Stockholm University, 1999
The project supervisor was Professor Erling Wande. The interpretation corpus was recorded by Birgitta Romppanen, who also participated in the initial stage of the project, and the transcription of the tapes was done by Miriam Sissala. The final analyses and the compilation of the report was made by Helge Niska.
Comments, criticism and suggestions for improvement can be sent to Helge.Niska@tolk.su.se. Address: Institute for Interpretation and Translation Studies, Stockholm University, S-106 91 Stockholm. Tel. +46 8 162000. Fax: +46 8 161396.
|In the activity of interpreting, two aspects are especially significant:
orality and interaction. The interpreter translates oral discourse (which
may or may not have been prepared as written text) in various communicative
situations, where messages are exchanged, through the interpreter, between
Text linguistics has emerged from the 1970s and onwards as the study of the property of texts (written or oral) and their uses in communicative interaction. The modern conception of text linguistics is a broad one, encompassing discourse analysis and pragmatics, as well as influences from cognitive sciences, communication studies and artificial intelligence. Written text is still the usual object of study within text linguistics, and monographs and scholarly papers abound with examples of written texts. But underneath, or behind, the written text, there are cognitive and other kinds of processes that can take the form of either written or spoken or signed discourse.
In working on this study, the contemporary text linguistic approaches of de Beaugrande & Dressler (1981) and van Dijk & Kintsch (1978, 1983) have been especially inspiring. Two other researchers, who are not text linguists, but who take utterances and whole texts as their point of departure in describing the interpreting process, have likewise been very influential, namely Chernov (1978, 1979, 1985, 1994) and Alexieva (1985, 1988). Of these two, Alexieva is probably the most text-linguistically oriented, while Chernov relies more heavily on psychological, especially cognitive research. Nevertheless, we feel that their models, apart from being highly relevant as descriptions of the interpreting process, in this context serve both as valuable complements and correctives of the more "explicitly" text linguistic model proposed by Kintsch and van Dijk.
|Internationally speaking, simultaneous interpreting is a relatively
new research field, and in Sweden there is virtually no empirical or theoretical
research in this area. The aim
of this project was to make a preliminary study of the process of simultaneous
interpreting, as a pilot study. The objectives were both to assess some
of the text linguistic models for the description of the process of simultaneous
interpreting that had been presented in previous research, and to test
a hypothesis as to the existence, in the simultaneous interpreting situation,
of a special variant of translation, for which we coined the term "translatorese".
In order to do this, analyses of the interpretations had to be done on
several linguistic levels. As a result of this, we also expected to get
a preliminary inventory of linguistically and theoretically interesting
topics for future research.
The languages involved are Finnish and Swedish. This is interesting from a typological point of view, as the bulk of all research on simultaneous interpreting so far has been conducted on Indo-European languages.
The pilot project nature of this project implies that another of our primary aims was, on the basis of specified theoretically-founded studies of samples from the empirical material, to develop a better point of departure for larger future projects in the area.
|The material for this study was collected at two conferences in Finland
in the autumn of 1990. Altogether, it consists of about 15 hours of audio
recordings from two conferences. Within the time frame of the pilot study,
we were only able to transcribe approximately five hours from one of the
conferences, the participants of which were female authors from Sweden
and Finland. From this conference we have analysed speeches delivered by
government officials and professional authors, and the interpreting conducted
by three, likewise female, professional conference interpreters.
The audio recordings were copied to four-channel audio tapes where the original speaker input occupies one channel pair, and the interpretation the other pair. In this way it is possible to listen to the speaker and the interpreter either separately or concurrently. The analyses have been made mainly on the basis of the transcribed material, and the original recordings have been used for control purposes.
|The transcription of the material was done by PhD student Miriam Sissala.
In the transcriptions, only lower case letters have been used, and there
is no punctuation. The transcription is orthographic, and spoken language
forms have been used extensively, if not exclusively; e.g. ’ja’ for Swedish
’jag’ [I], ’å’ for ’och’ [and], ’med’ for ’med’ [with], ’hitta’ for
’hittade’ [found], ’e’ for ’är’ [is, are].
Because of technical limitations, there is no measurement of time, e.g., length of utterances, duration of pauses or lags between original speech and interpretation. The transcriptions have, however, been marked for pauses within the respective utterances, where a single slash ’/’ denotes a short pause, and double slash ’//’ denotes a long pause. False starts and mispronunciations are recorded in the transcripts. The transcription of both originals and interpretations are printed in two (roughly) synchronised parallel columns. For the sake of clarity, in the samples printed in this report translation into English of the original and the interpretation has been added. We have aimed at using spoken language forms in the translations as well.
The code within parenthesis at the bottom of the first column of the samples (T 20) shows the location of the extract in the transcribed corpus.
Figure 1-1 Sample transcription with English translation
|The basic method in our study was to compare, on the basis of the transcripts,
the output of the original speaker to that of the interpreters. Although
the text-linguistic approach was, obviously, the most important one, it
was natural to make analyses on both "micro" and "macro" levels.
On the "micro level" we took note of anomalous intonation and mispronunciations, grammatical "errors" and possible interference from the source language in the target language, like changes in word order and other syntactic changes, changes in use of pronouns, lexical changes and mistranslation. The analyses on this level were made on short discourse segments, the equivalents of "sentences" in written texts. This is also a kind of analysis that we were used to do when assessing the performance of community interpreters in university examinations and state accreditation tests.
Already at this stage it was possible to match the findings of our analyses to the models that we were about to assess. Both Alexieva’s and Chernov’s models are well suited to cater for the phenomena that we noticed at this level (cf. section 6.3 for Alexieva’s four stage simultaneous interpreting model).
The analyses on the "macro" level had to do with issues like the interpreters’ "editing" of the texts, e.g. changes in the order of subtopics, their handling of special terminology, and strategies for coping with cognitive problems.
Since the aim of the study was to assess the applicability of text-linguistic models to the study of simultaneous interpreting, the main task for us was to try to apply the models to samples from the transcribed material. The results of these "tests" will be reported in the respective sections below.
|Traditional linguistics sees the sentence or possibly clauses within
sentences as its basic unit of research. Even in text linguistics, which
supposedly works with larger units, the sentence, or "the orthographic
unit that is contained between full stops" (Halliday 1985:193), is often
the unit quoted in examples etc. This is natural when dealing with written
texts, but when working with a spoken corpus like ours, things are more
complicated. In a forthcoming paper, Robert de Beaugrande proposes a new
way of defining the sentence (de Beaugrande forthc.):
Our prime question would then be: which sets of criteria might be relevant for making (or not making) a given stretch of discourse into a sentence, or for recognising it to be a sentence? At least the following sets of criteria might be considered:
This page was last updated on April 1, 1999
Please send comments or questions to Helge.Niska@tolk.su.se.