Web of Science (WoS) Core Collection data based was used in our study. We set the search query as: TS = “translation training” OR TS = “translator training” OR TS = “interpreting training” OR TS = “interpreter training” OR TS = “translation teaching” OR TS = “interpreting teaching” OR TS = “translation curriculum” OR TS = “translation classrooms” OR TS = “translation students” OR TS = “translation teachers” OR TS = “translation pedagogy” OR TS = “interpreting pedagogy”; language is English; document type is Article or Review; Time ranges from 2011 to 2020. 652 records were retrieved.

In this study the number of publications per year, co-authorship network, network of co-authors’ institutions, network of co-authors’ countries, co-occurring keywords, document co-citation network, author co-citation network, and journal co-citation network were analyzed. The tools used were CiteSpace and VOSViewer. In this study, we reclassified Taiwan, Hong Kong and Macao into China, England, Scotland and Wales into UK.

From 2011 to 2020, there were in total of 652 publications. As shown in Figure 1, the number of publications fluctuated. The number reached a low point in 2013 with 25 contributions, and a high point in 2019 with 105 contributions. In general, the number of publications per year increases with fluctuation, which indicated a continuous interest in the field of translation education.

We analyzed the scientific collaboration network at three levels: author, institution, and country. We used VOSViewer to do the analysis. 1051 authors contributed in the field of translation education. We only chose authors with more

than 3 publications. After the program finished, we got the co-authorship network shown in Figure 2. The larger the node was, the more publications an author contributed. Different colors meant these authors were in different clusters. From Figure 2, we could see there were few links among authors, which meant weak collaborations among authors. There were three major clusters. The first one consisted of Nour, Soudabeh, Stengers, Helene and Struys, Esli. The second one included Galan-Manas, Anabel, Hurtado Albir, Amparo and Olalla-Soler, Christian. The last one consisted of Pan, Jun, Wang, Honghua and Yan, Jackie Xiu. In terms of the number of publications, shown in Table 1, Olalla-Soler, Christian ranked the first with 8 documents. Dong, Yanping was the second with 7 documents. Lee, Jieun was the third with 6 documents.

The next was to analyze the co-authorship network at the institution level. There were in total 558 institutions involved. We set the minimum document as 5, and 40 met the threshold. Density visualization was shown in Figure 3. The more documents were, the closer the color was to yellow. Together with Table 2, Univ Autonoma Barcelona ranked the first with 22 documents, Guangdong Univ Foreign Studies the second with 16 documents, Islamic Azad Univ the third with 15 documents. From the results we could conclude that the above three universities focus on the research of translation education.

Thirdly we analyzed the country co-authorship network. There were in total 66 countries contributing to the field of translation education. We set the minimum number of documents as 5, then 33 met the threshold, shown in Figure 4. The larger the node was, the more documents a country contributed. The same color indicated the countries belong to the same cluster. From Figure 4, we can see China, Spain, the UK and the USA contribute more. China collaborated with UK, Japan, Ireland, and Australia. Spain collaborated with Russia, Saudi Arabia, Slovenia and Switzerland. The USA collaborated with Jordan and Ukraine.

A total of 2579 different keywords were used in 652 articles in the field of

translation education. We excluded the “translation training”, “translator training”, “interpreting training”, “interpreter training”, “translation teaching”, “interpreting teaching”, “translation curriculum”, “translation classrooms”, “translation students”, “translation teachers”, “translation pedagogy” and “interpreting pedagogy”, since they were the keywords used in the search query. In the co-occurring analysis of keywords, we set the minimum occurrence as 5, then 98 met the threshold. These keywords formed 7 clusters. The cluster of keywords can help us find the relative topics.

As seen in Figure 5, the largest cluster was in red with 22 keywords, such as communication, education, consecutive interpreting, ethics, and impact. The second cluster was in green with 18 keywords, including language, interpreting, working memory, sight translation and individual differences. The third was in blue with 17 keywords, such as translation competence, intercultural competence, competence, legal translation and audiovisual translation. In terms of total link strength larger than 30, translation, English were in the yellow cluster; language, working memory, interpreting, cognitive control were in the green cluster; competence was in the dark blue cluster; performance, students were in the light blue cluster; machine translation and post-editing were in the purple cluster.

These findings showed that the importance of translation education lies in translation, language, working memory, competence, interpreting, English, performance, machine translation, students, cognitive control, and post-editing.

The overlay visualization map of keyword co-occurrence was shown in Figure 6. A color bar was shown in the bottom of the map indicating different years. The colors ranged from blue (2011) to green to yellow (2020). Therefore, Figure 6 can help us identify the trend research topics. From Figure 6, we can see in the field of translation education, the keywords such as competence, machine translation, working memory, post-editing, eye-tracking, individual-differences, experience, interference, computer-assisted translation, feedback, CAT tools, directionality, and medical translation have stood out in recent years. This result suggested that scholars have to research on the mentioned topics. In order to strengthen this result, co-occurring keyword analysis was conducted using CiteSpace. Keyword timeline view map was generated in Figure 7. We mainly focused on the year 2019 and 2020. We can see in these two years, deep learning, directionality, explicit teaching, pronunciation, translation quality, medical translation, eye-tracking, simultaneous interpreting, reliability, system, literacy, neural machine translation, translation process research, program, market, terminology, academic writing, corpus analysis, 2^nd language proficiency, language control, interference, physical ergonomics, educational ergonomics, validation and Google translate have become the new topics.

We conducted the author co-citation analysis to find the most influential ones. It was observed that 108 authors met the threshold of at least 20 citations. The top 15 authors with highest TLS were listed in Table 3. Kiraly, Don was the most influential author in terms of both citations and TLS. Pym, Anthony ranked the

second and Kelly, Dorothy the third in terms of both citations and TLS. In terms of TLS, Beeby, Allison ranked the fourth, while in terms of citations Gile, Daniel was the fourth. Among the 15 most influential authors, four was in Spain; two were in Austria. This result can cast light on researchers newly into the field of translation education.

Among 10051 publications, 36 stood out with at least 50 citations. Top 15 with strongest TLS were listed in Table 4. From Table 4, we can see in terms of TLS, Interpreter and Translator Trainer ranked the first. While in terms of total citations, the first was Meta. Among the 15 publications, 13 were journals, accounting for 87%. Only 2 were books. If researchers want to publish their articles, Interpreter and Translator Trainer and Meta could be their good choices.

To recognize core references in the field of translation education, it is necessary to focus on the frequency of citations. We used CiteSpace to detect the most cited references. As shown in Table 5, the top 1 most cited reference was that of Amparo Hurtado Albir “The acquisition of translation competence: competences, tasks, and assessment in translator training” [16] . In this paper, examples of how to design translator training curriculums were provided. The second was that of Laura Babcock “Are simultaneous interpreters expert bilinguals, unique bilinguals, or both?” [17] . This paper found interpreters did not show bilingual benefits as previous studies suggested, but they appeared to possess larger verbal and spatial memory spans. The third was that of Federico Gaspari “A survey of machine translation competencies: Insights for translation technology educators and practitioners” [18] .

Among the 15 references, four focused on interpreting. Babcock and Vallesi [17] found interpreters did not show bilingual benefits as previous studies suggested, but they appeared to possess larger verbal and spatial memory spans. Their paper was from a cognitive way. Cai, Dong, Zhao and Lin [19] investigated what competencies impacted beginner student interpreters most in consecutive interpreting. Morales, Padilla, Gomez-Ariza and Bajo [20] conducted experiments to show the transfer of simultaneous interpreting experience was “specific to the cognitive processes more closely involved in the interpreting tasks”. Davitti [21] studied dialogue interpreting.

Three focused on machine translation. Gaspari, Almaghout and Doherty [18] conducted surveys to show “an improvement in quality assessment methods, tools, and training” were strongly needed. Kenny and Doherty [22] argued for “a holistic approach to the integration of Statistical Machine Translation into translator training programs”. Pym [23] identified the skills required for translation work in the age of integration of data from statistical machine translation into translation memory suites.