L2 Learner Use of Machine Translation:�Using What We Know to Harness the �Pedagogical Benefits of (Automated) Translation��

Dr. Luciane Maimone

Dr. Jason Jolley

Missouri State University

IALLT Webinar Overview

• About us
• The monsters at the door (past, present, future)
• Translation and L1 in the language classroom
• What we know about learner use of MT
• Harnessing MT for pedagogical benefits – A framework
• From MT to Generative AI?
• Q&A

​

​

2

About Us

Dr. Luciane Maimone

• Ph.D. in Hispanic Linguistics from Georgetown University
• Associate professor of Applied Linguistics at Missouri State University
• Graduate program director for Master of Applied Second Language Acquisition program
• Co-editor of Portuguese Language Journal and AOTP board member
• Research interests: L2 Portuguese acquisition, language assessment, machine translation

Dr. Jason Jolley

• Ph.D. in Spanish from Penn State University
• Associate Dean of the College of Arts, Social Sciences and Humanities and professor of Spanish at Missouri State University
• Treasurer of the Mid America Chapter of the American Translators Association
• Research interests: Latin American literature, machine translation, self-directed language learning

​

​

3

The Monsters at the Door

Past: Reliance on L1 and conventional translation�(excesses of the GTM)

Present: Machine translation (e.g., Google Translate)

Future: Generative AI�(e.g., ChatGPT, Claude AI)

4

Translation and L1 in the L2 Classroom

5

G. Cook (2010)

Kerr (2014)

See also:

Cook, V. (2001). Using the first language in the classroom. The Canadian Modern Language Review, 57(3), 402-23.

Vermes, A. (2010). Translation in foreign language teaching: A brief overview of pros and cons. Eger Journal of English Studies, 10, 83-93.

Vinall, K., & Hellmich, E. (2022). Do you speak translate? Reflections on the nature and role of translation. L2 Journal, 14(1), 4-25.

​

​

“The is little doubt that translation and contrastive analysis are not just an alternative but an indispensable tool for developing communicative competence, in which meanings are negotiated not just within one language but also across languages and cultures. Translation is increasingly coming to be seen as a natural and necessary competence in its own right. In this respect, a case has been argued in favor of introducing translation as a viable pedagogic tool at the earliest levels of SL instruction.” (Bratož & Kocbek, 2013)

Machine Translation: Student Use

The takeaways from numerous articles exploring how often students use MT, what they use it for, and why they use it are that…

• Almost all students use MT, and high percentages use it frequently, especially on writing tasks.
• Learners report using MT most frequently to translate individual words or short phrases, as opposed to longer segments (e.g., entire sentences, paragraphs).
• Some studies suggest that lower proficiency learners may translate longer segments more frequently.
• Students also use MT to double-check usage and to aid L2 comprehension.
• Students say they use MT because it’s convenient and fast.
• Psycholinguistic factors, such as low confidence and anxiety provoked by high-stakes L2 writing tasks, may be another reason some students turn to MT.
• Evidence suggests students use MT more when tasks are graded (high-stakes).

​

6

See, for example, Clifford et al. (2013), Ata et al. (2021), Farzi (2016), Hellmich & Vinall (2023), Jolley& Maimone (2015), Larson-Guenette (2013), Merschel & Munné (2022), O’Neill (2019), Tight (2017).

​

Harnessing the Benefits of MT Use

7

“It will be important for educators across the upper elementary, secondary, and postsecondary spectrum to understand how students are using translation tools and to teach them to do so in a responsible way that promotes, rather than circumvents, their progress toward more sophisticated language proficiency.”

Ducar, C., & Schocket, D. H. (2018). Machine translation and the L2 classroom: Pedagogical solutions for making peace with Google Translate. Foreign Language Annals, 51(4), 779-795.

​

Pedagogical Recommendations: New Directions

Recommendations for addressing MT use in the classroom have shifted from:

Discouraging the use of MT and penalizing unauthorized uses to

8

Learning about the capabilities of MT tools, modeling appropriate uses, and integrating it into instruction

Moving from Reactive Practices to a Curriculum Design Approach

A reactive approach manages issues once they emerge or are encountered in the classroom. It may involve unplanned interventions and immediate solutions to simple problems or adaptations of class materials and activities to address issues not grounded in the curriculum.

While reacting to students' needs and to unexpected issues is a common and necessary strategy for successful classroom management, the widespread availability of MT and AI tools and their frequent use by learners demand more than reactive instructional practices. Considering that our students already have and will always have access to these rapidly developing technologies, we really need to be including them in our language curriculum design in deliberate, principled ways.

9

Support for MT curriculum integration can be found in Carré et al. (2022), Knowles (2022), Pellet and Meyers (2022), Urlaub and Dessein (2022), Sugiyama and Yamanaka (2023), among others.

From the Literature: Recommendations

1. L2 writing instruction should emphasize technology integration, explicit writing strategies, and comprehensibility and pragmatic appropriateness (above grammatical accuracy).
2. Instructors should help learners understand that language development, rather than perfection, is the goal.
3. Course designs should include more formative, low-stakes, and ungraded chances for feedback.
4. Reading assignments should ensure true engagement beyond simple comprehension.
5. Assignments and assessments should be aligned with learners’ production abilities.
6. Instructors should expand performance assessment beyond presentational writing.

​

10

Curriculum Design in the Era of MT / AI

From the Literature: Recommendations (cont.)

11

7. Instructors should learn more about MT capabilities and principled uses of L1 and translation.
8. MT/digital literacy should be an emphasis of language curricula at all levels.
9. MT use policies should specify allowed and prohibited uses, means of substantiating allegations of cheating, consequences/sanctions, and appeals procedures.
10. Instructors should refocus attention on the root causes of cheating (as opposed to just detection).

Curriculum Design in the Era of MT / AI

Integrated Approach to MT and Instructed Language Learning

Three components

1. Instructor technology literacy
2. MT Integration into curriculum design
3. Evaluation

​

12

Backward Design Steps

• Definition of MT literacy objectives
• Definition of task pedagogical purposes
• Definition of task types
• Design of classroom and online/home assignments
• Design of assessments (MT-assisted or MT-deterrent)

​

Learner/Central Design Steps

• Digital literacy diagnostic
• Co-constructed goals and criteria for MT use
• Formal communication of classroom policies and agreed upon criteria (e.g., syllabus)
• Digital literacy training

​

MT Literacy Objectives

Refers to knowledge, skills, and critical awareness that students need to ethically engage with machine translation tools in language learning contexts

13

Beginners

• Equip students with a basic understanding of how MT systems function, along with their strengths and weaknesses
• Students will understand ethical boundaries of MT use (cases in which it would and would not be appropriate)

Intermediate/Advanced

• Students will develop critical evaluation skills for MT output
• Students will understand ways in which MT can aid their independent or self-directed language learning goals

​

MT literacy objectives should be appropriate for the learners' proficiency level and learning context

Conceptualizing Tasks

Task Purpose

• Feedback
• Pronunciation practice
• Vocabulary learning/comprehension
• Metalinguistic awareness
• Developing L2 writing skills
• Improving grammar control
• Domain-specific MT training (medical, legal, etc.)
• Developing cultural competence

​

14

​

Task Types

• Pre-editing
• Post-editing
• Text comparison – MT output vs. L2 direct writing OR human translation
• Pronunciation practice
• Hypothesis testing
• Grammatical analysis
• Error detection and/or correction

Task Design

• Consider modality
• Consider cognitive demands based on the learners' proficiency level

From the Literature: Pedagogical Uses

Most activities proposed in the literature focus on using pre- and post-editing tasks designed to raise metalinguistic awareness, enhance control of grammar structures, and improve L2 writing and translations abilities

15

• Analysis of raw MT output or comparison of HT and MT outputs to facilitate awareness of differences between and features of L1 and L2 (contrastive analysis)
• Analysis of raw MT outputs with a focus on error detection and correction
• Pre-editing of MT input so that output matches HT-generated “model texts” or direct L1 writing ideals to improve L2 writing performance (using the MT as a real-time feedback generator)
• Post-editing of raw MT output to enhance metalinguistic awareness, to practice L2 grammar, and to improve L2 writing performance
• Post-editing of MT output resulting from instructor-manipulated input (allows instructor to “plant” errors highlighting certain structures)
• A handful of studies have also looked at using MT to support other skills, such as reading and listening comprehension, and vocabulary acquisition.

Assessments w/ MT in Mind

MT-assisted

• Better suited for formative assessments
• MT may be allowed in one of the assessment's components (e.g., task planning, research, brainstorming phase, vocabulary checking, etc.)
• MT integrated to revision processes

​

16

MT-deterrent

• In-class writing tasks/Browser lockdown online tests
• "Smart assignments" (that make difficult to use MT)
• tasks that go beyond content (e.g., defining # of words and specific conditions and grammar points, etc.)
• penalize overreliance on MT and reward originality, creativity, complexity, and depth of language use
• Process-oriented assignments (that remove incentives to cheat)
• not a high-stakes assessment
• adequate for the learner's level
• reflect text types practiced enough in class
• not focused on accuracy or perfection
• students are not rushed and can use their notes
• focus on functional uses of language or on discourse elements (e.g., text organization, cohesion, coherence, clause relations)
• submission of multiple drafts in which students respond questions about linguistic choices

​

​

(see Ducar & Schocket, 2018; Knowles, 2022; McCarthy, 2004; Steding, 2009; etc.)

Digital Literacy Diagnostic

17

Examples

• Guided discussions about how students use MT, for what purposes, and how much they know about web tool capabilities
• Groups discussions (combined or not with sample tasks) on the effectivity of MT for different types of purposes and text types
• Simple questionnaires asking students how, when, and what for they use MT and their perceptions about the usefulness and ethicality of MT use in educational settings
• Polls about students' interests regarding MT tools uses
• Non-graded quiz on ethical aspects of MT use followed by discussion
• Diagnostic tasks designed for students to freely try different strategies accomplishing learning goals with the assistance of MT tools, followed by discussion and self assessment
• Critical analysis of MT outputs followed discussion and co-constructed MT learning objectives

​

​

Co-constructed MT Goals and Communication

Seated classes

• Discussions about authorized uses based on their benefits to L2 learning (L1 or L2)
• Reflections on the use of MT (L1)

​

Online classes

• Online forum discussions (L1)
• Online polls (L1)
• Reflections on the use of MT (L1)

18

Communicating policies and agreed upon rules formally

• Classroom posters
• Syllabus policies
• Content available in Learning Management Systems

What to include?

• Appropriate an inappropriate uses of MT
• Process for establishing and appealing violations
• Institutional policies for violations of academic integrity

​

See Correa (2014), Jolley & Maimone (2015, 2023); Loyet (2018), Mundt & Groves (2016), Somers et al. (2006), Steding (2009)

​

Digital Literacy Training

Digital literacy training activities should be integrated to curriculum planning, but instructors should also entertain learner needs and interests that are identified through classroom diagnostic

• Need for (general) digital literacy (Hubbard, 2004; Peters & Frankoff, 2014; Williams et al., 2014)
• Argue in favor of MT literacy in language courses (Bahri & Mahadi, 2016; Bowker & Buitrago Ciro, 2019; Goldwing-Jones, 2015, 2022; Groves & Mundt, 2015; Knowles, 2022; O'Brien & Ehrensberger-Dow, 2020; Stapleton & Leung, 2019 )
• Report that students are willing to receive MT training (Chan, 2017)

​

​

19

MT Literacy Workshop (Bowker, 2020)

1. Privacy/confidentiality
2. Academic integrity/ Policies
3. Potential for algorithmic bias
4. Awareness of different tools
5. Different translation tasks (uses of MT)
6. Pre-editing strategies

​

​

​

​

Simply using MT is easy. Using it effectively, being able to assess whether, when, and how to use it, requires critical thinking and technical skills (Bowker, 2020, p.28)

20

The Machine Translation Literacy Project and Infographics (Bowker, 2022) https://sites.google.com/view/machinetranslationliteracy/home/teaching-resources

Online Resources

https://sites.google.com/view/machinetranslationliteracy/home/research-output?authuser=0Online

From MT to Gen AI…

• MT and Gen AI platforms use different machine learning models and have different strengths
• We still lack research (data) comparing the effectiveness of MT and other Gen AI models in translation tasks
• Google Translate and DeepL built specifically for translation in general still perform translation tasks with greater accuracy compared to free versions of web-based Gen AIs
• Most likely, Gen AI models will soon be able to perform the same tasks carried out by DeepL MT tools with the same accuracy level
• Due to their specialized and narrower focus, MTs like Google Translate can be easier to use in the classroom (avoiding distraction)
• The usefulness of Gen AI outputs depend on a higher level of digital literacy and the ability to create adequate prompts
• Gen AI platforms are designed for creative, interactive tasks. For example, Gen AI chatbots can offer personalized feedback and tailored suggestions
• Curriculum considerations and tasks designed for MT tools can be adapted to Gen AI
• Ultimately, the purpose of the activity or task should determine the tool

​

21

Q&A

What else would you like to talk about?

Thanks for attending! Feel free to reach out to us!

lucianemaimone@missouristate.edu

jasonjolley@missouristate.edu

22

References

Ata, M., & Debreli, E. (2021). Machine translation in the language classroom: Turkish EFL learners’ and instructors’ perceptions and use. IAFOR Journal of Education: Technology in Education, 9(4), 103-122.

Bahri, H., and T. S. T. Mahadi. (2016). Google Translate as A Supplementary Tool for Learning Malay: A Case Study at University Sains Malaysia. Advances in Language and Literary Studies, 7 (3),161–167. https://10.7575/aiac.alls.v.7n.3p.161

Bowker, L., & Buitrago Ciro, J. (2019). Machine translation and global research: Towards improved machine translation literacy in the scholarly community. Bingley, UK: Emerald Publishing

Bratož, S., & Kocbek, A. (2013). Resurrecting translation in SLT: A focus on young learners. In D. Tsagari & G. Floros (Eds.), Translation in language teaching and assessment (pp. 135–153). Cambridge Scholars Publishing.

Clifford, J., Merschel, L., & Munné, J. (2013). Surveying the landscape: What is the role of machine translation in language learning? @tic Revista d’Innovació Educativa, 10, 108-121.

Cook, G. (2010). Translation in language teaching. Oxford University Press.

Cook, V. (2001). Using the first language in the classroom. The Canadian Modern Language Review, 57(3), 402-23.

Eisenstadt, M., & Vincent, T. (2000). The knowledge web: Learning and collaborating on the Net, UK: Kogan Page.

​

23

References

Farzi, R. (2016). Taming translation technology for L2 writing: Documenting the use of free online translation tools by ESL Students in a writing course [Unpublished doctoral dissertation]. University of Ottawa.

Hellmich, E. A., & Vinall, K. (2023). Student use and instructor beliefs: Machine translation in language education. Language Learning & Technology, 27(1), 1-27.

Hirumi, A. (2013). Three levels of planned elearning interactions: A framework for grounding research and the design of elearning programs. Quarterly Review of Distance Education, 14(1), 1-16.

Jolley, J. & Maimone, L. (2015). Free online machine translation: Use and perceptions by Spanish students and instructors. In A. J. Moeller (Ed.), Learn Languages, Explore Cultures, Transform Lives (pp. 181-200). Central States Conference on the Teaching of Foreign Languages.

Kerr, P. (2014). Translation and own-language activities. Cambridge University Press.

Larson-Guenette, J. (2013). “It’s just reflex now”: German language learners’ use of online resources. Die Unterrichtspraxis, 46(1), 62-74.

Loyet, D. (2018). Is Machine Translation a Threat to Language Learning? The Chronicle of Higher Education, April 13, B23

Merschel, L., & Munné, J. (2022). Zooming in on machine translation use in L2 online classes: Reflecting on the future of L2 writing. The FLTMAG.

24

References

Merschel, L., & Munné, J. (2022). Zooming in on machine translation use in L2 online classes: Reflecting on the future of L2 writing. The FLTMAG.

O’Brien, S., M. Simard, & M.-J. Goulet (2018). Machine Translation and Self-post-editing for Academic Writing Support: Quality Explorations. In J. Moorkens, S. Castilho, F. Gaspari, and S. Doherty (Eds.), Translation Quality Assessment: From Principles to Practice (pp.237–262). Springer. https://10.1007/978-3-319-91241-7_11.

O’Neill, E. M. (2019). Online translator, dictionary, and search engine use among L2 students. Computer-Assisted Language Learning-Electronic Journal, 20(1), 154-177.

Tight, D.G. (2017). Tool usage and effectiveness among L2 Spanish computer writers.” Estudios de Lingüística Inglesa Aplicada, 17, 157-182.

Vermes, A. (2010). Translation in foreign language teaching: A brief overview of pros and cons. Eger Journal of English Studies, 10, 83-93.

Vinall, K., & Hellmich, E. (2022). Do you speak translate? Reflections on the nature and role of translation. L2 Journal, 14(1), 4-25.

​

25