Schedule of Events | Symposia

Neuro-Generative Grammar: Universal Physics, Natural Syntax and Language-Chunk Assembly Circuits (LCACs)

Poster Session E - Monday, March 31, 2025, 2:30 – 4:30 pm EDT, Back Bay Ballroom/Republic Ballroom

Donald O'Malley¹ (d.omalley@neu.edu), Jessy Xue¹; ¹Northeastern University

Generative Grammars, because they lack substantive Semantic Construction (SemCon), are unable to actually generate sentences. Moreover, the neuronal machinery of SemCon is opaque to both introspection and current neurotechnologies. The complexity of modern sentences poses further challenges, but an alternative is to focus instead on the minimum-viable neuronal circuits needed at the outset of hominid language and subsequently for neocortical evolution and expansion. Language arises from Tagging: our uniquely hominid ability to attach signs to Real-World Items (RWIs). Both the symbolic Tag-store and the naturally-associated RWIs are richly interconnected with other semantic items including actions, properties, auditory/visual events and relationships. These innate neocortical constructs constitute a Universal Physics and enable a Natural (proto) Syntax which can be implemented by prefrontal, working-memory sequencing circuits. Subsequently, order-dependent meaning, based upon clan convention, required novel adaptations/circuits to understand, produce and disambiguate phrases like Hyena kill Thag vs. Thag kill Hyena. We present LCACs: candidate neuronal circuits to implement proto-syntax as well as increasingly complex constructs in line with Jackendoff and Audring’s Relational Morphology (RM; Texture of the Lexicon, 2020). By means of compact schemas, RM avoids GenGram’s most elaborate complexities: RM-schemas are more readily translated into neuronal word-sequencing algorithms. A virtue of our neuro-generative grammar / LCACs approach is that basic pattern-recognition machinery, in conjunction with the evolved Tagging/Symbolic layer, directly enables many kinds of schemas such as add -er (bake => baker) or -en (thick => thicken). Our lengthy, modern sentences also require an expanded, Tag-specific “sentence memory” (WM) system.

Topic Area: LANGUAGE: Syntax