source language ( Pöchhacker 2016: 11, bold in the original). In Pöchhacker’s definition, translation is written with a capital T and refers to translation in a general sense. The definition emphasizes immediacy and volatility as characteristic features of interpreting that distinguish it from translation of written texts, but the orality of interpreting is not specifically highlighted, which makes this definition possible to apply also to explorations of signed language interpreting. Following this definition, the inter preting individual, or the cognitive operations involved when producing renditions, are the primary objects of study. In this definition, interpreting means producing renditions of previously spoken utterances in a source language. This implies that those speaking with the assistance of interpreters basically are perceived of as producing utterances to be rendered in a target language. In other words, this definition tends to disregard participants’ other communicative projects ( such as organizing the interaction, searching for words, provid ing minimal responses like ‘ mm’ and ‘ aha’ and so forth), as well as situations and contexts. Seemingly, this definition of interpreting has grown out of an interest in simultaneous interpreting, where the performing interpreter is working with prepared ( often in written form) talk, rather than in the conditions of spontaneous spoken interaction. No doubt, to interpret written speeches places different demands than to interpret spontaneous spoken exchanges. Interpreting can also be defined as interaction, as suggested by Wadensjö ( 1992; 1998). With such an approach, the primary object of study is the interpreter- mediated situation. In order to explore situations of this kind, Wadensjö ( 1998) suggests to apply sociologist Erving Goffman’s ( 1981) notion of situated system of activity. In other words, in order to fully under stand participants’ behaviour in a social encounter of some kind, the researcher considers not just their talk, but also the social, cultural and institutional frames which this situa tion entails. Defining interpreting as interaction allows for studying interpreters’ utterances not only on the basis of their function as translations, as source texts in relation to target texts but also on the basis of their coordinating function in the situated activity.
