Dynamical systems and cognitive linguistics: toward an active morphodynamical semantics

Area of Science:

• Cognitive Science
• Computational Linguistics
• Neuroscience

Background:

• Spatial prepositions like 'in' apply to diverse container shapes, a phenomenon not fully explained by traditional semantic studies.
• Gestalt principles suggest spatial prepositions are invariant to object shape and size, hinting at underlying abstract mechanisms.

Purpose of the Study:

• To develop a dynamical system model for cognitive linguistics using cellular automata and spiking neural networks.
• To categorize the vast diversity of schematic visual scenes into a limited set of grammatical elements.
• To elucidate the topology of language and the cognitive mechanisms of spatial schematization and categorization.

Main Methods:

• Proposed a novel dynamical system approach integrating cellular automata and spiking neural networks.
• Introduced morphodynamical transforms to abstract visual details into virtual structures (singularities).
• Modeled singularities using coupled excitable units exhibiting spatiotemporal pattern formation, specifically traveling waves.

Main Results:

• Demonstrated how morphodynamical transforms create invariant representations, explaining the neutrality of spatial prepositions.
• Showcased how singularities arising from coupled excitable units can form the basis of grammatical elements.
• Linked spatial schematization and categorization to expansion processes like wave propagation.

Conclusions:

• The proposed 'active semantics' paradigm offers a computational explanation for spatial language understanding.
• The model bridges the gap between low-level visual processing and high-level linguistic abstraction.
• This dynamical system approach provides a framework for understanding the interface between vision and language.