COMPLEXITY Nature of Intelligence, Ep. 2: The relationship between language and thought
208 snips
Oct 9, 2024 
Join Evelina Fedorenko, an MIT Associate Professor exploring the brain's language functions, along with Steve Piantadosi, a UC Berkeley expert on child language acquisition, and Gary Lupyan, who investigates language evolution at UW-Madison. They delve into the dynamic interplay between language and thought, discussing whether intelligence emerges from language or the other way around. The conversation touches on the challenges posed by large language models to traditional linguistics, individual cognitive processes, and the fascinating disconnect between language abilities and reasoning skills.
AI Snips
Chapters
Books
Transcript
Episode notes
Emergent Language Rules
- Chomsky's theory of universal grammar posits innate language rules.
- LLMs challenge this, demonstrating grammatical accuracy without pre-programmed rules, suggesting emergent rules.
The "That" Rule
- Steve Piantadosi uses the example of embedding sentences using "that."
- LLMs learn this rule through exposure, not pre-programming, unlike Chomsky's theory suggests.
Social Drive for Language
- Human language's complexity might come from a drive to socialize.
- This drive, combined with inherent trust, may explain language's evolution, says Gary Lupyan.
