Their talk looks at the findings of a pilot study of CorpusChat in an undergraduate class for Arts at the University of Hong Kong. #corpuslinguistics #research Info and registration 👇👇👇 um.es/languagecorp...

This talk discusses how corpus linguistics and generative artificial intelligence (GenAI) may work together in impactful ways, acknowledging the potential of integrating GenAI and corpora while working to address their shortcomings.

Abstract & registration: um.es/languagecorp... Event organized by @letrasumu.bsky.social & The Education University of Hong Kong.
#corpuslinguistics #appliedlinguistics #research #AI

Read the paper here: www.cambridge.org/core/journal...

Implications for language teaching:
🔁 Learners rely on what they see early and often.
📘 Early exposure to emotionally charged or simpler verbs boosts acquisition.
👩‍🏫 Carefully sequencing complex structures in the curriculum could better support long-term learning.

The study introduces input-exposure shaping factors like:
-#Accessibility (how early/easily a word is encountered)
-#Proficiency level
-Prototype formation (early-learned meanings tend to stick)
-These strongly influence what learners produce, modulating the effect of input frequency.

Main finding: Learners do show signs of statistical learning even in complex nonadjacent constructions.
Their usage patterns align with native speaker input in terms of #frequency and #contingency (the likelihood two elements occur together), especially as proficiency increases.

Second language acquisition (SLA) research has long focused on how #language learners pick up adjacent language #patterns. But what happens when the words that form meaning aren’t next to each other? This study dives into nonadjacent #constructions in Mandarin Chinese—an overlooked area in SLA.

Methodology spotlight:
📊 Large-scale #corpuslinguistics analysis using native and learner corpora
🧠 Custom NLP tools to extract nonadjacent #patterns
📈 Mixed-effects negative binomial regression—perfect for skewed linguistic data.

Statistical language learning is a cognitive process by which people unconsciously detect patterns and regularities in the #language they hear or read. This includes how often certain words appear together, in what order, and in what contexts.