The Trinity Lancaster Corpus is the largest corpus of spoken English as a second language, created through a collaboration between Lancaster University’s Centre for Corpus Approaches to Social Science (CASS) and Trinity College London.

Based on data from over 2,000 Graded Examinations in Spoken English (GESE) candidates, the corpus contains 4.2 million words of transcribed interactions between L2 English speakers and L1 English-speaking examiners, reflecting diverse cultural, linguistic, and sociolinguistic backgrounds.

Trinity's Graded Examinations in Spoken English (GESE) assess English speaking and listening skills through interactive, one-to-one conversations with an examiner, tailored to the candidate's proficiency level.

We’re proud to highlight groundbreaking research that showcases the practical applications of the corpus. We spoke with Geisa Davila Perez, a postgraduate researcher whose PhD focused on examiner-candidate interactions using the Trinity Lancaster Corpus (TLC). Her findings offer valuable insights for teachers, learners, and examiners. 

Key Takeaways

• Rich Dataset: The Trinity Lancaster Corpus (TLC) captures interactive conversations, ideal for studying English as a Lingua Franca (ELF).
• Focus on Conversation: GESE conversation tasks enable natural, interactive communication across all levels.
• Unique Test Design: Unlike fixed-question formats in other tests, GESE examiners follow a flexible test plan. This minimizes power imbalances, allowing ELF features to emerge naturally.
• Accommodation Strategies: Geisa identified six accommodation strategies, including repetition, paraphrase, trans-languaging, clarification requests, and comprehension checks, which help interlocutors adapt their communication.
• Teacher Training Value: Corpus data can enhance examiner and teacher training with real-world examples.
• Theoretical Impact: The research links ELF concepts to language testing, offering insights for improved test criteria.

Why focus on Trinity's GESE exam and the Trinity Lancaster Corpus?

Geisa explains that her research was guided by both practical and personal factors:

"Working with the Trinity Lancaster Corpus was a natural choice because it’s an existing dataset with rich metadata—like speakers’ linguacultural backgrounds, candidate level of language proficiency and exam performance—that’s invaluable for research. As an English as a Lingua Franca (ELF) speaker myself, I was especially interested in how ELF features manifest in the test context."

The conversation task in the GESE exam stood out as particularly relevant for her study:

"The conversation task is included across all grades, making it suitable for comparing language proficiency levels. The flexible, co-constructed interaction between examiner and candidate reflects authentic communication, minimizing the power imbalance relationship typically seen in other speaking tests."

What did the research reveal about English as a Lingua Franca?

Geisa's work demonstrates that ELF features may naturally be present in speaking tests, even when they’re not part of a test’s explicit construct:

"Accommodation strategies—like repetition, translanguaging, and paraphrasing—are key features of ELF communication. I expanded this term in my research to include strategies like clarification request, comprehension check and confirmation check, which also serve to adapt communication to interlocutors."

These insights contribute to the growing body of ELF research and represent the first empirical study to provide an in-depth exploration of ELF communication in the context of high-stakes language testing.

Implications for teachers and examiners

Geisa sees significant potential for using the Trinity Lancaster Corpus beyond research, particularly in teacher training and examiner development:

"Annotated corpus excerpts could be used for examiner training, showing how examiner language use interacts with candidate language use and the rating scales. For example, this helps examiners identify interactionally challenging moments, like topic shifts, strategic language use that may enhance communication in those moments and its impact on candidate score."

For teachers, Geisa highlights the benefits of using corpora for data-driven learning in language teaching:

"Corpus data offer rich insights into language use, including the pragmatic level. Teachers and learners alike can benefit from these resources to understand real-world language use and improve their skills."

This research underscores the unique qualities of Trinity’s exams and resources, such as the Trinity Lancaster Corpus. As Geisa's findings show, these tools have broad applications for enhancing teaching, learning, and assessment practices.

For teachers, researchers, and examiners looking to explore corpus data or understand how authentic communication unfolds in a testing context, Trinity’s innovative resources are an excellent place to start.