Master
Reference
Closing the Gap Between Academic Offer and Industry Demand : A Qualitative Research Study on Translation Technology-related
Competences Within the EMT Framework
GUARNÉ I AYUSO, Ona
Abstract
The European Master's in Translation (EMT) Framework released its second set of competences in 2017. The first set had not been revisited since 2009. This master's thesis examines whether the technological competence within the 2017's EMT Translation Competence Model meets the translation and localisation's industry demands. The investigation was carried out following a mix-methods approach in order to evaluate the current trends of the market, as well as to identify technological contents in the EMT MA programmes accepted for the period 2019-2024. The analysis concentrates on job requirements and the offer of theoretical and practical content in courses. The results of comparing the technology-related course guides of the EMT MA programmes against real market job requirements disclose some gaps between academia and the industry. Albeit courses in CAT tools and Machine Translation appear to be the most taught, knowledge concerning these is not as demanded in the market. In contrast, knowledge in QA and Localisation seemed to be the most demanded in the industry, but QA and Localisation contents did not seem to be given as [...]
GUARNÉ I AYUSO, Ona. Closing the Gap Between Academic Offer and Industry Demand : A Qualitative Research Study on Translation Technology-related Competences Within the EMT Framework. Master : Univ. Genève, 2020
Available at:
http://archive-ouverte.unige.ch/unige:145376
Disclaimer: layout of this document may differ from the published version.
CLOSING THE GAP BETWEEN ACADEMIC OFFER AND INDUSTRY DEMAND: AQUALITATIVE RESEARCH STUDY ON TRANSLATION TECHNOLOGY-RELATED
COMPETENCES WITHIN THE EMTFRAMEWORK
Directrice : Silvia Rodríguez Vázquez Jurée : Lucía Morado Vázquez
Mémoire présenté à la Faculté de traduction et d’interprétation (Département de TIM, Unité d’espagnol) pour l’obtention de la
Maîtrise (Ma) universitaire en traduction et technologies, mention localisation et traduction automatique
Année Académique 2019-2020 / Août 2020
Acknowledgements
A l’Eulàlia, moltes gràcies per haver-me ajudat a emprendre un viatge llarg però necessari, durant el qual m’he fet més forta i que m’ha portar a ser capaç d’acabar la tesina, i culminar així els meus estudis.
Thank you, Christina, for having been by my side from the beginning and staying on it, especially during the tough times. Thank you for having spoken the hard truths when they were difficult to swallow. The minute you knew this paper was going to be written in English you offered your help and I am beyond grateful for having been my very own thesis reviewer.
A Cristina y a Xènia, os estoy muy agradecida por haberos convertido en mi sistema de apoyo durante toda mi estancia en Suiza, especialmente en Ginebra. Gracias por haber sido pacientes conmigo y haberme ayudado con el máster cuando no podía yo sola. No podría haber pedido mejores amistades, que fueran tan humildes, generosas y buenas como vosotras dos.
To Ruth, thank you for having been the light in the dark during the cataclysm that represented my first year in Switzerland. I am very obliged to you for having offered your tiny little studio and your big kind heart to me when I was in dire need of help. You have contributed to my growth in so many ways.
A la Miriam, moltíssimes gràcies per haver-me donat suport durant tota la investigació i per haver-me guiat quan estava encallada. Estic increïblement contenta d’haver trobat per casualitat en tu una familiaritat natural, un riure extremadament càlid i honest i un cor que no et cap al pit. Moltes gràcies per haver-me ajudat tant durant el camí.
Y, finalmente, muchísimas gracias a Silvia y a Lucía, vuestro apoyo durante la redacción de la tesina ha sido crucial para mí. Vuestra forma de supervisión tan cercana y honesta ha sido la perfecta para alguien como yo que estaba dando sus primeros pasos en investigación.
Abstract
The European Master’s in Translation (EMT) Framework released its second set of competences in 2017. The first set had not been revisited since 2009. This master’s thesis examines whether the technological competence within the 2017’s EMT Translation Competence Model meets the translation and localisation’s industry demands. The investigation was carried out following a mix-methods approach in order to evaluate the current trends of the market, as well as to identify technological contents in the EMT MA programmes accepted for the period 2019-2024. The analysis concentrates on job requirements and the offer of theoretical and practical content in courses. The results of comparing the technology-related course guides of the EMT MA programmes against real market job requirements disclose some gaps between academia and the industry. Albeit courses in CAT tools and Machine Translation appear to be the most taught, knowledge concerning these is not as demanded in the market. In contrast, knowledge in QA and Localisation seemed to be the most demanded in the industry, but QA and Localisation contents did not seem to be given as much attention in the EMT MA programmes.
Keywords
Translation technology – Translation competence – European Master’s in Translation – Master programme – Postgraduate programme – Language Service Provider –
Translation and localisation industry
Table of Contents
List of Tables ... 8
List of Figures ... 9
List of Abbreviations ... 10
1. Introduction ... 12
1.1. Research Context ... 12
1.2. Motivation ... 13
1.3. Goals, Research Questions and Methods ... 14
1.4. Structure of This Thesis ... 15
2. Literature Review ... 17
2.1. Historical Perspectives and Technological Trends ... 17
2.1.1. The Evolution of Translation Technology ... 17
2.1.2. Technological Trends in the Translation and Localisation Industry .. 23
2.1.2.1. General Overview: Translation and Localisation as a Sector ... 23
2.1.2.2. Technology within the Translation and Localisation Industry ... 25
2.1.2.2.1.Technology-related Sectors in Expansion ... 25
2.1.2.2.2. Perceptions of Technology ... 27
2.1.2.3. Translation Technology within the Translation and Localisation Industry ... 27
2.1.2.4. Forecasts and Watchlist ... 30
2.2. Developments in Translation Competence Teaching ... 32
2.2.1. Translation Competence ... 32
2.2.2. Translation Competence Models ... 33
2.2.2.1. PACTE’s Translation Competence Model ... 33
2.2.2.2. Göpferich’s Translation Competence Model ... 36
2.2.2.3. Kiraly’s Translation Competence Model ... 38
2.2.2.4. EMT Framework Translation Competence Model ... 40
2.2.3. Translation Technology Competence ... 41
2.2.3.1. Translation Technology in the Classroom: an Overview ... 42
2.2.3.2. Technology-related Competence in the EMT Framework Translation Competence Model ... 43
2.2.3.3. Prior Work on the EMT Framework Translation Competence Model ... 46
3. Methodology ... 49
3.1. Research Questions, Goals, and Hypotheses ... 49
3.1.1. Research Question (R1) and Hypothesis Statement (H1) ... 49
3.1.2. Research Question (R2) and Hypothesis Statement (H2) ... 50
3.2. Data Samples ... 51
3.2.1. Data Sample 1: Selection of LSPs and Position Postings ... 51
3.2.1.1. Data Sample 1a: Selection of LSPs ... 51
3.2.1.1.1.LSPs Collection Procedure ... 52
3.2.1.2. Data Sample 1b: Selection of Position Postings ... 53
3.2.1.2.1.Position Postings Collection Procedure and Process ... 54
3.2.2. Data Sample 2: EMT MA Programmes and Technology-related Course Guides ... 55
3.2.2.1. Data Sample 2a: Selection of EMT MA Programmes ... 55
3.2.2.1.1.EMT MA Programmes Collection Procedure and Process ... 55
3.2.2.2. Data Sample 2b: Selection of Technology-related Course Guides .. 56
3.2.2.2.1.Technology-related Course Guides Collection Procedure and Process ... 56
3.3. Data Analysis ... 59
3.3.1. Phase A: Qualitative Analysis ... 59
3.3.1.1. Thematic Analysis Procedure ... 59
3.3.1.1.1. Coding ... 62
3.3.1.1.2. Tools Employed ... 63
3.3.2. Phase B: Quantitative Analysis ... 64
3.3.2.1. Descriptive Analysis Procedure and Simple Graphical Analysis .... 64
3.3.2.2. Tools Employed ... 64
4. Findings and Discussion ... 67
4.1. Position Postings ... 67
4.1.1. Location and Type of Position ... 68
4.1.2. Languages Desired ... 70
4.1.3. Qualifications ... 72
4.1.4. Background and Experience ... 74
4.1.5. Desired Soft Skills ... 78
4.1.6. Technology-related Knowledge and Specific Tools and Devices ... 81
4.1.7. Summary of Findings and Discussion ... 91
4.2. Technology-related Course Guides ... 102
4.2.1. General Overview: Universities and Allocated ECTS ... 102
4.2.2. Nature of Courses ... 105
4.2.3. Theoretical Content ... 106
4.2.4. Software Content ... 111
4.2.5. Summary of Findings and Discussion ... 116
5. Conclusions ... 126
5.1. Overview of the Research ... 126
5.2. Main Findings ... 126
5.3. Limitations and Future Work ... 127
5.3.1. Limitations of Data Sample 1b ... 127
5.3.2. Limitations of Data Sample 2b ... 128
5.3.3. Future Work ... 129
6. Bibliography ... 131
Appendices ... 140
Appendix A – Details of Selected LSPs ... 140
Appendix B – List of Selected Position Postings ... 142
Appendix C – List of EMT Members 2019-2024 ... 145
Appendix D – List of Selected Universities ... 154
Appendix E – List of Selected Technology-related Course Guides ... 156
Appendix F – Overview of Software in Technology-related Course Guides by Course Type ... 163
List of Tables
Table 2-1: Technological Competences Published in 2009 and in 2017 ... 44
Table 3-1: Details of LSPs Selected for the Study ... 52
Table 3-2: Position Postings per LSP ... 54
Table 4-1: Languages Asked Once in Position Postings ... 72
Table 4-2: Number of EMT MA Programmes by Country ... 103
Table 4-3: B Languages Offered in EMT MA Programmes ... 103
List of Figures
Figure 3-1: Codes Grouped under Soft Skills Theme at an Early Stage of Analysis ... 61
Figure 3-2: Codes Grouped under Software Theme at an Early Stage of Analysis ... 61
Figure 3-3: Coding a Position Posting in Atlas.ti 8 (PDF Format) ... 62
Figure 3-4: Coding a Technology related Course Guide in Atlas.ti 8 (PDF Format) .... 63
Figure 3-5: Code Search across Position Postings in Atlas.ti 8 ... 65
Figure 3-6: Code Search across Technology-related Course Guides in Atlas.ti 8 ... 65
Figure 3-7: MS Excel Results Export ... 66
Figure 4-1: Position Postings Released per Location ... 68
Figure 4-2: Position Postings per Type ... 69
Figure 4-3: Languages Desired in Position Postings Asked More Than Once ... 71
Figure 4-4: Qualifications Desired in Position Postings ... 73
Figure 4-5: Qualifications per Position Type ... 74
Figure 4-6: Background Asked in Position Postings ... 75
Figure 4-7: Background per Position Type ... 76
Figure 4-8: Previous Experience Asked in Position Postings ... 77
Figure 4-9: Experience per Position Type ... 78
Figure 4-10: Soft Skills Valued in Position Postings ... 79
Figure 4-11: Soft Skills per Position Type ... 80
Figure 4-12: Desired General IT Skills ... 82
Figure 4-13: Desired General IT Skills per Position Type ... 83
Figure 4-14: Spreadsheets and Word Processors per Position Type ... 84
Figure 4-15: Sought-after Programming Languages ... 84
Figure 4-16: Programming Languages per Position Type ... 85
Figure 4-17: Operative Systems Identified in Position Postings ... 86
Figure 4-18: Operative Systems per Position Type ... 87
Figure 4-19: Multimedia and Creativity Software Identified in Position Postings ... 87
Figure 4-20: Multimedia and Creativity Software per Position Type ... 88
Figure 4-21: Translation-related Software Identified in Position Postings ... 88
Figure 4-22: Translation-related Software per Position Type ... 89
Figure 4-23: Other Software Identified in Position Postings ... 90
Figure 4-24: Other Software per Position Type ... 90
Figure 4-25: Technical Skills and Experience of an MT Scientist Position ... 93
Figure 4-26: Roles and Responsibilities of an MT Scientist Position ... 93
Figure 4-27: Number of Allocated ECTS in Course Guides ... 104
Figure 4-28: Nature of Course Guides ... 106
Figure 4-29: Theoretical Content Identified in Course Guides ... 108
Figure 4-30: Theoretical Contents Arranged per Nature of Course ... 110
Figure 4-31: CAT Tools Identified in Course Guides ... 112
Figure 4-32: Tools for Localisation Identified in Course Guides ... 112
Figure 4-33: Programming Languages Identified in Course Guides ... 114
Figure 4-34: Multimedia and Creativity Software Identified in Course Guides ... 115 Figure 4-35: Word Processors and Grammar Correctors Identified in Course Guides 115
List of Abbreviations
ACE: Automated Content Enrichment AI: Artificial Intelligence
API: Application Programming Interface AV: Audio-visual
CAT: Computer-assisted Translation
CEFR: Common European Framework of Reference for Languages CMS: Content Management System
CNGL: Centre for Next Generation Localisation DL: Deep Learning
DTP: Desktop Publishing
EAGLES: Expert Advisory Group on Language Engineering Standards ECTS: European Credit Transfer System
EMT: European Masters in Translation FAUT: Fully Automatic Useful Translation GMS: Globalisation Management Systems IT: Information Technology
LSP: Language Service Provider MA: Master of Arts
ML: Machine Learning MSc: Master of Science MT: Machine Translation
NLP: Natural Language Processing NMT: Neural Machine Translation OCR: Optical Character Recognition OS: Operative System
OTT: Over-the-top
PACTE: Process in the Acquisition of Translation Competence and Evaluation PC: Personal Computer
PE: Post-editing
PM: Project Management QA: Quality Assurance
S2S: Speech-to-Speech Translation
SDH: Subtitles for the Deaf and the Hard-of-hearing SMT: Statistical Machine Translation
SWOT: Strengths, Weaknesses, Opportunities and Threats TAUS: Translation Automation User Society
TM: Translation Memory
TMS: Translation Management System TTS: Text-to-Speech Translation UI: User Interface
1. Introduction
This master’s thesis examines the gap between the translation and localisation market and academia regarding technological skills. More specifically, it investigates whether the technological competences released by the European Master’s in Translation (EMT) Framework for the period 2018-2024 are aligned with the needs of the translation and localisation market in 2019-2020. For that purpose, it compares the desired requirements and skills listed in position postings released by the largest Language Service Providers (LSPs) with the technology-related contents of course guides of EMT universities. This master’s thesis hopes to contribute to the general knowledge on this area of the language industry and identify trends of valued skills in the sector.
Furthermore, it seeks to provide information to help improve technology-related courses in universities, ultimately both at undergraduate and postgraduate level.
The following sections offer contextual information about the research topic (Section 1.1.), describe the reasons for conducting this research (Section 1.2.), and detail the goals and research questions that resulted in this thesis (Section 1.3.). To conclude the Introduction section, the structure of this thesis will be outlined (Section 1.4.).
1.1. Research Context
The translation and localisation industry moves millions of USD and is growing each year. A key factor of the industry is technology: from general Information Technology (IT) software, such as email or word processors, to translation technologies, such as Computer-Assisted Translation (CAT) Tools, Machine Translation (MT), Quality Assurance (QA) checkers, Content Management Systems (CMS), or specialised software for localisation.
The presence of technology in the translation field has been steadily increasing since the appearance of CAT tools in the 1980s (Sin-wai, 2015). Translators are asked to complete and deliver complex translations in less and less time in order to give answer to market demands. Currently, the advancements in Neural Machine Translation (NMT) and its introduction to the mainstream make it seem like translators will have to develop new and more advanced technological skills in order to be able to incorporate MT and other new technologies into their day-to-day work. Only by making the most out of new tools will they be able to stay competitive in the translation and localisation sector.
In view of the above, “if graduates are to be competitive in the market, translation education programmes need to continue to improve the way in which […] technologies are taught and learned” (Marshman & Bowker, 2012). Having in mind the employability of graduates and to ensure quality in translation training programmes, the European Master’s in Translation (EMT) Framework released a translation competence model in 2009. Its relevance emerges from the result of consensus from a set of European experts and provides the structure of the training plan for over 50 universities in Europe (Pym, 2013). The EMT Translation Competence Model appears to be the most successful translation competence model proposed by translation scholars (Esfandiari et al., 2019).
It had remained unchanged until 2017, when the EMT Framework released a second set of competences valid for the period 2018-2024, in order to give answer to the many technological changes occurred in the translation industry since 2009. These competences replace the first set and are divided into five main areas: language and culture, translation, personal and interpersonal, service provision, and technology. These new competences were conceived with the future translation graduate employability in mind (EMT Expert Group, 2017).
1.2. Motivation
Although in the last years many reports sought to investigate the translation and localisation market’s needs and trends (Massardo & van der Meer, 2017; memoQ Translation Technologies, 2020; Pielmeier & O’Mara, 2020; Tirry, 2019), fewer studies investigate the relationship between the technological contents taught in the EMT Master’s (MA) programmes and the demands of the translation and localisation market.
The EMT label seeks to improve the quality of translator training to facilitate the integration of graduates into the labour market, and this is the main reason why the EMT Translation Competence Model has been taken as the reference model in the present study.
Rodríguez de Céspedes (2020, p. 20) recently published an article concerning the current use of translation technologies and digital tools by LSPs in the UK, and one of the conclusions was that LSP employers “acknowledge the so-called gap between academia and the professional world.” Several scholars have investigated topics related to the EMT label focusing on the 2009’s set of competences (Chodkiewicz, 2012; Plaza Lara, 2019; Ester Torres-Simón & Pym, 2017), while others have studied technological
contents at an undergraduate level in translation programmes (Piqué Huerta & Colominas, 2013; Sayaheen, 2019). Nevertheless, based upon the review of the literature and to best of our knowledge, there are no research studies to date which compare the current situation of the translation and localisation market from the employer’s perspective with the EMT’s proposal of 2017 regarding translation technologies.
1.3. Goals, Research Questions and Methods
This section describes the aims of this thesis. As previously discussed, the scope of this particular investigation is narrowed down to the EMT competences of 2017, particularly the technological competence. To delimit the scope of the investigation, the focus was put on two goals:
GOAL 1: Identify trends of valued skills in the translation and localisation sector in 2019-2020, with a particular focus on the technological skills.
GOAL 2: Determine whether the technological competences published by the EMT Framework in 2017 line up with the needs of the translation and localisation market in 2019-2020.
To achieve the aforementioned goals, the researcher adopted a mixed-methods approach. The study conducted combines both thematic analysis and descriptive analysis strategies in order to answer the following research questions, which will be reviewed in detail in Chapter 3:
(R1) What technological skills are the largest LSPs in the translation and localisation market seeking out in job applicants in 2019-2020?
(R2)Do the technological competences in translation released by the EMT for the period 2018-2024 meet the translation and localisation market’s needs in 2019-2020?
To address R1, position postings released by eight of the largest LSPs in the translation and localisation sector were collected and were analysed using Atlas.ti 8, a qualitative data analysis software. Only LSPs with both yearly revenue higher than USD 100 million/year in 2019 and office presence in Ireland and/or Switzerland were considered. Following these criteria, 92 position postings were collected between September 2019 and October 2019. The data collected were analysed following a thematic analysis approach, and the resulting themes and units were presented using
descriptive statistics. The mixed-methods analysis of this data set provided the necessary quantitative data to answer R1.
In the framework of R1, the following hypothesis stemmed from the literature review:
(H1) Most LSPs will seek out proficient general IT skills and knowledge on MT.
The answer to R1 needed to be obtained so that R2 could be answered, since it is the core research question of this investigation. To address R2, the technology-related course guides of all EMT programmes accepted for the period 2019-2024 were first included for study. Only completed course guides (which contained at least the course objectives and contents) were included in the data set. These were analysed using the same qualitative data analysis software, Atlas.ti 8. After completing the thematic analysis, the resulting themes and units were presented using descriptive statistics. The mixed- methods analysis of this data set, combined with the comparison against the results from the data set of position postings, provided the necessary quantitative data to answer R2.
In the framework of R2, it was deemed necessary to formulate the following hypothesis, which stemmed from the literature review:
(H2) The technological competences published by the EMT for the period 2018-2024 fail to align with the translation and localisation market’s needs in 2019-2020, even though they represent an improvement compared to the set published in 2009.
In order to support or reject H2, the results from the analysis of the selected position postings were compared against the results of the technology-related course guides.
1.4. Structure of This Thesis
To conclude this introductory chapter, an outline of how the work is organised in its entirety will be briefly described. This thesis is structured in five sections. Chapter 1 offered contextual information about the research topic, discussed the relevance, the goals and the research questions of the study, and gave an overview of the applied methods.
Chapter 2 reviews fundamental concepts related to this research, as well as prior work conducted on the topic. Firstly, the historical perspectives of CAT tools and
technological trends in the translation and localisation sector in 2019-2020 are reviewed in Section 2.1. Then, translation competence models are discussed in Section 2.2. This subsection also provides an analysis of technological competence in the EMT model.
Chapter 3 is dedicated to the methodology adopted. The chapter starts discussing in detail the research questions, goals and hypotheses introduced in Chapter 1. The chapter then is divided into two main sections: the description of the two main data samples collected for the study (Section 3.2.) and the chosen methods of analysis (Section 3.3.).
Chapter 4 presents the quantitative findings of the investigation. To facilitate understanding, the results are divided into position postings and technology-related course guides. Chapter 4 also discusses the findings presented and seeks both to (i) answer R1 and R2 and (ii) examine whether H1 and H2 are supported or rejected.
Chapter 5 gives an overview of the investigation and briefly summarises its findings. The thesis is concluded by highlighting its limitations and by proposing future research directions.
2. Literature Review
In order to locate the current state of translation technology, this chapter will first describe how technology became intrinsically tied to the translation activity over time:
from its first appearance in the 1960s until recent years. Moreover, the technological trends that can be observed in the translation and localisation market at the time of writing will be discussed with the aim of identifying the significance of technology within the sector.
Then, the place of translation technology within translator training will be studied, translation competence will be overviewed and four of the most influential translator competence models will be described. Lastly, the teaching of translation technology at university level and the technological competence within the EMT Translation Competence Model will be outlined.
2.1. Historical Perspectives and Technological Trends
Since the idea of CAT systems was first brought up in the 1950s, translation technology struggled to make a difference on the translation activity and the translation industry in the early years. In spite of this, the panorama of translation technology has changed dramatically ever since the first CAT system was released. In 2019, over 500 technological tools for translation and interpreting, media localisation, MT, machine intelligence, and marketplaces and platforms have been mapped by Nimdzi (2019b). This number represents an increase of 125% compared to those mapped for 2018 (ibid).
The following sections will provide an overview of the rapid intrusion of technology into the translation process, and the technological trends in the translation and localisation market in recent years.
2.1.1. The Evolution of Translation Technology
Even though there is a myriad of translation technology tools currently being commercialised and available for the translator, they would probably be significantly different if the concept of CAT systems had never been invented.
Essentially, CAT systems allow linguists to reuse legacy translations found in translation memory (TM) databases and automatically apply terminology contained in
terminology databases. Other supplemental functionalities such as alignment tools, term extraction tools, creating databases from previous translations, or compiling termbases from TMs, bilingual glossaries and other documents can be added to these basic characteristics.
The birth of CAT systems, or more accurately, the idea of computers assisting the translation process (as described by Garcia (2015)), traces back to circa 1949 as a result of the unfruitful experiments with MT. The concept of CAT systems resulted from the ambition of producing quicker, cheaper, and still high-quality translations. However, due to the constraints of technological advancements at the time, the concept of TMs did not appear until the 1980s.1 When Personal Computers (PC) were unreachable by the broad public, translators hand typed their translations and most likely stored copies in paper format (Garcia, 2015, p. 3). However, the introduction of the IBM Personal Computer in 1981 marked the beginning of real productivity gains for translators (Massardo et al., 2016, p. 7). The 1980s represent the start of word processors, automatic spell and grammar checkers and computer-based glossaries (ibid).
By the early 1990s, the necessity of offering products and services to potential clients that spoke other languages increased, which in turn ramped up the translation industry (ibid). Due to the high volume of translations and tight turnaround times, teams of translators were required to work simultaneously on the same source material (Garcia, 2015, p. 3). Therefore, maintaining the consistency in terminology and the possibility of reusing legacy translations became crucial. In view of this, two technology-proficient, business-driven translators (the German Hummel and Knyphausen, who in 1984 founded Trados) released the terminology database MultiTerm in 1990 (ibid). Later in 1992, the first commercial CAT system was launched, Translator’s Workbench I and Translator’s Workbench II, which was the first-ever system to integrate TM and alignment features into its workstation (Sin-wai, 2015, pp. 5–7). Additionally, the American company IBM released its in-house developed system Translation Manager 2 in 1992, and the German LSP STAR AG commercialised Transit, its own in-house system (Garcia, 2015, p. 3).
1In the mid-1980s, ALPS (Automated Language Processing Systems) in Salt Lake City, Utah, developed the Translation Support System (TSS), which is considered to be the first prototype of a CAT system.
However, it was not ready for commercialisation (Garcia, 2015, p. 3).
Until 1993, only three CAT systems were available on the market: Trados Translator’s Workbench II, IBM Translation Manager/2, and STAR Transit 1.0. Soon similar products were released to the market. By 2003, approximately 20 systems were being commercialised (some are still commercialised today, while others have been discontinued). Among these systems one can find the well-known Déjà Vu, Eurolang Optimizer, Wordfisher, SDLX, ForeignDesk, Trans Suite 2000, Yaxin CAT, Wordfast, Across, OmegaT, MultiTrans, Huajian, Heartsome, and Transwhiz (Sin-wai, 2015, pp.
8–11). It was at this stage that Trados positioned as the preferred tool by many stakeholders and, eventually, became the default industry standard (Garcia, 2015, p. 3).
To the basic functionalities of CAT systems (typically TM, terminology management, and translation editor) more components were gradually incorporated (Sin- wai, 2015, p. 12). Over time, MT and Project Management (PM) features were also included (ibid). By the mid-1990s, the more advanced systems included, at least, TM, terminology and translation management, alignment and term extraction tools, file conversion filters, and QA (Garcia, 2015, pp. 3–9).
CAT systems also adapted to the great variety of document formats, dealing directly, or with filters, with Adobe InDesign, FrameMaker, HTML, Microsoft PowerPoint, Excel, Word, QuarkXPress, and even PDF by 2003 (Sin-wai, 2015, p. 12).
As described by Bowker & Corpas Pastor (2015, p. 2):
“The increased interest in CAT has been needs-driven – on the part of both clients and translators – as recent decades have witnessed considerable changes in our society in general, and in the translation market in particular. Most texts are now produced in a digital format, which means they can be processed by computer tools. Largely as a result of globalization, there has been a significant increase in the volume of text that needs to be translated into a wide variety of languages. In addition, new types of texts, such as web pages, have appeared and require translation. Furthermore, because companies want to get their products onto the shelves in all corners of the world as quickly as possible, and because electronic documents such as web pages often contain content that needs to be updated frequently, deadlines for completing translation jobs seem to be growing ever shorter.”
In line with the expansion of document formats supported, after the release of Microsoft Office 2000, the implementation of Unicode allowed the gradual resolution of obstacles in language processing. This implied that CAT systems could handle more languages (Sin-wai, 2015, pp. 12–13).
While CAT systems were well-developed to deal with general documentation and web content in general, they did not hit the mark regarding software user interfaces (UIs).
This new type of content display, radically different from the traditional documents and extremely visually oriented, included graphical controls such as menu bars, drop-down menus, toolbars, windows, buttons, pop-up messages, error messages and dialogue boxes, and all of which could be selected using a mouse or keyboard. Identifying and extracting the translatable text from the software code presented itself as a tricky challenge.
Moreover, given that the conventional punctuation rules used to identify segments were of no use for this type of content, localisers came up with a new approach revolving around “text strings” rather than segments (Garcia, 2015, p. 10). Additionally, a visual panel was included to quickly check that the display of translated text fitted the space designated for it (ibid). These two features established the limits between tools for localisation and CAT systems. The principle behind localisation tools was similar to CAT systems, and also included TMs, term bases, alignment and term extraction tools, PM and QA features (ibid). The well-known localisation tools Passolo2 and Catalyst3 are still commercialised today and are included in the Nimdzi Language Technology Atlas 2019, among others such as Multilizer,4 Sisulizer,5 Lokalize6 (which replaced KBabel) or PO-Edit.7
As described by Garcia (2015, p. 10):
“Eventually, as industry efforts at creating internationalization standards bore fruit, software designers ceased hard-coding translatable text and began placing it in XML-based formats instead. Typical EXE and DLL files give way to Java and .NET, and more and more software (as opposed to text) files could be
2 Available from: https://www.sdl.com/fr/software-and-services/translation-software/software- localization/sdl-passolo/. Accessed on Apr 4, 2020.
3 Available from: https://www.alchemysoftware.com/products/alchemy_catalyst.html. Accessed on Apr 4, 2020.
4 Available from: https://pdf.multilizer.com/en/. Accessed on Apr 4, 2020.
5 Available from: https://www.sisulizer.com/. Accessed on Apr 4, 2020.
6 Available from: https://lokalise.com/. Accessed on Apr 4, 2020.
7 Available from: https://poedit.net/. Accessed on Apr 4, 2020.
processed within conventional CAT systems. […] Nowadays, the distinctions which engendered localization tools are blurring, and they no longer occupy the field exclusively.”
CAT systems continued to be developed throughout the 2000s. While until the late 1990s only LSPs and corporate buyers benefitted from leveraging and savings resulting from CAT systems, independent translators started to own CAT tools at a gradual pace (Garcia, 2015, p. 10). Owning a CAT system and being proficient at it made them visible to the localisation industry, which essentially broadened their market horizons (ibid). From 2000, professional associations and training institutions played a key role in the promotion of CAT systems among freelancers (ibid).
Key factors of the evolution of CAT tools have been both computer processing power and connectivity (Garcia, 2015, p. 12). CAT systems that were developed between 2004 and 2013 already included common features such as PM, spell check, QA, and content control, and were server-based, web-based and even cloud-based. By 2012, “there were fifteen cloud-based CAT systems available on the market for individuals or enterprises, such as Lingotek Collaborative Translation Platform, SDL World Server, and XTM Cloud” (Sin-wai, 2015, p. 22). The commercial offer also became wider during the aforementioned period so the potential buyer could choose from a different range of packages, functions, operation systems and prices. About 30 new systems were released and the old systems kept being upgraded regularly (Sin-wai, 2015, p. 13).
The TAUS Translation Technology Landscape Report (2016, p. 9) defined the 2010s as “the age of web services.” Internet brought innovations to translation technology such as “crowdsourcing, community translation platforms, hybridisation of TM and MT, and easy-to-use web-based translation platforms” (ibid, p. 10).
Currently, translation technology includes a broad selection of tools and it is no longer an exclusive synonym of CAT systems. As mentioned in the introduction of the section, over 500 translation technology tools have been mapped by the Nimdzi Language Technology Atlas in 2019. The Globalization and Localization Association (GALA)8 (2015) states both that “the translation and localisation industry today is driven by
8 Additionally, GALA’s website provides an overview of the main sectors within language technology (CAT, MT, TMS, TM and Website Translation Technology); what they consist on and provides examples of specific tools. Available from: https://www.gala-global.org/language-industry/language-technology.
Accessed on Apr 4, 2020.
technology” and “virtually all language professionals use localization and translation software in their daily work.”
According to Roturier (2015, pp. 116–149), professionals may currently count on the following tools:
- Translation Management Systems (TMS) and workflows, which proved to be popular among large LSPs due to their level of flexibility between stakeholders (content owners, project managers, translators, reviewers and Desktop Publishing (DTP) specialists).
- Translation Environments, which may be web-based or desktop-based.
- TMs, which integrated in a CAT tool allow to reuse legacy translations.
- Terminology Management, which comprises term extraction and glossary creation.
- MT, which is becoming mainstream among many LSPs,9 and includes Post-Editing (PE) tools.
- QA tools, which may include rules-based checks, statistical checks, machine learning-based checks, and handling quality standards.
To the aforementioned translation technology tools, the TAUS Translation Technology Landscape Report adds (Massardo et al., 2016, pp. 21–27):
- Globalisation Management Systems (GMS), which is typically provided through an Application Programming Interface (API) and is used to optimise enterprise localisation automation workflow systems (as described by Schäler (2008, p. 207)).
- Localisation Project Management systems, which integrate significant business processes of a translation business, and Proxy-based Localisation Management Platforms, which exploit the principle of proxying.
- App localisation systems, which are usually “localisation proxies with limited TMS capabilities” (as described by Massardo et al. (2016, p. 23)).
9 For instance, 2019 market leaders TransPerfect (https://www.transperfect.com/blog/dia-session-recap- automated-translations-pharma-companies), Lionbridge
(https://www.lionbridge.com/whitepaper/machine-translation/) and RWS (https://www.rws.com/what-we- do/rws-moravia/go-global-model/localize/) already offer MT services. All aforementioned websites were accessed on Feb 27, 2020.
- Captive Translation Management Systems, which can be “either implementations of commercial SaaS systems for the exclusive use of a company’s clients […] or proprietary platforms specifically developed by a translation business to enhance process automation” (as described by van der Meer (2019, p. 298)).
- Community translation platforms, which is a way to “integrate open knowledge into a collaborative network following a social media model”
(Massardo et al., 2016, p. 24).
- Middleware, which is a “suite of connectors providing services for the integration of […] CMSs and TMs by managing content flow from one system to the other” (ibid).
- Controlled authoring tools, which is “the process of applying a set of predefined style, grammar, punctuation rules and approved terminology to content (documentation or software) during its development” (Ó Broin, 2009).
- Audio-Video captioning tools, which are software programmes that allow text to be added to a sequence of video (Massardo et al., 2016, p. 25).
- Translation apps, which “translate signs and print-outs using barcodes or QR codes, spoken text through megaphones […] or a mobile device microphone, or messages within the most common messaging apps” (ibid, p. 26).
- Speech-to-Speech translation (S2S), the goal of which is to “enable instant oral cross-lingual communication between people not sharing a common language” (ibid, p. 27).
2.1.2. Technological Trends in the Translation and Localisation Industry
This section will provide an overview of the language industry at the time of writing, the perceptions of technology in general and of translation technology, and the forecast for the translation and localisation industry.
2.1.2.1. General Overview: Translation and Localisation as a Sector According to a recent report that gathers the size and state of the language services industry in 2019 (Nimdzi, 2019a, p. 21):
“The language industry in 2019 is ramping up to deal with increased volumes. Leading translation buyers in software, pharmaceutical, intellectual property, and manufacturing sectors add languages, start localizing videos for marketing and training. They experiment with machine translation to deal with volumes of user content far too large for professional translators to cope with. […]
In media localization, the explosion of online streaming production leads to unprecedented challenges of scaling and opens the space to new providers. […]
Technology startups appear and get funding every year, and artificial intelligence and neural computing have become an inevitable part of nearly every business conversation, but they have not disrupted the industry in a meaningful way — yet.”
Nimdzi (2019a, p. 23) estimated that the market size for language services10 would be USD 53.5 billion for 2019, and while the translation services accounted for over 50%
of the total revenue in 2018, technology barely reached 1.5%. However, according to Slator (2020), the global language services and technology industry grew to a USD 24.2 billion market in 2019 (based on buyer spend). The Public Sector, Technology, and Travel & Retail were the biggest contributions, which together made up for more than 46% of the market (ibid).
For 2020, Nimdzi (2020, p. 21) estimated that the industry size would be of USD 57 billion11, which represents an increase of 6.5% compared to the estimated for 2019. As a result of the increase in the demand, the market is expected to keep growing and reach USD 77 billion by 2025 (ibid). Additionally, Nimdzi’s data (2020, p. 23) show that media and entertainment generated 17.8% of revenue in the industry, while 13.2%
came from IT and software.
Regarding personnel, over 398,000 professionals were estimated to be employed by the language industry in 2019 (78,000 of which were LSP employees and an estimated 250,000 were full-time professional translators, interpreters and subtitlers) (Nimdzi,
10 In the Nimdzi report (2019a), language services in this understanding are defined as core services such as translation, interpreting, localization of software, website and multimedia, including film and TV series, and many smaller ones including multilingual marketing, multilingual DTP, eDiscovery, linguistic testing and respective technologies.
11 The estimated effects of COVID-19’s pandemic (both positive and negative) have not been reflected in Nimdzi’s report. Slator (2020) estimates that the COVID-19 outbreak could lead to an 8% decline in the overall market over the course of 2020.
2019a, p. 23). The top 100 LSPs listed in Nimdzi’s report (2020, p. 32) concentrate over 50,000 employees.
2.1.2.2. Technology within the Translation and Localisation Industry Technology experiences a positive trend in the translation and localisation sector.
According to a recent report by the market research company Technavio (2020) “the adoption of technology to enhance language translation process efficiency is one of the primary factors driving language services market growth.” Futurologist Ray Kurzweil predicts the industry to move towards Fully Automatic Useful Translation (FAUT) by 2030, even though a few scholars and industry professionals reject his statement (Massardo & van der Meer, 2017, p. 10).
2.1.2.2.1. Technology-related Sectors in Expansion
The demand for data training is high worldwide and this acts as a growth driver for the top two LSPs, TransPerfect and Lionbridge, which are after AI support services such as data annotation and data labelling (Nimdzi, 2020, p. 28).
In line with memoQ’s predictions (2020) and Technavio’s (2020), the media localisation sector experienced a dramatic change and growth in 2019 according to Nimdzi (2020, p. 31). There is more audio-visual (AV) content being produced and
“traditional roles are shifting in the media industry. As traditional distributors like Netflix started to produce original content, traditional producers like Disney realized that to stay competitive, they needed to enter into direct-to-consumer services” (ibid). While the first half of 2019 endured a slowdown in production in preparation of the release of over-the- top (OTT) media services like Disney+ and Apple TV+, the other half experienced an explosion of content as soon as these platforms were launched (ibid). Then, consolidated
“players” such as Netflix and Amazon released even more content. All these events resulted in an increase of work for the media localisation sector (ibid). However, “the real surge is said to come in late 2020 and early 2021 when the internationalization of all this content is expected to begin happening” (ibid).
In 2020, technology is also expected to be used to optimize services. Regarding this, Nimdzi (2020, p. 32) states that:
“[…] Technology is playing a crucial role in enabling LSPs to optimize their internal workflows and to deliver better services to their clients. Buyers are
increasingly demanding language service providers to deliver not only localized assets but also to deliver them in the specific format that is required for their distribution channels. Offering an end-to-end distribution platform for localized assets that optimizes this process is key for localization companies to stand out from their competitors.”
Another sector that is experiencing a shift is MT. According to a report published in 2018 about MT market size and trends, the global MT market size is expected to reach USD 983.3 million by 2022 (Research and Markets, 2018). The TAUS report about the translation industry in 2022 (2017, p. 7) ranked MT as one of the six drivers of change in the world of translation. Additionally, TAUS recognised NMT as the most popular
“hype” within the industry (Massardo & van der Meer, 2017, p. 11):
“Organizations that have worked with the technology are convinced that it [NMT]
offers significant advantages over phrased-based SMT [Statistical Machine Translation] in terms of output fluency and accuracy. Unlike […] SMT which uses look-up tables to “learn” comparable phrases, NMT learns to translate sentences by using technology called neural networks – several layers of linked “neurons”
that operate in symbiosis to roughly imitate the cognitive processes used in the human brain to recognize and learn patterns of information. Like SMT, NMT is trained using parallel data, but due to the “deep learning” carried out by the neural network, the engines are capable of far richer sentence modeling than SMT engines.”
Research and development of NMT have been ongoing since at least 2012, and it was considered “mature” when it became available in 2016 (ibid). Both Smartcat (2019) and memoQ (2020) agree that the demand for NMT is growing, and even though large and medium-sized LSPs have increased the use of MT, “MT engines have reached such a high level, that big tech players like Google, Microsoft, and Amazon don’t have enough incentive to invest in optimizing their MT engines even further” (Nimdzi, 2020, p. 33).
As the use of MT is now well established, the traditional role of translators is transforming. Often, they are required to work as post-editors. According to TAUS’s report (2017, p. 12), “as long as NMT will not fully replace SMT, post-editing will remain essential to track and understand issues in the performance of SMT engines as well as to make the necessary adjustments to improve them.” Translators will need to adapt and develop other skills if they want to keep being competitive.
2.1.2.2.2. Perceptions of Technology
There have been two recent reports that provided interesting insights regarding translation technology: Tirry (2019) and Pielmeier & O’Mara (2020). Both documents analysed the trends and characteristics of the language industry (including translation technology) by means of a survey. Tirry gathered the opinion of over 1400 professionals of the translation and localisation sector from 55 countries and analysed the trends of 2019 in Europe. Pielmeier & O’Mara received over 7300 responses from translators and interpreters from 178 countries and analysed the trends of 2019 globally.
According to Tirry (2019, p. 31), technology-related trainings were the most popular type of training among companies, independent professionals, training institutes and translation departments in 2019. This reality goes in line with the necessity of staying up to date with technological advancements reflected in Pielmeier & O’Mara’s report (2020, p. 38). Additionally, in Tirry’s study (2019, p. 31), the majority of respondents affirmed that master-level graduates are perceived to have either partially developed or sufficiently developed skills of technology usage in general. In contrast, according to the results of a recent study by DeVry University’s Career Advisory Board (2017) about the technological skills gap between curricula and the industry in the USA, organisations seemed to believe that there was a notorious skills gap in technology. Only 11% of organisations considered that graduate institutions provided graduates with skills to meet the current’s tech needs (ibid, p. 1). Moreover, nearly 60% of organisations thought that job applicants usually lacked key technological skills for success (ibid, p. 2).
Nevertheless, according to Pielmeier & O’Mara (2020, p. 38), 51% of respondents considered themselves to be tech savvy and willing to be introduced to new language software. In contrast, only 7% felt rather uncomfortable with regards to technology, and the rest fell somewhere in between (ibid). Attention should be drawn to the fact that the rapid technological advancements acted as a motivation for working with LSPs. Linguists believed that assistance regarding the software used was more difficult to obtain when working with direct clients than when working for LSPs (ibid).
2.1.2.3. Translation Technology within the Translation and Localisation Industry
Translation technology is experiencing a positive trend with regards to development, investment and usage. The 2019 Nimdzi 100 identified 531 different
technology brands devoted to language services commercially available for 2019. This number represents an increase of at least 32.75% compared to the previous year and the available brands are expected to keep growing for 2020 (Nimdzi, 2019b).
According to the answers to Tirry’s survey (2019, pp. 20–21), technology investment plans of LSPs experience a very positive trend on recent and future technology developments and seem to be advancing towards MT and automated workflow.
Moreover, particularly among larger LSPs, the determination to use MT is much stronger than in past years (ibid). Interestingly enough, however, less than 20% of the translation companies report to be using MT frequently (ibid). In Tirry’s results (2019, p. 21) a generally positive trend to invest in CAT tools can be observed as well. While QA tools were only popular with translation companies, subtitling tools, optical character recognition (OCR) and dictation tools scored lowest (ibid).
The relationship between translation technology skills possessed by master-level graduates was also covered in the aforementioned study. In particular, translation technology skills possessed by master-level graduates were scored the lowest byLSPs and translation departments, “despite the stronger cooperation between universities and translation professionals, and the efforts made by translation tool providers” (Tirry, 2019, p. 27). This result may come across as striking since many translation tool providers attempt to bring the professional world closer to academia by offering partnerships that may include lower prices for CAT tools and free certifications for students. For instance, this is the case of memoQ12 and SDL13: the academic partnership of the latter congregates more than 550 universities.14 Because of this, both memoQ’s translation environment and SDL Trados Studio may have been introduced into undergraduate and postgraduate curricula.
The results in Tirry’s survey (2019, p. 22) highlighted remarkable distinctions between the 20 most known tools by independent professionals and companies, and the most used tools. For instance, while at least 90% of respondents knew about Google Translate, MS Office and SDL Studio, at least 80% of companies used MS Office and
12 Available from: https://www.memoq.com/portal/es/programa-academico-de-memoq. Accessed on Apr 26, 2020. The list of institutions adhered to the academic partnership can be accessed from the following link:
https://www.google.com/maps/d/u/0/viewer?mid=12c5mWL_8HzHyhK33HkLZyspD4hs&ll=20.717013 795179337%2C14.457170900000051&z=2
13 Available from: https://www.sdltrados.com/education/. Accessed on Apr 26, 2020.
14 Available from: https://www.sdltrados.com/education/partners/list.html. Accessed on Apr 26, 2020.
>60%, SDL Studio, but only 17% used the NMT engine, Google Translate. At least 80%
of respondents reported knowing the CAT tools memoQ, Wordfast and Déjà Vu, and while memoQ was used by 43% of companies, Wordfast and Déjà Vu were used by
<10%. Similarly, at least 70% knew the CAT tools Across, and Memsource;
>60%, OmegaT; and >50%, Transit, Wordbee, SmartCAT and MateCAT; but only Memsource (>20%), Across (15%) and Transit (>10%) made it to the top 20 most used by companies. Multiterm and Passolo (both under the SDL stable) were known at least by 70% of respondents. In contrast, >30% of companies reported using Multiterm, and only >10% used Passolo. The MT engines Systran and DeepL were known by at least 60% of respondents, but only DeepL was used by >10% of companies. Systran did not make it to the top 20 most used tools by companies. Other widely known tools were Linguee (>80%), Plunet (>50%) and Catalyst (>50%). However, only the first two were used by >20% of companies. The rest of the most used tools were: self-developed tools (>20%), client-specific tools (20%), XTRF (>10%), Translation workspace (<10%), SDLTMS (<10%), SDL Worldserver (<10%), XTM (<10%), and Smartling (<10%).
Similarly to the most used tools by companies reported by Tirry (2019, p. 22), independent professionals mostly used MS Office (>70%), SDL Studio (>40%), Linguee (>40%), Multiterm (>20%), memoQ (<20%), Google Translate (10%), Plunet (10%), Memsource (10%), client-specific tools (10%), DeepL (10%), XTM (<10%), Across (<10%), XTRF (<10%), SDLTMS (<10%), self-developed tools (<10%), Transit (<10%), and a translation workspace (<10%). The only tools more used by independent professionals than by companies were the CAT tools Wordfast (10%), Wordbee (<10%), and OmegaT (<10%). In contrast, the software SDL Worldserver, Smartling and Passolo (all >10%) were reported to be used more frequently by companies than by independent professionals.
The results of translation-related technology usage by linguists in Pielmeier &
O’Mara’s report (2020, pp. 41–42) showed that linguists used TM and CAT tools in most of their projects, followed by quality checkers and terminology management tools.
Nevertheless, despite the positive interest raised by LSPs, the usage of MT scored lowest between linguists (ibid).
Regardless of which software were most known or most used, Tirry (2019, p. 22) concluded that all involved stakeholders used translation software on a day-to-day basis and, therefore, should be proficient when it comes to their usage.
2.1.2.4. Forecasts and Watchlist
The translation and localisation industry is swiftly adapting to give response to emerging businesses, such as the media and game localisation, which are on the point of focus for 2020 (memoQ Translation Technologies, 2020; Nimdzi, 2020, p. 36).
Localisation for IT, MT and machine dubbing are other forecasted sectors to keep growing (ibid).
According to memoQ’s report (2020), the gaming industry alone generated more than USD 120 billion in 2019, consolidating the industry as one of the largest globally, and Nimdzi (2020, p. 36) forecasted that it would reach USD 196 billion by 2022.
Additionally, the entertainment AV sector is expected to keep growing as more streaming platforms, such as Apple and Disney+, introduce their services (memoQ Translation Technologies, 2020). Nimdzi (2020, p. 36) estimated the global revenues of the media industry in 2019 at approximately USD 522.2 billion. This growth in the media and game sectors is expected to have a direct impact on the localisation industry.
Another sector that will keep growing according to Nimdzi (2020, p. 40) is the IT industry, in which “software remains the fastest-growing sector, growing at a rate of 8.3%.” Localisation for the IT sector represented 7% of the total localisation market for all industries (ibid).
Both the trend reports of memoQ (2020) and Nimdzi (2020) agree that the use of MT is also expected to increase in 2020; in particular, NMT. Nimdzi affirms that the IT sector represents the highest demand for MT, on top of e-commerce, Life Sciences and social media sectors (Nimdzi, 2020, p. 40). According to memoQ (2020), “it will affect everything: business, process, technology, people, quality management, decisions on rates, and the kinds of professional roles needed to do this work well. […] The human role will be more important than ever, with new skills required for working with NMT.”
Nimdzi (2020, p. 41) also affirms that machine dubbing, in the area of natural language processing (NLP), is expected to become centre stage in 2020. In addition to the fact that several companies, such as Amazon AI, Evernote, or Omniscien, are currently working on ways to enhance automatic dubbing (ibid):
“[…] modern neural TTS (text-to-speech) engines from Google and IBM are becoming good enough – to the point where users stop noticing any imperfections.These and similar developments suggest that, in the near future,
neural TTS technologies will be available in a wider variety of languages, and we can expect to see more companies offering auto-dubbing tools.”
Moreover, TAUS’s report The Translation Industry in 2022 (2017, pp. 13–23) provides a broader overview of the industry and predicts six drivers of change: Machine Learning (ML), MT (discussed in Section 2.1.2.2.1.), Quality Management, Data, Interoperability and Academy (human resources and training). ML went mainstream in 2016 and TAUS predicted it to have “a great impact on every organization’s strategic planning as it keeps showing promise in delivering a high degree of competitive advantage over the next five to ten years” (ibid, p. 14). TAUS foresees ML to make data- driven technologies “astonishingly disruptive”, which might also be the case of MT (ibid).
ML will be applied to MT, translation quality assessment and translation data, and will transform these areas completely (ibid, p. 15–19). The translation industry “has always suffered from a lack of interoperability, […which] has been costing a fortune for years, both on the client side (in translation budgets) and on the vendor side (in revenues)” (ibid, p. 19). TAUS expects technology vendors to fulfil this gap and perhaps introduce their own solutions on the market (ibid). Lastly, the sixth driver of change according to TAUS is academia, which should reflect the changes in the industry (ibid, p. 22).
TAUS published another report (2016) which laid out in trends in the translation and localisation industry. These included cloud computing, S2S, Deep Learning (DL) technology and the aforementioned translation data, and quality assessment (ibid, p. 75–
88).
Regarding technology to keep an eye on in the foreseeable future, Pielmeier &
O’Mara (2020) highlighted Automated Content Enrichment (ACE). ACE parses source and target text and scans the content to identify concepts. The software then generates links to relevant external resources (e.g. articles, images or audio media) and supplies further information such as terms, places, products or dates. Since the content is more interactive and intelligent, the linguist may then use the links to research pertinent information without having to leave the translation environment, thus saving research time. According to the results of Pielmeier & O’Mara’s survey (2020, p. 45), even if this technology holds promise for the translation activity, most linguists were not aware of its existence and only a few had the opportunity to try it out.
2.2. Developments in Translation Competence Teaching
As seen in the previous section, translation technology has made a significant impact on the translation industry, which demanded stakeholders to keep abreast with the ever-changing technological side of the profession by learning new tools on their own (Pielmeier & O’Mara, 2020, p. 38). The technologization of the profession forced educational institutions to modify the contents taught in undergraduate and postgraduate programmes to include translation technologies. This section will discuss what is understood by translation competence and will examine four of the most influential translation competence models. PACTE’s translation competence model will be described first, followed by Göpferich’s and Kiraly’s. To finish with, the EMT Framework’s proposal of translation competence model will be reviewed.
2.2.1. Translation Competence
As laid out by Presas in her article Bilingual Competence and Translation Competence (2000, p. 19), “for many authors a bilingual is a natural translator, because the bilingual, in addition to acquiring competence in both languages also acquires the ability to translate from one language to the other.” However, even though linguistic competence is crucial for the translation activity, it “is not in itself sufficient to guarantee translation competence” (ibid). For Neubert (2000, p. 3), “translation involves variable tasks that make specific demands on the cognitive system of the translator [and] what enables translators to cope with these tasks is their translational competence.” According to the aforementioned scholar, translation competence entails complexity, heterogeneity, approximation, open-endedness, creativity, situationality and historicity; and all of these features are intertwined with each other (ibid, p. 5). Both scholars agree that translation competence is much more than linguistic knowledge alone.
In simple words, translation competence is the system of knowledge needed for translation (Presas, 2000, p. 28). The translation process requires broad knowledge and skills: “knowledge of the two languages, knowledge of the real world and of the material, the ability to use tools such as dictionaries and other sources of documentation, cognitive qualities such as creativity and attention, or the capacity to resolve specific problems”
(ibid). This heterogeneity of kinds of knowledge has resulted in several translation
scholars proposing various models of translator competence (Bell (1991), Campbell (1991), Cnyrim et al. (2013), Eser (2015), Hönig (1991), Shreve (1997), amongst others).
2.2.2. Translation Competence Models
As stated above, several models of translator competence have been presented over the past decades and still influence translator training approaches adopted at higher education institutions. In this thesis, four of the most influential ones (PACTE’s, Göpferich’s, Kiraly’s and the EMT Framework’s Translation Competence Model) will be described. For the purposes of this research study, the Translation Competence Model proposed by the EMT Framework has been taken as the main reference, as explained in Section 3.2.2.
O’Brien & Rodríguez Vázquez (2019) have recently published an article titled Translation and Technology included in The Routledge Handbook of Translation and Education that reviewed the four aforementioned translation competence models (amongst others) from the perspective of translation technology. After the review of said translator competence models and others, they concluded that competence in using translation tools is perceived as a core element (ibid, p. 7). Furthermore, in their opinion, other aspects are included within the translation technology competence. These are, for instance, knowledge concerning the market and the deployment of tools, and other personal aspects, such as perseverance and critical spirit (ibid). These aspects are further endorsed by the EMT Framework translation competence model (ibid). In view of the relevance of their publication, O’Brien & Rodríguez Vázquez’s views will be included in this section in order to see how these four translation competence models understand translator competence and how translation technology, in particular, comes into play.
2.2.2.1. PACTE’s Translation Competence Model
The Process in the Acquisition of Translation Competence and Evaluation research group from the Facultat de Traducció i Interpretació of the Universitat Autònoma de Barcelona, Spain, commonly known as PACTE, has been investigating the acquisition of translation competence in inverse and direct written translation since it was established in 1997. Since then, PACTE has published about twenty articles reflecting the outcomes of their research on how to improve translation programmes and their evaluation. Their goal is to build both a model that defines translator competence and a model that lays out