Natural language syntax complies with the free-energy principle

Area of Science:

• Computational neuroscience
• Theoretical linguistics
• Psycholinguistics

Background:

• Natural language syntax generates complex, hierarchical structures.
• The free-energy principle (FEP) explains how biological systems maintain organization.
• Previous work linked FEP to speech segmentation and linguistic communication.

Purpose of the Study:

• Extend the FEP to explain the computations underlying syntactic object generation.
• Argue that language design principles, like "minimal search," align with the FEP.
• Introduce Turing-Chomsky Compression (TCC) as a principle for efficient syntactic computation.

Main Methods:

• Conceptual integration of theoretical syntax and the FEP.
• Modeling syntactic computations within the FEP framework.
• Using tree-geometric depth and Kolmogorov complexity (Lempel-Ziv) for preliminary analysis.

Main Results:

• Principles of economy in language design adhere to the FEP.
• Tree-geometric depth and Kolmogorov complexity predict syntactic operations aligned with variational free energy minimization.
• TCC provides a principle for efficient syntactic computation.

Conclusions:

• The FEP offers a powerful framework for understanding higher language functions and computability in linguistics.
• TCC unifies linguistic concerns with the FEP's account of self-organization.
• Evidence from linguistics and psycholinguistics supports grounding syntactic computation in active inference.