Co-written by Gust Burns, Kevin deVoss, Katie King, Sumyat Thu
Rise of Translation Apps in Education
At the Praxis 2018 conference, Gust Burns, Kevin deVoss, and Katie King facilitated a roundtable discussion based on their research on the implications of the rise of using translation apps in higher education. Burns discussed a Marxist critique of machine translation app like Stepes. In 2015, the multinational localization company CSOFT debuted Stepes, the first on-demand and just-in-time human translation platform. Self-described as the “Uber of translation,” the platform contracts individual bilinguals to translate texts on-demand for business customers. As the Stepes website puts it: “By seamlessly connecting businesses and individuals with the world’s linguists through our on-demand translation eco-system, we help break down language barriers between businesses and their customers, and among nations and people.” In Burns’ words, we can understand the mechanics of Stepes, along with both machine translation in general, by thinking with Marx in the Grundrisse, about machines, as manifestations of fixed capital that function not to “transmit the worker’s activity to the object,” as we normally think about labor in early industrial and even Fordist production, but rather so that “the machine’s work, the machine’s action” is maintained and supplemented by the human activity, the latter which becomes “a mere living accessory of this machinery” (692, 693). In other words, as human activity becomes subservient to the active determining functionality of machinery, human capacity is increasingly inseparable from its position within machines, it becomes fully subsumed, in a real way, under capital.
In deVoss’s work on computer-assisted language learning (CALL), he mentions much of pedagogical literature on CALL seems to focus on questions of “best practices,” and ways to implement particular tech products into the classroom. But deVoss points out that what is missing is sufficient critical inquiry into these products, and of the ideological assumptions behind them as they spread into the wider academic ecology. Tech commodification has certainly touched postsecondary English language education significantly. For instance, Duolingo, which is probably the most popular language learning application currently on the market, has already set up English placement testing services within a lot of well-known universities. Many of the most popular massive open online courses (or MOOCs) through for-profit services like Coursera and edX, and who sponsor coursework at many American universities, have developed a lot of English language learning content both domestically and internationally.
Using discourse analysis methods, deVoss looked at media statements made by the creators of these kinds of tech-ed companies and found that the word “disruption” kept recurring with regard to the higher education system. “Disruption” is a kind of loosely-defined buzzword within Silicon Valley and associated with a variety of meanings, including: “innovation,” “marketization,” and even “destruction.” Of course, in these accounts, tech-ed industry figures unanimously viewed disruption as a necessary, positive project for the higher education system. But deVoss also finds that the “disruption” of English language education in higher ed was associated with a lot of troubling trends, including: the introduction of more student fees, higher course costs, low course completion rates, and disinvestment in faculty and academic student employment and compensation.
In King’s research on Google Translate as a tool for literary translation, she argues that the process for “training” an AI algorithm is not only something we humanists can understand but also should understand, and teach. A machine translation tool is just that, a tool that has been created by human beings, and not all the available tools perform equally. Google Translate not only did not perform well translating a complex literary paragraph from Spanish into English, it also did not learn from King’s feedback over time. But King found other researchers in Europe who have experimented with building their machine translation tool specifically trained to translate literature between French, German and English. Their results were limited but much better. King’s advice to researchers and instructors is to inform themselves for the benefit of their own knowledge about these evolving tools, and to introduce to their students the concept that they must check the results of translation tools carefully.
Translation Pedagogy in Writing Classroom
While teachers should be aware of the possible downsides of using translation technology as part of their pedagogy, encouraging multilingual students to use their human translation capacity has been found to be a helpful writing pedagogy. For example, Eun-Young Kim, an English language and second-language writing scholar at Southwestern Adventist University, asks her Korean and English speaking multilingual students to do translation exercises in order to raise students’ metacognitive awareness about how they write and things they may have missed in their writing process. In the first exercise, students were asked to read the final versions of their English essay writing and translate their own writing into Korean. Kim emphasizes that they should translate their writing as if it had been written by someone else. After finishing the translation, the students were asked to hand in a written response describing what they felt and learnt while translating their own writing into Korean.
Some students reflect:
“After I translated my writing into Korean, I discovered a lot of mistakes. There were many unconnected sentences. I think translating our own writing into Korean is important in order to measure our true writing skill.”
“Before I translated my writing into Korea, I had thought that my writing was pretty good. But when I tried to translate it, I was surprised that my expression was inadequate and the content was insufficient. Some of my sentences were not exactly what I meant.”
“I had to revise my writing four times while I was translating. Some sentences are smoothly connected, but others were a little too forced. I also realized that I missed some details because I was concentrating on my English. I like my introduction better this time. I could discover my mistakes more easily when I translated my composition into Korean.” (Please see the rest of the quoted students’ reflections on the article here).
It’s a great result for students’ self-assessment through translation that students are reporting they could have had better sentence transitions, logical connections, and more details in the writing. This level of self-assessment seems to engage students in a process of possible substantial revision that goes beyond merely checking for lower-order concerns such as conforming to Standard Edited English grammar or mechanics.
In the second exercise, Kim asks her students to work with a peer partner and translate their partner’s English essay into Korean. Students were then asked to respond to their partner’s translation to see how accurate the translated essay was. Some students respond:
“I was very much afraid if my composition would clearly deliver my message to my partner. When I read my partner’s translation of my composition, I found out that she understood my composition differently from what I meant in the writing. I feel that my English was unclear; that’s why my partner didn’t understand it correctly …”
“The translated text is not what I meant. I wish my partner had translated my writing better. Professor [addressing me], I added some words to the translation in order to make my meaning complete. I think this [miscommunication] is because my composition is not clear.”
Though the second exercise is more complicated with the possible room for misreading and misinterpretation by a peer-partner, both of the exercises indicate that when multilingual students engage in translation between their languages, it heightens their rhetorical awareness and self-assessment ability to more clearly see where the students might have fallen short in composing their text and the different choices they can make because of this awareness. Multilingual students might also begin to critically reflect on cultural rhetorics and contexts behind using a particular language which might help them become sharper language users and writers.
Eun-Young Kim, “Using translation exercises in the communicative EFL writing classroom,” ELT Journal, Volume 65, Issue 2, April 2011, Pages 154–160.