51ÉçÇøºÚÁÏ

TILT program: Teaching and Learning Development Grant (TLDG)

Principal Investigator: Ashley Farris-Trimble, associate professor, Department of Linguistics, Faculty of Arts and Social Sciences

Project team: Danica Reid, research assistant, Department of Linguistics, Faculty of Arts and Social Sciences

Timeframe: June 2022 - August 2023

TILT Support: $5000

Course addressed: LING 321 â€“ Phonology

Final Report: View Ashley Farris-Trimble's final report (PDF)

Description:

This project examined the effectiveness of introducing artificial language dataset creation as a pedagogical tool in the third-year Linguistics' undergraduate phonology course. Traditionally, the course relies on data from real languages, but this study explored whether having students construct their own artificial language datasets would improve their understanding of core phonological concepts. 

A mixed-methods design combined quantitative and qualitative data to evaluate the impact of artificial language dataset creation. Quantitative measures included a comparison of fall 2022 grades with three prior semesters taught by the same instructor and three surveys measuring students’ self-reported familiarity with phonological concepts, while qualitative data came from two end-of-semester focus groups. Grade analysis distinguished between high-stakes exams and low-stakes assignments and showed no overall differences between semesters, but exam scores were lower in fall 2022. Survey results showed increased familiarity with all concepts across the semester, indicating overall learning gains. Focus groups revealed generally positive perceptions of dataset creation, highlighting its hands-on and analytical benefits, alongside concerns about reduced exposure to real language data, misalignment with exams, and uncertainty about its role within the discipline.

Additional findings showed little improvement in student curiosity or playful engagement with course material, and no major changes in overall attitudes toward coursework. Instructor reflections emphasized the importance of strong scaffolding, noting that clearer step-by-step guidance significantly improved student experiences. Overall, the study suggests that dataset creation is best used as a supplement to traditional data analysis rather than a replacement, with careful alignment to assessment and learning goals.

Questions addressed:

• Was student mastery of phonological concepts affected by the dataset creation activities?
• What were students' perceptions of the dataset creation activities?
• Did dataset creation arouse students' curiosity and inspire them to play with language?
• Did dataset creation change students' attitudes toward coursework or study habits?
• How should dataset creation activities be scaffolded and integrated into phonology instruction?

Knowledge sharing: Project findings were discussed informally with colleagues. Plans to discuss results at upcoming departmental teaching brown-bag.

Keywords: Phonology, artificial language datasets, dataset creation, active learning, scaffolding, phonological rules, student-generated content, linguistics pedagogy, constructive learning, hands-on assignments