Patrick J. Burns

Associate Research Scholar, Digital Projects @ Institute for the Study of the Ancient World / NYU | Formerly Culture Cognition, and Coevolution Lab (Harvard) & Quantitative Criticism Lab (UT-Austin) | Fordham PhD, Classics | LatinCy developer

Latin in an Environment of Infinite Extensive Reading

Abstract for workshop presentation given at CANE2024

Abstract

That emergent Latin readers can become stronger Latin readers by reading more Latin—magnitudes more—is the central pedagogical principle behind extensive reading (Day and Bamford 1998; also, Cooper 2022; Olimpi 2019). Such an approach, drawing on research and practices in modern secondary language acquisition (Carlon 2013), promotes open-ended, voluntary reading, and specifically reading that is level-appropriate and “comprehensible” (M. Patrick 2019; R. Patrick 2015; Venditti 2021; Rogers 2019), as an important supplementary classroom activity alongside other language instruction (Piantaggini 2019). At the same time, in contrast to modern languages (VanPatten 2014), the use of extensive reading in the Latin has long been constrained by the total amount of material available, as well as the even more limited amounts once that toal is split by subject variety, level-appropriateness, and related factors. Even with the surge of Latin novellas (Hunt 2022) and other student-oriented Latin content (Gruber-Miller 2023), it is still the case that, as one researcher recently noted (Olimpi 2019), “scarcity of available texts is the largest obstacle” in adapting extensive-reading practices to the Latin classroom. Large language models (LLMs) like GPT-4—and its popular “conservational” interface, ChatGPT—upend this dynamic, producing with minimal effort an inexhaustible supply of improvised, on-demand reading (Poole 2022). This article has three aims: 1. I show examples of ChatGPT-produced novel short-form passages as well as the simultaneous production of pedagogical apparatus like vocabulary lists and reading comprehensive questions (in English or in Latin); 2. I argue that this is a profound and unique shift in the teaching of Latin, at least within recent centuries, namely that we have available to emergent readers what amounts to infinite extensive, comprehensible input for classroom use; and 3. I argue that while current LLM implementations are not adequate for classroom use without intervention, there are two steps that teachers can take to improve this situation: a. develop an “editorial” mindset, as opposed to a content-creation mindset, that is they can rely on LLMs to produce first drafts of classroom-tailored materials from stories to quizzes to handouts (so, in a certain sense an AI-assisted complement to “untextbooking”, Ash 2019) and expend their energy on pedagogical “fine tuning” aimed at so-called last-mile improvements to direct output from LLMs.

Works Cited

• Ash, R. 2019. “Untextbooking for the CI Latin Class: Why and How to Begin.” Journal of Classics Teaching 20(39): 65–70. doi:10.1017/S2058631019000114.
• Carlon, J.M. 2013. “The Implications of SLA Research for Latin Pedagogy: Modernizing Latin Instruction and Securing Its Place in Curricula,” Teaching Classical Languages 4(2) 106-122.
• Cooper, M. 2022. “To Read or Not to Read: Trialling an Extensive Reading Program in a Year 10 Latin Classroom.” Journal of Classics Teaching 24(47): 44-51. doi:10.1017/S2058631022000587.
• Day, R.R., and Bamford, J. 1998. Extensive Reading in the Second Language Classroom. Cambridge: Cambridge University Press.
• Gruber-Miller, J. 2023. “Reading in a Multisensory Environment: The Visual Latin Reading Library.” Classical Outlook 98(1): 22–25.
• Hunt, S. 2022. “Novellas and Free Voluntary Reading: An Overview and Some Starting Points for Further Research into Practice.” Journal of Classics Teaching 23(46): 176–83. doi:10.1017/S2058631022000186.
• Moran, J. 2022. “Comprehending Comprehensible Input (CI): Some Observations.” Journal of Classics Teaching 23(46): 124–25. doi:10.1017/S2058631021000799.
• Olimpi, A. 2019. “Legere Discitur Legendo: Extensive Reading in the Latin Classroom.” Journal of Classics Teaching 20(39): 83–89. doi:10.1017/S205863101900014X.
• Patrick, M. 2019. “Free Voluntary Reading and Comprehensible Input.” Journal of Classics Teaching 20(39): 78–82. doi:10.1017/S2058631019000126.
• Patrick, R. 2015. “Making Sense of Comprehensible Input in the Latin Classroom.” Teaching Classical Languages 6(1): 108–36.
• Piantaggini, L. 2019. “Input-Based Activities.” Journal of Classics Teaching 20(39): 51–56. doi:10.1017/S2058631019000084.
• Poole, F. 2022. “Using ChatGPT to Design Language Material and Exercises.” The FLTMAG. December 13. https://fltmag.com/chatgpt-design-material-exercises/.
• Ro, E. 2019. “How Learning Occurs in an Extensive Reading Book Club: A Conversation Analytic Perspective.” Applied Linguistics 40(1): 152–75. doi:10.1093/applin/amx014.
• Rogers, K. 2019. “Comprehensible Input FAQs.” Journal of Classics Teaching 20(39): 33–36. doi:10.1017/S2058631019000059.
• Suk, N. 2021. “Developing a Sensitive but Generalizable Measurement of Vocabulary Gains from Self-Selected Extensive Reading.” System 101: 102614. doi:10.1016/j.system.2021.102614.
• VanPatten, B. 2014. “Creating Comprehensible Input and Output.” The Language Educator (October/November): 24–26.
• Venditti, E. 2021. “Using Comprehensible Input in the Latin Classroom to Enhance Language Proficiency.” Journal of Classics Teaching 22(43): 22–28. doi:10.1017/S2058631021000039.