LIGN 117
The Language Code
Natural languages are codes used for communication. This course examines the design features of language that allow them to satisfy two desiderata: communicative effectiveness and efficiency. Topics covered include elementary probability theory (conditional probability; Bayes’ Rule), binary representations, elementary information theory (entropy), data compression, error correction, n-gram language models, noisy channel models, and large language models (e.g., ChatGPT).
Topics
- elementary probability theory (conditional probability
- Bayes’ Rule)
- binary representations
- elementary information theory (entropy)
- data compression
- error correction
- n-gram language models
- noisy channel models
- and large language models (e
Topics explicitly listed in official catalogue description
Sources
- Course catalogue ↗2026-2027