• 9.1 The relevance of the models
• 9.2 Text typology for interpreting
• 9.3 References, understanding and production strategies in the interpretation of specialist discourse
• 9.3.1 Lexical problems
• 9.3.2 Interpretation strategies

9 Conclusions for future research
9.1 The relevance of the models	Our study has shown that models based on a text-linguistic analysis can give an increased understanding of what is happening during simultaneous interpreting. This applies to both the macro level (text structure and message content) and the micro level (morphology and syntax). A combination of text-linguistic and cognitive approaches seems to be especially fruitful.
9.2 Text typology for interpreting	The development of an adequate text typology for interpreting is important for educational purposes in the training of interpreters, as well as the theoretical development within interpreting research.  Such a typology would have to take into account the following factors: a) the discourse function, i.e. whether the text is, in Reiss' (1976) terms, informative, expressive, operative (argumentative, persuasive) or phatic; b) situation types, i.e. the setting, e.g. type of institution where interpreting is conducted; the number of parties involved in the interaction and their roles in society, purpose of interaction etc.; c) type of action that the situation demands, i.e. choosing the (appropriate) genre or in Foucault's terms, the text produced within a "fellowship of discourse"; cf. Swales 1990; d) textual strategy types, i.e. macrostructure, rhetorical types and "aesthetic" features; etc. that the genre requires; e) prototypical formulation matrices, i.e. conventional phrases, conventions for interaction, e.g. politeness, etc.
9.3 References, understanding and production strategies in the interpretation of specialist discourse	The study in a broader context of interpreters' understanding and production strategies of expert discourse would be highly interesting not only for translation theory, but also for the development of the theories of text linguistics and terminology, as the relation between terminology, LSP phraseology and LSP discourse has barely been the object of research.  To round up this paper, I would therefore like to briefly return to one of the most important issues in this paper and in interpreting: understanding, and discuss strategies for coping with problems of reference in technical / specialist texts[15]. To be able to interpret, the interpreter has to understand what the speaker says. But the interpreter is usually not a subject specialist. Consequently, various types of problems of understanding occur — terminological problems, and difficulties in understanding the structure of the text. In expert discourse references may not be explicit in surface structure, and coherence is often upheld only with the use of implicit references. How and when does the interpreter "begin to understand" in such a situation? An example of a terminological problem which is solved by the interpreter during the interpreting session is reported in section 7.2 above. In that example, the speaker’s term ’matrilineal’ was treated in the following way by the interpreter: 1. not translated 2. ’matrial’ 3. ’matriarchal’ In this case, the interpreter chooses not to translate the term the first time it occurs; the reason for this may be a) that she does not have time to search for the correct target language term, b) she does not know the target language term, or c) does not understand the source language term. However, as the speaker continues using the term, it becomes impossible not to translate it, so the interpreter chooses the rendering ’matrial’, (which is in fact, in the actual situation, used twice in a row). Finally, the interpreter settles for ’matriarchal’. (None of the translations is correct, but that is not the point here.) How and when does the interpreter develop ’understanding’ in these situations? According to Sørensen (1992) the main purpose and the very raison d’être of LSP (Language for Special Purposes) texts is deduced from the chain of action of which they form part. When special pragmatic-communicative problems occur, the translator must be able to refer to this chain of actions in order to make relevant decisions. The more consciously the translator is able to do this, the better the translation is. Specialist texts are often tailor-made to satisfy the chain of actions as briefly as possible; knowledge that is supposed to be known by the recipient is not rendered, explicit coherence is not upheld, and the texts abound with special terminology. An LSP text can therefore be like a coded message which can only be decoded by someone who has the necessary expert knowledge, and who possibly takes part in the same chain of actions as the message sender, or at least has a knowledge about this type of chain of actions. "Lexical cohesion", according to Halliday and Hasan (1976; cf. section 3.2.4), is achieved by selection of vocabulary, using semantically close items, e.g. general nouns, synonyms and hyperordinate concepts. Because lexical cohesion in itself carries no indication whether it is functioning cohesively or not, it always requires reference to the text, to some other lexical item to be interpreted correctly. An especially complicated form of reference in texts is "collocation", which is defined as "any pair of lexical items that stand to each other in some recognisable lexico-semantic relation" (ibid.). Such a relation is indirect, more difficult to define and based on the readers’ associations. The interpretation of such relations is based entirely on the knowledge of the subject. Relations between lexical entities may be implicit in the surface structure of the text; they are presupposed to be known by the readers/listeners ("given information"). The recipients have to make inferences on the basis of the textual context and their own subject knowledge (Lahdenmäki 1989).
9.3.1 Lexical problems	The conference interpreter has a special problem: as a mediator between subject specialists the interpreter also has to be able to make inferences in spite of his/her incomplete knowledge of the subject under discussion. But even if the interpreter does know the subject and has a good understanding of subject terminology (often, if not often enough, the interpreter has an opportunity to prepare for assignments by reading the manuscript in advance, looking up special terminology etc.), there may be problems in rendering special terms in the other language. "Knowledge transfer" from one language area to another calls for development of new terminology as well as standardisation of neologisms and harmonisation between languages. As linguistic mediators, interpreters — as well as translators — also take part in these activities. Not always is the original speaker "helpful" when it comes to neologisms. In this sample, the speaker introduces the neology (Swe.) ’okärlek’ [’unlove’] , which the interpreter easily converts to (Fin.) ’epärakkaus’ [’unlove’]. Later on, as the speaker starts to elaborate on this word, the interpreter is forced to do the same: Swedish original Translation of original Finnish interpretation Translation of interpretation förlamningen som kommer när ja upplever okärleken / (W 48) the paralysis that comes when I experience unlove lamaantumista / jonka / kokee silloin kun / tuntee tämän / epärakkauden the paralysis / which / you feel when / you experience this / unlove okärlek e ju ett s_ e ju ett ord som / man kan använda ja brukar använda de just för de som ja upplever som okärlek de vill säja rejektion / (W 49) unlove is [?]_ a word which / you can use I keep using it precisely for what I experience as unlove that is rejection /  siis tämä okärlek rakkaudettomuus / j_ okärlek sitä voi käyttää / silloin kun koen jotakin tällaista rakkauden vastakohtaa / well this okärlek lovelessness / [?]_ okärlek you can use that / when I experience this kind of opposite of love / Bold: "unlove" and its translations Figure 9-1 Lexical solutions: Unlove In the second extract the interpreter changes (Fin.) ’epärakkaus’ to ’rakkaudettomuus’ [lovelessness]. This new translation is probably compelled by the speaker’s explanation of the Swedish word. Note that in Finnish, like in Swedish and English, the use of the antonym prefix ’epä-’, ’o-’ and ’un-’, respectively, is an acceptable way of coining a new, but completely understandable word. The interpreter’s (Fin.) ’rakkaudettomuus’ [lovelessness] is an existing word, corresponding to (Swe.) ’kärlekslöshet’ [lovelessness], which is not what the speaker means. Apparently, the interpreter is aware of this, since she changes her translation to (Fin.) ’rakkauden vastakohta’ [opposite of love]. The term (Swe.) ’rejektion’ [rejection] — which is left unrendered by the interpreter — has been used by the interpreter earlier in this session, when the speaker used the verb form (Swe.) ’rejekterad’[rejected]. The words ’rejektion’ and its verb form ’rejekterad’ are probably equally unusual words in both Swedish and Finnish. ’Rejektio’ in Finnish is still acceptable as a special term, but the verb form *rejektoitu would probably be incomprehensible: Swedish original Translation of original Finnish interpretation Translation of interpretation / om ja f_ blir rejekterad å ja upplevde mej / rejekterad ... (W 31) / if I [?]_ get rejected and I felt myself rejected ...jos minut torjutaan niin minulle käy / ja / tämä oli siis rejektio siis eeh / eeh / minut torjuttiin kokonaan ... ... if I get rejected then I / and / this was a rejection that is eeh / eeh / I was rejected completely ... Bold: 'rejected' and its translations Figure 9-2 Lexical solutions: Rejection
9.3.2 Interpretation strategies	The study in a broader context of interpreters' understanding and production strategies of expert discourse would be highly rewarding. Such research would have to cover not only questions about texts per se, but also issues like norms an attitudes within the interpreter group and society at large in the use of textual strategies.  Reports on a more restricted study of strategies for solving lexical problems in the interpreting of neologisms and culture-specific terms are published in the documents "Translational creativity: Strategies for interpreting neologisms" and "The interpreter as language planner".