Generative Linguistics, LLMs, and the Social Nature of Scientific Success (arxiv.org)

A forthcoming paper by Chesi highlights a crisis in generative linguistics, which has been criticized in the wake of advancements in large language models (LLMs) that challenge Chomsky's theories. Coining the dilemma as not merely one of rigor but of social ambition, Chesi argues that generative linguists must elevate their formal practices while simultaneously engaging more broadly with the community to secure their discipline's relevance. The paper references Piantadosi's assertion that LLMs, through their capabilities, undermine traditional linguistic frameworks. This discussion is significant for the AI/ML community as it underscores the intertwined nature of linguistic models and social dynamics in science. Chesi posits that while intellectual merit is crucial, the current successes in language modeling hinge on broader social engagement. This suggests a need for generativists to expand their focus beyond insular academic rigor, advocating for collaboration and inclusivity in research. The implications could lead to a transformative approach in how linguistic models are developed, potentially fostering innovations that align closer with societal needs and advancements in AI technology.