Rethinking inner speech through linguistic active inference.

Area of Science:

• Cognitive Neuroscience
• Linguistics
• Computational Psychiatry

Background:

• Inner speech is crucial for cognitive functions, but its precise role in predictive brain processes remains debated.
• Existing theories often address specific functions of inner speech without a unifying framework.

Purpose of the Study:

• Introduce and elaborate on the linguistic active inference theory.
• Provide a unifying framework for understanding inner speech's diverse functions and mechanisms.
• Propose how inner speech augments predictive processing for perception and action.

Main Methods:

• Theoretical synthesis integrating principles of active inference and linguistic theory.
• Conceptual analysis of language's properties (efficiency, extensibility, generativity) in relation to predictive coding.
• Framework development to explain inner speech's role in reducing prediction error and uncertainty.

Main Results:

• Inner speech transforms prior expectations and sensorimotor predictions to minimize prediction error.
• Language enables predictive processes to encode sensory experiences and decode goals for action.
• Inner speech dynamically adapts form and function based on computational demands and prediction errors.

Conclusions:

• Linguistic active inference theory offers a unifying explanation for inner speech's functions, phenomenology, and development.
• The theory advances foundational cognitive science by integrating perception, action, and language.
• It provides a roadmap for future research, generating testable hypotheses and new perspectives on symbolic cognition.