On the compatibility of generative AI and generative linguistics

Area of Science:

• Cognitive Science
• Artificial Intelligence
• Linguistics

Background:

• Generative linguistics, pioneered by Chomsky, has significantly influenced cognitive science and symbolic AI.
• The emergence of neural language models has sparked debate regarding their compatibility with generative linguistics principles.

Purpose of the Study:

• To explore the alignment between generative artificial intelligence (AI) and generative linguistics.
• To demonstrate how generative AI can support core concepts of generative linguistics.
• To establish generative linguistics as a framework for evaluating and enhancing neural language models.

Main Methods:

• Conceptual analysis of generative linguistics core ideas.
• Identification of alignment points between generative AI and generative linguistics.
• Proposal of generative linguistics as an evaluative framework for neural language models.

Main Results:

• Three key areas of alignment between generative AI and generative linguistics were identified.
• Generative AI demonstrates support for fundamental principles of generative linguistics.
• Generative linguistics offers a robust framework for assessing neural language models' cognitive and linguistic validity.

Conclusions:

• Generative AI and generative linguistics share fundamental principles, fostering mutual support.
• Generative linguistics provides essential criteria for advancing neural language models.
• This synergy enhances our understanding of human language and cognition through AI.