Abstract

After a slow start, Computer-Assisted Interpreting (CAI) is receiving growing attention from practitioners and researchers alike. While research on CAI tools is still somewhat limited, a few studies have tried to shed some light on aspects such as the tools’ usability and their impact on interpreting quality in terms of terminological performance, mainly focusing on their use in the booth. The fact that many of these studies were conducted as part of BA or MA theses shows a certain interest among trainee interpreters. The new generation of interpreters is certainly more technology-savvy than their older colleagues; personal computers, tablets or other kinds of technological tools have become a staple in the interpreting booth. The prerequisites for the introduction of CAI tools in the curriculum of trainee interpreters are there, but how have training programmes reacted to these new terminology management solutions for interpreters? To answer this question, in this paper we present the results of a survey conducted among 25 interpreter-training institutions. The survey shows that, while technology applied to interpreting is present in all programmes object of the survey, only some universities have integrated CAI tools in their curriculum. Overall, despite the growing interest in this emerging field, there is still some confusion and lack of information among trainers. With the aim of providing an example of how to introduce CAI tools to students, this paper briefly describes how some universities have included these solutions in their training and presents a pilot study conducted at the University of Mainz/Germersheim (Germany), which involved trainee interpreters as test subjects.

Keywords: computer-assisted interpreting, simultaneous interpreting, interpreter training, terminology management

1. Introduction

ICTs applied to interpreting have been playing an increasingly relevant role not only in the profession, but also in interpreter training. As shown by Berber-Irabien’s survey of ICTs in interpreting (2010), which addressed both the professional and the educational fields, the most relevant technologies in interpreter training are setting-oriented solutions such as technologies for remote interpreting, videoconferencing and remote simultaneous interpreting (RSI), and computer-assisted interpreter training (CAIT). The latter involve e-learning platforms, online resources for interpreter training (such as ORCIT[1] and Interpreter Training Resources[2]), online speech databases like Speechpool[3] and the DG SCIC Speech Repository[4], video corpora like ELISA (Braun 2006) and the Backbone corpus (Kohn 2012), software programmes such as BlackBox (Sandrelli 2007) and SCICrec, and even virtual reality environments such as IVY (Chmiel et al. 2012; Tymczyńska et al. 2013).[5] These solutions provide the infrastructure for blended or distance learning and complement traditional, in-class training by providing additional material for self-study and additional tools for instructor, peer and self-assessment.

The relevance of setting-oriented technologies and CAIT for interpreter training is reflected in research, with a focus on the integration of the above-mentioned technologies in the interpretation curriculum (Braun et al. 2012; Kajzer-Wietrzny and Tymczyńska 2014; Fantinuoli and Prandi 2018) or on the development and use of CAIT in interpreter training (Merlini 1996; Sandrelli 2007, 2015; Sandrelli and de Manuel Jerez 2007; Ko 2006, 2008; Chmiel et al. 2012; Tymczyńska et al. 2013). Additionally, Orlando (2010, 2015, 2016) explored the use of digital pen technology for process-based assessment in note-taking training for consecutive interpreting.

A third, emerging area of technologies applied to interpreting can be identified in computer-assisted interpreting (CAI), which is gaining popularity among professional interpreters and is receiving increased attention in interpreting research. While very broad in its scope, Berber-Irabien’s survey did not reveal the current state of CAI inclusion in interpreter training, which can be easily explained by the relatively recent introduction of this technology. To our knowledge, no contribution can be identified in the body of interpreting research on the inclusion of CAI tools in interpreter training. With the aim of gaining a first understanding of how interpreter-training courses have responded to the emergence of computer-assisted interpreting, we carried out a survey involving 25 interpreter-training institutions.[6] In this paper, we discuss the current state of research on CAI tools and we present the results of the survey. We provide selected examples of the integration of CAI tools in interpreter training that we believe can provide inspiration for further development of interpreter training courses in this direction. As an example of how to introduce CAI to trainee interpreters, we present the training phase of a process- and product-based study on CAI tools. Finally, we conclude by discussing the results of the survey and addressing future work in the field of computer-assisted interpreting research.

2. State of the art: CAI tools

Computer-assisted interpreting (CAI) tools are software solutions developed specifically to suit the needs of interpreters in terms of terminology and knowledge management. In their most advanced form, they represent an interpreter’s workstation that supports each phase of the interpreter’s workflow, especially conference preparation, but also the peri-process, in-process and post-process operations (Kalina 2005) necessary to ensure efficient management of the terminological resources and a high-quality interpretation in language for special purposes (LSP) conferences.

Although not as numerous as computer-assisted translation (CAT) tools, there is a growing array of CAI tools available on the market for interpreters to choose from. Some examples[7] are Flashterm[8], Glossarmanager[9], Glossary Assistant[10], Interplex[11], InterpretBank (Fantinuoli 2012)[12], Interpreter’s Help[13], Intragloss[14] and Terminus[15].

Some recent publications present an overview of these solutions (Costa et al. 2014a; Rütten 2017) and provide a set of criteria for their evaluation (Costa et al. 2014b; Will 2015). Fantinuoli (2018) takes a different approach and classifies CAI tools into first-generation and second-generation tools. While first-generation CAI tools only provide  a more systematic way of organizing terminological resources than traditional tabular glossaries[16], second-generation CAI tools offer a much wider range of functionalities, drawing from computational linguistics and providing solutions such as automatic terminology extraction from preparation documents, support for the memorisation of specialised terminology and access to terminological databases not only before and after the interpreting task, but also in real time, during interpreting itself.

If they can be considered the equivalent to what CAT tools are for translators, CAI tools have not enjoyed the same level of integration in the workflow of professional interpreters, who still tend to use more rudimentary and non-interpreter-specific solutions, as shown by several surveys (Valentini 2002; Berber 2008; Berber-Irabien 2010; Bilgen 2009; Projektgruppe KoDoTools des Sprachen & Dolmetscher Instituts München 2007; Corpas Pastor and May Fern 2016). The main reason for this may be that such tools are more often than not met with scepticism (Drechsel 2013b) or are simply unknown to most professional interpreters. Some practitioners perceive them as somewhat “unnatural” to the very same act of interpreting (Donovan 2006) and potentially distracting (Tripepi Winteringham 2010) during simultaneous interpreting (SI).

Despite these points of criticism, which mainly concern the in-process phase, CAI tools have received heightened attention in most recent years, as testified by the increasing number of workshops and webinars offered on the topic by professional associations (e.g. by the ITI[17], AITI[18], AIIC[19], VKD[20]). This is an encouraging sign for the further diffusion of such tools, which have the potential of making preparation more efficient, reducing the time spent on it. The author’s experience as a CAI trainer showed that the functions pertaining to the pre-process phase, such as terminology extraction and memorization, are generally judged positively by expert interpreters. The speed of the look-up functions provided by some tools is also appreciated. Additionally, terminology systematization, glossary export and import in various formats, and solutions for collaborative work are all perceived as useful innovations when compared to traditional, non-interpreter-specific solutions.

As for any tool, much depends on the user’s individual preferences. Trainers would do well to raise awareness on the inherent limitations and issues of CAI tools, while pointing out their usefulness for the professional practice and their potentially positive impact on interpreting quality - at least in terms of terminological accuracy. To this aim, we recommend following a constructivist approach with the active involvement of trainees, who should be given the chance to test and compare the different tools and features. Fantinuoli and Prandi (2018) provide guidelines for CAI use in interpreter training and suggest practical activities and resources for trainers[21].

Interpreting research has supported CAI tools developers by defining theoretical models that describe the ideal structure and functionalities of a CAI tool. Rütten (2004, 2007) offers a comprehensive theoretical model of CAI tools, while Will (2007) and Fantinuoli (2006) provide models of terminology work conducted by interpreters. Recently, Fantinuoli (2017) developed a prototype for the integration of Automatic Speech Recognition (ASR) into CAI tools, thus paving the way for the third generation of CAI.

Another type of investigation on CAI tools can be identified in several BA, MA and PhD theses, which have dealt with various aspects of this technology. Some of them analyse the usefulness of CAI tools for conference preparation, comparing them with more traditional solutions and evaluating their main features (see for example De Merulis 2013; Gacek 2015). Others take an empirical approach and investigate their impact on the terminological quality of interpreters’ performance (Biagini 2015; Prandi 2015a, 2015b). In particular, Prandi explores the reception of CAI tools by trainee interpreters, providing a qualitative evaluation of their usage in the booth and highlighting the need for practical experience in addition to theoretical instruction on the topic. Her research also highlights the potential risk for students of relying too heavily on such tools, a risk which should be taken into account and addressed by trainers. In his study, drawing on Prandi’s methodology, Biagini focuses on a quantitative analysis of the terminological quality of the interpreted text comparing the use of a paper glossary and the CAI tool InterpretBank. It is the first attempt at collecting quantitative data to provide an empirical basis for the ongoing discussion on such tools and it suggests that CAI tools may provide an advantage in terms of terminological quality when compared to paper glossaries. In her study, Xu (2015) found that computer-assisted terminology extraction - a feature present, for instance, in the CAI tools InterpretBank and Interpreter’s Help - can improve conference preparation, increase terminological accuracy and reduce omissions. Even though these studies, with the exception of Xu’s, have a preliminary character and are limited in scope, they all seem to indicate that CAI tools can have a positive effect on the quality of terminology in the interpreter’s delivery and that the tools are well received among students. A doctoral study underway at the University of Mainz/Germersheim (Prandi 2017, 2018) aims at expanding the analysis by exploring the effects of CAI tools not only on the terminological quality of an interpreter’s rendition, but also on cognitive load during SI with CAI, an aspect which has not yet been addressed in CAI research.

While more research in this field is certainly needed, these studies already provide good grounds for the inclusion of CAI tools in the interpreting curriculum, especially since they have the potential to help interpreters address the fast-paced changes occurring in the profession (Donovan 2006).

3. Where are we now? A survey on CAI tools in interpreter training

The goal of the survey is to explore the current landscape of the diffusion of CAI tools in conference interpreter training at university level and to gain a better understanding of the factors influencing their inclusion in curricula. The questionnaire (see Annex 1) is based on the survey conducted by Berber-Irabien (2010), but has a narrower scope, as it focuses only on process-oriented CAI tools as a subset of technologies applied to interpreting and only on the educational setting.

3.1. Sample

The survey was administered via Google Forms between October and December 2017 and sent to the participants via email. The institutions to be contacted were selected on the basis of the CIUTI member list[22], which provides the contact information either of the dean of the institute or of the head of the interpreting department, or both. Whenever the persons contacted specified that they were not experts in the field or not directly responsible for the inclusion of technologies in the curriculum, we either contacted the person they suggested or asked for their contact details, to make sure that the respondents had the necessary knowledge of interpreting technologies. In order to reach a higher number of interpreter trainers, the survey was also posted on the Facebook Group “Interpreter Technology Group”[23], a forum which holds in-depth discussions on the topic of technologies applied to interpreting.

Figure 1: training institutions per country

Of the 85 interpreter training institutions contacted, 25 questionnaires were returned from 15 countries. The graph above (Figure 1) shows the distribution of the institutes per country. The most represented countries are Belgium, Germany, Italy, Austria and Spain. Overall, the sample is quite representative of the European landscape. The only non-European country is Lebanon. All institutions offer a master’s degree in Conference Interpreting.

3.2. Results

All respondents taking part in the survey reported including technologies for interpreting in their curricula. The following technologies were listed in the question concerning the inclusion of technologies in the curriculum: remote interpreting, videoconferencing, telephone interpreting, computer-assisted interpreter training tools, computer-assisted interpreting tools, online terminological databases and dictionaries, and search engines.

We also wanted to verify whether universities included technologies in a specific subject, and if not, how they included them in the curriculum. More than half of the sample (15 institutions) reports dedicating a specific subject to technologies for interpreting.

Table 1: Are technologies addressed in a specific subject?

The other 10 training institutions include technologies in the curriculum especially focusing on videoconferencing and remote interpreting, for example in the context of virtual classes with other institutions or with the EU. Some provide a general introduction to the topic. One institution reports using “speech databases for practice, sourcing terminology, using recordings and transcriptions (digital pen) for analysis etc.”. Another university adopts SCIC-Rec “for supervised self-training”, while in another institution a short seminar on technologies is offered within the course “Professional contexts and interpreting resources”. One institution also lists preparation tools as one of the technologies included in the curriculum.

As for CAI tools, more than half of the sample (13 institutes out of 25) includes them in the curriculum. In this regard, it should be noted that two affirmative answers had to be reclassified as negative, since from other questions it emerged that the respondents included technologies applied to interpreting, but not CAI tools. This is very interesting, as it suggests a certain degree of confusion around the concept of computer-assisted interpreting tool.

Table 2: Are computer-assisted interpreting (CAI) tools part of the curriculum?

In both cases, the respondents were asked to motivate their answers. As for the reasons for which these institutions do not include CAI tools in their curriculum, some common themes emerge from the responses collected.[24] Two respondents pointed to the trainers’ expertise as a decisive factor for the integration of CAI in the curriculum. In one case, there is “a lack of lecturers able to teach them”. In the other, the inclusion of CAI in training is left to the trainer’s personal initiative. This is not surprising and ties in with what emerges from surveys on the use of CAI tools by professional interpreters, since most trainers are also practitioners (see Section 2). Another reason that we expected to see mentioned was the difficulty of altering the curriculum to officially include CAI training. This was mentioned by one respondent, which indicated that curricula were shaped before such tools were widespread. Two respondents indicated that, while “related” technology or CAT tools were included in training, CAI was not. One other institution stated that they had not considered including CAI tools. This might point to a lack of interest in CAI but could also (if interpreted optimistically) be seen as a starting point for their future inclusion. A rather surprising claim was made by one institution, which stated that CAI tools are “too recent”. The first CAI tool prototype was created by Christoph Stoll in 1993, while the first publications on terminology management systems for interpreters can be dated back to the early 2000s. Given the current speed of technological development, a technology that first emerged 27 years ago can hardly be defined as “too recent”. It is possible that the respondent was actually thinking of other technologies when providing this answer. This also seems to be the case for 2 more responses which point to a lack of financial and infrastructural resources and “technical issues” as reasons for the exclusion of CAI. While budgeting issues are certainly plausible, the only infrastructure needed for working with CAI is a PC or a laptop and (for some functions and some tools) an internet connection, as well as a technician for the necessary maintenance, all things usually available at universities. It would certainly be interesting to explore this further with the institution in question, but we suspect this is another case of confusion on the actual definition of CAI.

When CAI tools are included in the curriculum, lessons usually include both a theoretical presentation and practice sessions. Three institutions report focusing on the practice sessions. This is confirmed by the question aimed at finding out whether students learn how to use the tools or whether they are just informed of their existence (Table 3). Apart from one institution where students are only informed of the existence of these tools and are provided with a short demo, students learn how to use the tools, even though, in one case, instructors focus mostly on theory and only little practice is offered.

Table 3: Are students taught how to use the tools, or are they just informed of their existence?

The most interesting question in the survey is perhaps the one that concerns the CAI tools presented to trainee interpreters. As shown by the table below, InterpretBank is the tool students are most often introduced to, followed by Interplex and Interpreter’s Help. Other tools presented to students include, in descending order of popularity, Terminus, Flashterm.eu, Glossary Assistant, Intragloss, LookUp and Glossarmanager.

Figure 2: Which CAI tools are presented to trainee interpreters? [n = 13]

InterpretBank is actively used by students in all the universities that include it in the curriculum. The same can be said for Terminus, LookUp, Glossary Assistant and Glossarmanager. Interplex, Interpreter’s Help, Flashterm and Intragloss are also often used actively by students and only in rare cases are they solely presented in theory (Figure 3). The respondents also mentioned additional tools that are indeed technologies that can be or are usually applied to interpreting, but which are not CAI tools. Some examples are SDL MultiTerm, CrossTerm, online databases and corpora, corpus creation and analysis tools, but also videoconferencing software (Zoom, Polycom, Adobe Connect).

Figure 3: Do students actively learn how to use the CAI tools, or do they just learn of their existence?

In 7 cases out of 13, 50% or more of the teaching hours on interpreting technologies are dedicated to CAI. In 2 cases, only CAI tools are taught. In 4 cases out of 13, one third of teaching hours on technologies for interpreters is dedicated to CAI, while in the remaining 2, CAI plays a marginal role (around 17%). The differences in the number of hours allocated to CAI can be explained with the personal preferences of the individual trainers, the ease of adapting the curriculum to newer technologies, the inclusion of CAI in regular interpreting classes, the preference given to setting-oriented technologies (such as those for remote interpreting), which are more present in the public debate about technological changes in the profession.

In the institutions where CAI tools are part of the curriculum, students use these tools and are encouraged to do so also during regular conference interpreting classes in eight cases out of 13 (Table 4). In three cases, students use CAI tools even though they are not explicitly encouraged to do so, in one case they are not encouraged to use them, and in one instance they do not use them even though they are encouraged to do so.

Table 4: Do students use CAI tools when practicing conference interpreting in the classroom?

The last set of questions aimed at finding out what the future prospects of the institutions involved in the survey are when it comes to CAI tools. Seven institutions out of 25 are planning to expand their curriculum to include CAI tools, by including a course on the subject, by increasing booth practice and teaching hours, or by “implementing high-performance strategies” as well as “validating empirical findings” on the topic. In one case, remote interpreting was also mentioned. Fourteen universities out of 25 state that they may expand or change the way CAI tools are included in their curriculum, and list teaching hours as well as financial and technical resources as critical factors in this decision, but also indicate they have different priorities, and that the evolution of market demands also plays a role. One university is thinking of including more activities on CAI tools, another one underlines that the curriculum needs to be changed in order to accommodate for such integration.

Four institutions do not plan to include CAI tools in their curriculum in the near future nor to change the way CAI tools are included in their curriculum. These respondents either believe they already cover the subject sufficiently (2 cases out of 4), or do not have an interest in these tools (1), or do not provide an explanation (1). We also asked the respondents to indicate whether any research project on this subject was being carried out at their institution. While 11 universities answered yes, five of them described research projects that have to do with technology applied to interpreting, but not involving CAI tools specifically, such as remote interpreting, telephone interpreting, real-time subtitling with speech recognition, or the preparation of multimedia material to be used to teach conference interpreting. Two institutions indicated that research was underway but did not provide more detailed information. This means that, in our sample, six institutions at best are actually conducting research on computer-assisted interpreting tools. Five out of six state that trainee interpreters are involved in these projects. Five respondents had no information on the matter and nine institutions stated that they are not conducting research in this area.

4. Where do we go from here?

CAI tools can be included in interpreter training in many ways, and some institutions already do this by integrating activities with CAI in regular interpreting classes, even though they do not offer a course specifically dedicated to CAI tools. This is very promising and can represent a good solution in those situations where the curriculum cannot be changed in order to make room for a new course dedicated only to interpreting technologies (and in particular to process-oriented technologies). After all, even though a stand-alone module can certainly prove beneficial, as it allows instructors and trainees to explore the topic in detail, technologies should ideally be integrated in regular interpreting classes, at least after trainees have developed a sound interpreting technique, for example in the last semester of their studies (Fantinuoli and Prandi 2018). This would ensure that students are equipped with the necessary interpreting “toolkit” (Drechsel 2013a) to keep up with the current interpreting market.

Among the survey respondents, some interpreter training institutions stand out for the way they have included CAI tools in their interpreting curriculum. In this section, we will select and describe the four most relevant examples of the inclusion of CAI in interpreter training through a course specifically dedicated to computer-assisted interpreting (4.1), hoping this will provide some inspiration to those institutions wishing to include such tools in their curriculum.[25] Moreover, institutions which provide a specific module on CAI tools also seem to include students in research projects on such tools. In 4.2 we will therefore describe a training course on CAI tools which is part of a pilot study conducted at the University of Mainz/Germersheim (Prandi 2017, 2018).

4.1. Selected examples from the survey

The University of Bologna (Italy) offers a master’s degree in conference interpreting at its Department of Interpreting and Translation, at Forlì campus. Within the programme, a course on “Methods and technologies for interpretation” is offered. It is a compulsory 40-hour course for all trainee interpreters, 25 hours of which are dedicated to CAI tools. The CAI tools presented to students are InterpretBank, Interplex and Interpreter’s Help. In Forlì, interpreting students also learn how to use SDL MultiTerm for conference preparation. During the course, students receive theoretical instruction but also have the opportunity to practice using these tools for conference preparation and in the booth. InterpretBank is also used and tested during events organised at the university, such as conferences or seminars. At the time of the survey, there was no research project specifically dedicated to CAI tools, but a study was underway on ASR for live subtitling, a technology also presented during the course.

The University of Heidelberg in Germany also offers its interpreting students a specific course dedicated to CAI tools. Eight of the ten hours of the course are devoted to CAI tools. Trainee interpreters are introduced to the following tools: InterpretBank, Interplex, Intragloss, LookUp, Interpreter’s Help and Flashterm. The tools are used during workshops and are considered part of the workflow of interpreting for a technical conference. There is currently a study underway about the empirical quantification of high-performance strategies - including CAI - and the automation of the interpreter’s workflow.

The University of Innsbruck (Austria) and the University of Mainz/Germersheim offer a course on CAI tools taught by the same lecturer, which is why we present them together. The courses offered in Innsbruck and Germersheim are both 22 hours long. The teaching hours dedicated to CAI are respectively six and eight. Students are introduced to InterpretBank, Interplex, Intragloss (only in Innsbruck) and Interpreter’s Help. Both courses combine theory and practice. The focus of the theoretical part is on the goals of CAI and on the features of interpretation during LSP conferences. Trainees then have the opportunity to practice interpreting with the support of InterpretBank in the booth. At the time of the survey, there was no study underway at the University of Innsbruck with a specific focus on CAI tools. In Germersheim, apart from the study currently being conducted by Prandi (2017, 2018), there is another project dedicated to the usability of a glossary for community interpreting in application forms.

4.2. Research as a way of bridging the gap: a pilot study on CAI tools in SI

The research project underway at the University of Mainz/Germersheim (Prandi 2017, 2018) combines process-oriented and product-oriented methods - in particular eye tracking, keylogging and the analysis of the terminology used by the test-subjects. It compares SI performance with the support of traditional glossaries, second-generation CAI tools and a prototype of what we could define as a third-generation CAI tool (with integrated ASR). The study aims at investigating how CAI tools influence the terminological quality of interpreters’ renditions and cognitive load during SI. Before taking part in data collection, the students participating in the study as test-subjects undergo training on the use of CAI tools in the booth during SI. For the purpose of this paper, we will focus on the training offered during a pilot study aimed at testing the research methodology and the experimental set-up in preparation for the main phase of the experiment. We will highlight the potential benefits of the training and describe the changes implemented in the main study. For further information on the results of this pilot study, the reader can refer to Prandi (2018).

The training took place between April and June 2017 at the University of Mainz/Germersheim. Seven students of the master’s degree in Conference Interpreting took part in the training (3 German natives and 4 Italian natives). Participation in the study was voluntary and the participants did not receive any form of compensation. One Italian native was then excluded from data collection because she had only received two semesters of training in SI up to that point, while the other test-subjects had received at least 3 semesters of training.

The participants took part in a preliminary meeting, which covered the basics of terminology management for interpreters and focused on the CAI tool InterpretBank, also used during the practice sessions and data collection. The one-hour workshop allowed the students to gain a first understanding of the goals of CAI tool usage. The students were provided with an overview of the tools currently available on the market and had the opportunity to ask questions and clarify any doubts. Over the following weeks, the students took part in 5 practice sessions. During each session, they interpreted three short speeches of around 10 to 12 minutes each, on a variety of topics ranging from biology to medicine, from English into their mother tongue. The students interpreted each speech with the support of a different tool. They could look up terminology in a glossary prepared with Microsoft Word, a glossary prepared with Microsoft Excel, and a glossary accessible with InterpretBank. Before the practice sessions, the trainer indicated which tool they were supposed to work with and prepared the glossaries, so that all students worked with the same material and equal time was dedicated to each tool. During the sessions, whenever a technical problem arose, the students could discuss it with the trainer in order to clear any doubts and solve potential issues before data collection. After the last practice session, the students took part in a test to check their proficiency in the use of the tools. The test included a series of tasks to be performed with the three tools (e.g. identifying and activating a certain function, selecting options, looking up terminology). Their laptop screen was recorded with the screen-casting software Active Presenter[26] and the recordings were then reviewed to verify whether the tasks had been correctly carried out. All the students passed the test and were therefore deemed ready for data collection.

A first observation of the data collected during the pilot study and recent advances in the field of CAI tools led us to replace one of the traditional glossaries with a glossary accessible with a CAI tool with integrated ASR. This change was implemented in the main study, where a PDF glossary was used as an example of traditional glossary in lieu of the Word and Excel glossaries. A further opportunity for training that was implemented in the main study and that could prove a valuable alternative for some trainers is offering the training remotely, for example as a self-paced Moodle course. In the main study, the students took part in an introductory webinar covering the same topics as the in-person workshop and were able to communicate with the trainer whenever needed, but this allowed them to individually pace their practice sessions. Additional benefits could be the possibility to adjust their training to their level of proficiency, as some students may already know how to use CAI tools or require different amounts of practice.

The inclusion of students in research projects on CAI could provide benefits both to trainees and to trainers. When extra hours cannot be added to the official curriculum, research projects on CAI can represent an additional opportunity to offer students a theoretical and practical introduction to new tools for conference preparation outside of the regular interpreting classroom. Such projects can represent the first direct experience of their use in the booth, which requires some time getting accustomed to. For the scope of our study we focused on InterpretBank, but it would certainly be interesting for students to experiment with different tools. Another opportunity would be involving students also in the preparation of the terminological resources, an activity which we believe would have a positive effect on the optimisation of their conference preparation strategies and also on their interpreting strategies in the setting of an LSP conference where they have access to their terminological resources. Finally, the extra effort required when using these tools would also be a way for students to test their capacity distribution and to work on attention allocation.

5. Conclusions and future work

In this paper, we presented the results of a survey aimed at drawing a picture of the inclusion of CAI tools in interpreter training. After describing the current state of research on CAI tools, which has mainly focused on their evaluation and on analysing their impact on terminological quality in SI, we discussed why there is a case for their inclusion in interpreter training.

What emerged from the survey is a multifaceted picture of the inclusion of CAI tools in curricula. The number of responses received and the fact that half of the responding institutions already include CAI in interpreter training is certainly a sign that training is starting to take account of this technology. In the survey, we did not provide a definition of CAI tools because we wanted to understand whether trainers were conversant with these new technologies for interpreting. When answering the survey, some respondents interpreted CAI as a synonym for ICT applied to interpreting, in particular of setting-oriented technologies such as remote interpreting and videoconferencing. This may point to a lack of information among conference interpreting trainers, probably due to the novel nature of these tools. Another reason might be a lack of interest, which could be explained by the fact that experienced interpreters usually already have their own terminology management system and therefore do not update their knowledge in this field or try out new tools or systems. Interest is however growing, as shown by the relatively high number of institutions represented in the survey, by the increasing number of studies on the topic, and by the desire to expand curricula so as to include CAI tools in training. On the basis of the survey, at present, InterpretBank is by far the most widely used CAI tool in interpreter training, but the variety of CAI tools presented to trainee interpreters is also a piece of good news.

A potential obstacle to the introduction of CAI tools in interpreter training is a lack of material and financial resources, but also of adequate preparation among instructors. For this reason, we believe that, especially where resources are limited and where the curriculum cannot be changed, research projects on CAI tools can represent a good way to top up official training, helping students improve their preparation strategies.

Our survey is quite representative of the European landscape, but further research is needed to expand this observation on a global scale. It would certainly be interesting to conduct this survey again in a few years, in order to carry out a diachronic analysis of the response of training institutions to the emergence of new computer-assisted interpreting tools on the market. Similarly, it would be useful to check whether students who have received training in CAI benefit from it also in their professional life, especially in terms of terminology management. Such analysis could shed new light on how CAI tools influence the work of professional interpreters and would certainly be beneficial in designing an interpreting curriculum that reflects professional practice and promotes an open, impartial and data-driven discussion on the role of process-oriented technologies for the interpreting profession.

About the author(s)

Bianca Prandi is a research associate at the University of Mainz (Germersheim campus), where until March 2021 she will collaborate on the research project on machine interpreting M.INT. She holds an MA in Interpreting from the University of Bologna/Forlì and is currently a PhD candidate at the University of Mainz/Germersheim. Her research focuses on the impact of computer-assisted interpreting (CAI) tools on terminological quality and cognitive processes in simultaneous interpreting. Her research interests include technology applied to interpreting, terminology management for interpreters, and cognition. She regularly provides training on CAI.

Email: [please login or register to view author's email address]

•  Previous article