Think-aloud protocol studies into translation: An annotated bibliography

The article outlines a pedagogically oriented research project to study translation processes from German into Portuguese. The pool of subjects would comprise a total of 24 subjects (12 from Portugal and 12 from Brazil). Each of the national groups would include professional translators and novice and mature students.

An overview article which, in addition to summing up the findings, suggests e.g. that more attention be paid to routineness, professionalism, translation into L2, as well as methodology in TAP research.

Preliminary results from a study with five students of translation (two with 2–3 years of work experience) translating excerpts of news texts from English into French. The analysis highlights three processes and focuses on (mis)comprehension (i.e. translation errors).

The article discusses text comprehension in translation by analysing the videotaped TAPs of three students of translation (with varying degrees of professional experience) and comparing the students’ comprehension strategies with their translation products. [ p. 113 ]The aims are both theoretical and pedagogical, i.e. to develop a small-scale model of text comprehension in a given situation and to increase the students’ awareness of their own processing.

The article discusses different approaches to process analysis from an epistemological point of view. The focus is on theoretical considerations, but concrete examples (from e.g. Dancette 1994) are included.

A theoretical contribution discussing the use of verbal reports as data on language processing. The author draws attention to the fact that methodological groundwork done by cognitive psychologists (see note 4) tends to ignore research evidence from language research. TAP research on translation (cf. Dechert and Sandrock 1986) is one of the language tasks used as examples.

The article discusses the main findings of the study which is reported in full in Sandrock (1982).

The article gives an outline of a research project which aims at investigating the semantic changes, particularly explicitation, which occur in translation from Russian into Swedish. The first stage of the study includes an analysis of (1) published translations and (2) interpreting performance (by professionals and students). The second stage will include a TAP experiment with professional and student translators translating from Russian into Swedish. A computer logging system will also be used.

The article reports on a study in which 12 community translators translated a leaflet in English into ethnic minority languages. The process data was collected by using [ p. 114 ]immediate retrospection. The article discusses the contribution of translation theory to community translating, identifies the strategies used by community translators, and evaluates the usefulness of the method. The community translators seemed to adopt a socio-cultural and reader-oriented approach to the translation task, i.e. their aim was to ensure the readers’ access to British society. The need for ‘user education’ for those using translation services is also discussed.

A pedagogically oriented study reporting on the salient process features identified in the TAPs produced by 21 freelance translators translating a journalistic text from French into English. The TAPs reveal e.g. the importance of the brief in translational decisionmaking; the role of context in determining meaning; the translators’ habits with using dictionaries; and the translators’ self-image. The author argues for a translation pedagogy based on professional practice.

An overview article which focuses on research with professional translators as subjects. The author discusses the limitations of the method as well as the contribution which TAP studies could make to translation theory, translator training, and translation practice. More research on professional translating is called for.

The article opens with a criticism of traditional translation theory which seems to offer little for the purposes of teaching translation. The author then describes features of professional practice identified in earlier research (Fraser 1993, Fraser 1994), such as how the brief guides the process, how professionals deal with cultural concepts, and how they see the translator’s role. A new approach to teaching translation is proposed.

An overview article based mainly on some of the papers presented at the AILA World Congress in Jyväskylä in 1996 and subsequently published in Tirkkonen-Condit and Jääskeläinen (2000).

[ p. 115 ]A pilot study on the translation processes of university students of French translating a text excerpt from French into English. The main aim is to develop methodological tools. Two categories of analysis are presented: one of identifying the unit of analysis and another for identifying subjects’ text processing strategies. Later developments are reported in Gerloff (1988).

A pilot study with focus on the development of methods for data analysis. TAP data from five intermediate-level students of French and one competent bilingual (English/French) are used to create a method for identifying the unit of analysis in translating. The study also illustrates the strategic use of repetition in problem-solving.

The subjects represent three levels of professional competence: novices (4), competent bilinguals (4), and professional translators (4). The process data are analysed in terms of external processing (time spent) and internal processing, e.g. units of analysis and “problem solving activities” the quality of the translations was also assessed. The findings show that professionals did not necessarily produce the best quality translations (cf. Jääskeläinen 1990b). Instead, quality seemed to be the outcome of the amount of time and effort invested in the task: the processes leading to the best results took more time and required more processing effort than the less successful processes. The author terms her finding as the “translation-does-not-get-easier” phenomenon (see Introduction). See also Jääskeläinen 1996a.

The book contains nine articles which report on the TRAP (“Translation Process”) project at the Copenhagen Business School. In this project the translation process has been approached from different angles and different language combinations. The researchers have used a combination of methods: computer logs created with the Translog program (Jakobsen and Schou 1999), retrospective and think-aloud protocols. The authors argue for triangulation methods in translation research. The book includes contributions by Jakobsen about using Translog, Lorenzo (1999a, 1999b) about methodology [ p. 116 ]and Danish-Spanish translation, E.H. Jensen (1999) about Spanish-Danish legal translation, Andersen about ‘metaphor phobia’, A. Jensen (1999) about time pressure in translation, and Livbjerg and Mees (1999) about the use of dictionaries in Danish-English translation.

The author takes a critical view of TAP research; in addition to methodological reservations, he criticises e.g. the use of language students as subjects and the absence of a translation brief in the experimental translation tasks. Hönig also discusses students’ irrational strategies in e.g. dictionary use (see also Krings 1986b, Jääskeläinen 1989a).

A didactically oriented article partly based on TAPs. The author argues that students of translation need to be weaned from their “school-translation” concept by e.g. simulating professional assignments in translation class. Otherwise students fail to develop professional attitudes and practices, which shows as a lack of appropriate translation strategies and a failure to take responsibility for the success of translated communication.

Drawing partly on TAP research, the author proposes a flow-chart describing the translation process. The central concepts are controlled (conscious) and uncontrolled (unconscious) mental workspace (Kiraly 1990) in processing, and micro- and macrostrategies, which correspond roughly to local and global strategies (e.g. Jääskeläinen 1993). The latter are the general principles guiding the whole translation process (e.g. foreignising or domesticating), while the former are the isolated rules or procedures used to solve individual translation problems. The claim is that micro-strategic choices should always be guided by the macro-strategy.

In addition to modelling translation process as a flow-chart and discussing controlled vs. uncontrolled workspace and strategies in translation (see Hönig 1991), the author discusses findings from neuro-physiological research. The emphasis is again on translation didactics.[ p. 117 ]

A study of ten language students thinking aloud alone (four students) vs. translating in pairs (three pairs). The author argues for pair translation as a methodological and didactic tool, as the dialogue protocols show evidence of more sophisticated problemsolving strategies.

The article reports on a study in which ten language students translated with or without access to reference books. The students seem to fall into two groups, high-risk-takers, who are confident in both situations, and low-risk-takers, who are lost without their trusted aids. The article also contains a discussion of the nature of consciousness. The author argues for joint translating both as a teaching method (“translation in and as interaction”) and as the research method in translation process analysis.

The study investigates non-professional and professional-like translation performance with four translation students (two first-years and two fifth-years) as subjects (translation from English into Finnish). Analyses of the data deal with external processing (time spent, dictionaries used) and internal processing (mainly problem-solving). Reactions to the translation brief (“assignment”) are also looked at. Tentative profiles of nonprofessional vs. professional-like behaviour are drawn to be tested in further research.

A description of student translators’ use of dictionaries and other reference works on the basis of Jääskeläinen (1987).

[ p. 118 ]A description of student translators’ use of dictionaries and other reference works on the basis of Jääskeläinen (1987) with a link to classroom experiences in translator training.

A description of student translators’ reactions to the translation brief on the basis of Jääskeläinen (1987).

See Jääskeläinen 1989c.

The study combines the student data from Jääskeläinen (1987) with new data from four professional translators and four “educated laypersons” (true non-professionals in translation). Contrary to the initial plan (“identify professional behaviour”), the aim is to identify process features which seem to go together with successful task performance, as not all professionals succeeded well in the experiment. The explanations relate to differences in the amount of time and effort invested in the task, as well as differences in the types of knowledge used: the least successful processes, irrespective of the professionality of the translator, seemed to be characterised by treating the task as a mechanistic code-switching operation.

A theoretical discussion (based on empirical work) which focuses on defining the notion of “translation strategies” for the purposes of empirical translation research. Strategies are divided into global vs. local strategies (see also Hönig 1991).

[ p. 119 ]An overview article (in Finnish) which sums up the author’s previous TAP research and outlines future directions.

A methodological article which describes different verbal report methods, their merits and limits from the point of view of language research.

A comparison of the findings reported by Gerloff (1988) and Jääskeläinen (1990b). Both studies indicate that official professional status does not automatically guarantee high translation quality. Instead, faced with a non-routine assignment, even professional translators may need to invest more time and effort to ensure translation quality. The article also draws attention to affective factors as a potential explanation to unexpected research findings.

Based on Jääskeläinen (1990b), the article discusses factors which may come into play in a TAP experiment and influence the results. Such factors include e.g. role behaviour as described by Erwin Goffman as well as the (professional) subjects’ work history.

An overview article of TAP research into translating.

A doctoral dissertation focusing on three main areas: methodology in TAP research, cognition in translation processes (types of knowledge used), and affect in translation processes (personal involvement in the task). The data includes TAPs from eight subjects: four professional and four non-professional translators who translated a short English text into the subjects’ mother tongue (Finnish).[ p. 120 ]

A methodological discussion partly based on Jääskeläinen (1999) focusing on the potential interference effects of verbalising and thinking aloud vs. joint translating.

The analysis of automatised processing is based on Königs’ (1987) notions of Adhocblock and Rest-block and draws on data from the experiments reported in Jääskeläinen (1987) and Tirkkonen-Condit (1987, 1988, 1989). The authors provide evidence of potentially automatised parts of the translation process; they also argue that while some processes become automatised with increasing experience, other processes are evoked into consciousness leading to a heightened awareness of potential problems.

A methodological study which investigates the effect of thinking aloud (TA) on translation processing; more specifically, translation speed, the number of revisions, and the number of text production segments per source-text unit. In the experiment four semi-professionals and five professionals participated in four translating situations: English into Danish translation in TA vs. non-TA conditions, and Danish into English translation in TA vs. non-TA conditions. Translog (see Hansen 1999) was used in all conditions. The findings indicate that (1) thinking aloud significantly reduced translation speed, (2) had little effect on the number of revisions, and (3) increased the relative number of segments per source-text unit.

See Hansen (1999).[ p. 121 ]

A preliminary report on the doctoral dissertation reported in full in Jensen (2000).

The study investigates the effect of time pressure on the translation processes of nonprofessional translators, young professionals, and expert translators. The subjects performed translation tasks with a time limit of 30, 20 and 10 minutes. The qualitative TAP data are supplemented with quantitative data, i.e. computer logs files (Translog), which were also subjected to quantitative analysis (analysis of variance and multiple regression). Statistically significant effects of time pressure include faster start-up times the less time there is for the process, as well as the application of coping tactics, particularly among non-professionals.

The article sums up the methods and findings reported in full in Jensen (2000).

See Hansen (1999).

The study describes (in Swedish) the text comprehension processes of four students of translation and two professional translators tackling a translation task from French into Swedish. As a case in point the author discusses an ambiguous ST expression (typical of French) and its interpretation by the subjects.[ p. 122 ]

In this article two professional translators’ (cf. Jonasson 1996) products and processes are analysed from the point of view of the Tourian notion of translation norms. Based on textual and TAP evidence, it seems that one of the processes is guided by the norm of adequacy, while the other one is aiming at acceptability. Interestingly, neither subject makes an explicit statement of his or her initial normor global strategy, but these can be inferred from their commentary on local decisions. Differences in the translators’ attitude are also discussed.

A preliminary report of a longitudinal study in which five students participating in a translator training programmeand three professional translators,who also acted as teachers in the course, were the subjects. They translated a text fromFrench into Swedish (L1). The author’s hypothesis is that the differences between students vs. professionals identified at the beginning of the programme would tend to decrease during the course. The article focuses on text awareness which refers to the subjects’ reactions, if any, to the origin or purpose of the ST, ST coherence, and the text as a whole.

A doctoral dissertation the main aim of which is to contribute to the development of a systematic approach to translator training. The author identifies a pedagogical gap in Translation Studies; he also draws attention to the translator’s role in interlingual acts of communication. The empirical part of the study comprises TAPs from 18 subjects; nine novice students of translation and nine professional translators.

A revised version of Kiraly (1990).

[ p. 123 ]With a clear pedagogical emphasis, the article draws on the findings reported in Kiraly (1990) and (1995). In the light of his TAP data, the author argues for the adoption of a constructivist approach in the teaching of translation.

After a discussion of the product-orientedness of translation theory, the author suggests a shift into investigating translational practice. He reports on a pilot study with five subjects translating from Spanish into German. The analysis covers lexical and syntactic problem-solving strategies.

The translation process is here divided into two \‘blocks’: the Adhoc-block includes automatic processing, e.g. automatic 1:1 equivalents and translational experience (procedural knowledge), while the Rest-block includes items requiring conscious processing, e.g. linguistic problems as well as considerations regarding ST author or TT recipients. The model is illustrated by TAP data from five subjects: two novice students, two advanced students, and one professional translator. Didactic implications are discussed with regard to both translator training and FL teaching. See also Jääskeläinen and Tirkkonen-Condit (1991).

The study describes an experiment in which four subjects carried out two tasks, an FL writing task and a translation task into FL; in both cases travel brochures for Spanish speakers. The TAP data on the two tasks were first analysed separately (e.g. global planning, problem identification and solutions, revisions, dictionary use, and cultural transfer), then the writing processes were compared with the translation processes (structure of plans, strategies, etc.). One of the findings is that neither process proceeds in a linear fashion; both are characterised by recursiveness.

[ p. 124 ]The authors aim at providing new perspectives into the mental activities which take place during non-professional translation for language teaching purposes. The data comprise the TAPs from three advanced students of French, and the analysis covers e.g. global textual considerations, revision processes, and the use of dictionaries.

The article reports on a study in which TAP data was collected in a subtitling task with six subjects (two beginners, two with some experience, and two experts). In addition to TAPs, the analysis contains text analysis (e.g. condensation strategies) and interviews.

A preliminary report on the study reported in full in Krings (1986b). This report focuses on the identification of translation problems and translation strategies.

The first major publication in the field. Krings’ Doktorarbeit discusses thoroughly the methodological foundations of verbal report methods and describes the translational behaviour of eight subjects. The subjects are all German students of French as a foreign language (FL), i.e. non-professional translators. The data includes translation processes into the subjects’ mother tongue (by four students) and into the foreign language (by four students). Analyses cover the subjects’ use of time and reference books, the nature of problems as well as problem-solving strategies.

The article discusses think-aloud methodology on the basis of the experiment reported in full in Krings (1986b).

[ p. 125 ]A case study in which one professional translator tackled the same translation task as the language students inKrings (1986b). The findings indicate that the professional engaged in more problem-solving activities than the language students; also, her processing was more concentrated on particular problem areas which she returned to over and over again.

An overview article of the first empirical studies into the translation process.

A report of an experiment with student pairs translating together a text which calls for creative input. The author introduces findings and methods from creativity research; he also draws attention to affective factors, e.g. creating a classroom atmosphere conducive to creativity.

A book on translator training which draws partly on TAP research.

The study focuses on the interpretation of ST meaning in four dialogue protocols by advanced translation students. The theoretical framework includes prototype and scenes-and-frames semantics. The findings indicate that, at the reverbalisation stage of the process, students may abandon the good solutions which they have come up with while analysing the ST. As a result, the author suggests that the division between ST analysis and reverbalisation be discouraged in translation teaching.

An overview article combining TAPs and dialogue protocols, summarising findings, and suggesting future directions.[ p. 126 ]

An overview article of TAP research in Europe.

An investigation of translation errors combining TT quality assessment with an analysis of process data. The data comprises two translations from German into French (L2) which were assessed by four native speakers of French. One of the subjects was a translation student. The findings indicate e.g. that the translation student’s information retrieval skills were better, which may stem from the subjects’ different educational backgrounds.

A case study of one professional translator tackling two kinds of tasks: one which belonged to her normal commissions (travel brochures fromFinnish into English), and another which resembled them in terms of text-type (advertising) but dealt with a different topic. The findings indicate that the two processes differed not so much in terms of cognitive processing, but in terms of the translator’s attitude and self-image which were more positive in the routine task. This, in turn, led to better translation quality. To investigate attitude, the author analysed the quality and quantity of evaluative utterances in the TAP data.

The article sums up the main findings of Laukkanen (1993) and suggests topics for further research.

The study compares three professional translators performing routine vs. non-routine tasks. The subjects’ attitude and their text production strategies were investigated. The findings show that the link between attitude and translation quality is complex: in this study positive attitude seemed to be linked with perseverance and maintenance of highquality text production goals, while negative attitudes led to giving up easily and settling for less satisfying solutions.[ p. 127 ]

See Hansen (1999).

See Hansen (1999).

See Hansen (1999).

A preliminary report of Lörscher’sHabilitationsschrift (1991a) focusing on the methodology for analysing the subjects’ translation strategies.

See Lörscher (1989).

A theoretical discussion which identifies the research gap in translation research to be filled by empirical translation process analysis.

Lörscher’s book (based on his Habilitationsschrift) is the second extensive TAP publication and it describes the oral translation processes (i.e. sight translation) of 52 German learners of English (non-professional translators).Methodological issues are discussed [ p. 128 ]at length within a psycholinguistic framework and a sophisticated method for analysing translation strategies is introduced. The findings indicate, for example, that FL learners prefer a sign-oriented approach in translation (as opposed to sense-oriented translation by professional translators, see Lörscher 1993).

A methodologically oriented article discussing the merits and problems of thinking aloud as a research tool.

The article describes a research project in progress; the first stage describes the translation strategies of 45 advanced learners of French (findings reported in Lörscher 1991a); the second stage includes data from 22 professional translators (bilingual children were included later). The analysis is based on the method introduced in Lörscher (1991a). Preliminary findings show that professionals prefer a sense-oriented approach and considerably larger units of translation than the language learners.

See Lörscher 1993.

The study compares two first-year and two fourth-year students of translation who translate in two conditions, with or without access to dictionaries. Access to dictionaries slowed down the advanced students’ processing, but had little effect on the global strategies they used. Access to dictionaries did not seem to affect translation quality.

[ p. 129 ]A summarising article which discusses the inherent multidisciplinarity of Translation Studies and the potential problems involved, particularly from the point of view of translation and interpreting process analysis.

The study reports on an investigation of metaphor translation (English-Finnish). The theoretical framework is cognitive linguistics, and the aim is to test the “cognitive translation” hypothesis (if SL and TL metaphors utilise different conceptual mappings, translation requires not only a linguistic, but also a conceptual shift which, in turn, requires more processing effort). The data comprise TAPs from five students of translation. The findings support the cognitive translation hypothesis; in addition, the best translations were produced by those who were able to make the conceptual shift in translating metaphors.

A comparison of thinking aloud alone vs. joint translating (in groups of three) with the same subjects translating two source texts in the two experimental conditions. The subjects are grouped into novices (three first-year students), advanced students (three thirdyears + three fourth-years), and experts (three professional translators). The theoretical framework is based onVygotsky’s notions of consciousness, metacognition, and the genetic method. The author argues for joint translating as the method to study translating.

The article discusses the findings fromthe study reported inMondahl and Jensen (1992, 1996); in addition, there is a brief discussion of six more subjects’ (divided into two groups: medium and high proficiency) lexical search strategies.

The article reports on a study in which TAPs, retrospection and translated texts (Danish- English)were used as data. The subjectswere two advanced students of translation and two non-professionals (military personnel) who need English for their own professional purposes. The analysis deals with identifying differences in information processing (e.g. procedural vs. declarative knowledge) and the subjects’ interlanguage.

[ p. 130 ]Drawing on the data introduced inMondahl and Jensen (1992), the authors describe the subjects’ behaviour in terms of translation sequences (spontaneous vs. problematic) and different search strategies (divided into achievement and reduction strategies) and evaluation strategies. The advanced students seem to use more knowledge-based and experience-based knowledge. The findings also highlight the need for good evaluation strategies to assess translation solutions.

In this study the subjectswere four students of translation: twowith Germanas their major subject in translator training (B-language), and two with German as their minor subject (C-language). All the subjects performed tasks both into and from German. As the quality of the translations was equally good, it is hypothesised that general strategic competence in translation may to some extent balance differences in language proficiency.

An exception among the first studies which used thinking aloud in their research design, this study has a narrower focus than the other contributions published in 1986. TAPs from five naive bilinguals are used as one source of data about noun-compound translation between English and Hebrew. The other sources include native speaker interpretations of the meaning of the compounds and six written translations produced by naive translators (i.e. not professionals).

The study compares the translation processes of one professional and one non-professional (subject specialist). The findings indicate that the professional translator gleans information from the text as a neutral mediator of information, while the subject specialist tends to be side-tracked by her own views and attitudes. In addition, the nonprofessional proceeds in a linear fashion, while the professional’s process is characterised by numerous backward and forward shifts and concentration on problematic items (see Krings 1988a).[ p. 131 ]

To the best of my knowledge, this is first research report of a study in which TAPs were used to study translating. The TAP data was collected as an assignment in a seminar and the author was among the students who took part in the experiment. The categories of analysis include automatic retrieval, self-correction, and lexical search strategies.

The article reports on a study in which translation students translated texts in fairly large groups (up to 20 students in a group). The data comprises five ‘general’ translations and five LSP translations from English into German, and six ‘general’ translations and four LSP translations from German into English.

A case study of a Canadian government translator performing a routine translation task (from English into French). The subject verbalised little, and her processing had to be inferred fromthe external behaviour recorded on videotape, which supported the early hypothesis of automatic processing by professional translators (cf. e.g. Tirkkonen- Condit 1989; Krings 1988a). The analysis deals with e.g. the subject’s global translation strategies.

Starting with a discussion of the limitations of TAPs as a research tool, the author then describes a study in which two translators, who regularly work together, were videotaped. The analysis of the joint translation process shows evidence of non-linear procession, parallel processing, and the iterative nature of the process (returning again to accepted solutions). Furthermore, sometimes “the meaning on which the translation operates can be self-generated”.

The author draws on a number of studies to account for the inherent variability in translation, which complicates experimentation in translation research. The nature of the translation process is such that it does not lend itself easily to rigorously controlled experiments with neatly isolated variables. The sources of variation include skills, strategies, and the organisation of memory.[ p. 132 ]

Based on the dialogic data also discussed in Séguinot (1996), the article draws attention to the seemingly irrelevant processing which goes on in the joint translation process. This type of processing tends to be suppressed in think-aloud conditions—or at least, based on my own experiences, such phenomena tend to be ignored in the analysis. The author argues that a theory which aims at explaining the mental processes underlying translation should be able to account for the types of phenomena observed here and in Séguinot (1996).

The article opens with summarising different approaches to translation and knowledge. Drawing on cognitive science and translation studies (incl. TAP studies), the author discusses e.g. the development of translator competence as well as the notion of expertise in general and translational expertise in particular.

A preliminary report on the experiment later discussed in Tirkkonen-Condit (1988, 1989, 1991). The analysis covers two TAPs, one first-year student and one fifth-year student of translation (labelled as ‘non-professional’ and ‘professional’). The study focuses on translation problems and shows, for example, that the mature student identified more problems but spent less time on the process, i.e. was more sensitised to potential problems but also more efficient in problem-solving.

The study reports (in Finnish) on three translation students’ (two beginners and one mature student) processing while translating a text from English into their native Finnish. The aim is to identify different ways of interpreting the source text. The subjects are divided into two groups on the basis of their text comprehension patterns: [ p. 133 ](1) globalists who add textual and extralinguistic knowledge to linguistic details of the text, and (2) localists who rely on the linguistic details only.

The study reports on the same experiment as Tirkkonen-Condit (1988). Here the analysis focuses on classifying the subjects’ decisions and their distribution to different stages in the process. The mature student’s process contained the highest number of (verbalised) decisions (cf. Séguinot 1989), which also showed a higher reliance on encyclopedic knowledge than did the beginners’ decisions.

See Tirkkonen-Condit (1989).

Two subjects’ (a professional and a non-professional from Pöntinen and Romanov 1989) translation processes are analysed from the point of view of the type of knowledge used in translational decision-making. The professional’s process is described as a topdown governed, holistic process which utilises textual knowledge, while the nonprofessional processes sentences in isolation and allows her extra-textual knowledge to interfere with the contents of the ST.

The study reports on an experiment in which three teachers of translation translated a text from Finnish into English (two subjects translating from L1 into L2 and one from L2 into L1). The purpose was to test the subjects’ sensitivity to ST relational structure, more specifically to the Finnish clitic particle -kin.

[ p. 134 ]The author discusses research evidence from several TAP studies in order to illustrate the ways in which professionality is reflected in translators’ decision-making. Professionality is defined as successful task performance in relation to the translation brief, which means that sometimes success equals speed, sometimes it requires the production of a publishable text. Professional translators seem to be able to invest their efforts strategically depending on their analysis of the translation brief. In addition, contrary to some other contributions, the evidence reported here indicates that professionals have access to a well-organised and articulate body of procedural knowledge.

The study focuses on the development of translational proficiency by looking at e.g. addressee-oriented verbalisations in TAPs fromfour different experiments (Tirkkonen- Condit 1987, 1993; Pöntinen and Romanov 1989; Jääskeläinen 1990b).

Based on the assumption that translation competence involves tolerance of ambiguity and uncertainty, the study examines ‘uncertainty management’, i.e. the ways in which professional translators deal with the uncertainty inherent in translation. An analysis of six translators’ TAPs suggests that uncertainty management includes e.g. deliberation of tentative solutions and postponement of final decisions.

Drawing on data from two TAP experiments with professional translators, the authors analyse evaluative utterances to identify the translators’ professional self-image and their subjective theories of translation. The findings seem to indicate a positive relation between confidence and translation quality (but cf. Laukkanen 1997), which would challenge translator training to facilitate the development of positive professional selfimage and confidence in students of translation.

[ p. 135 ]A critical overview article of empirical translation research, with TAP research (Sandrock 1982) as one of the cases discussed. The author raises the still unanswered issue of the potential interference effect of the need to verbalise on the end product.

A case study with three balanced bilinguals (English/German) with no formal translator training as subjects. The aim is to investigate Harris and Sherwood’s natural translation hypothesis5 (see e.g. Krings 1986b), although the subjects were asked to produce a written translation of a written text.