Essays on AI, Neuro-circuitry and Basal Ganglia Grammar
Joseph Galasso
June 2024
 

The most compelling evidence to date for involvement of the Basal Ganglia (BG) (Basal Ganglia Grammar) in natural language comes to us from theoretical movement operations (nested dependency, distant binding and trace-theory). This implication of BG overlaps with well-established evidence showing Broca’s involvement with movement. Dual pathways are a marked characteristic of BG insofar that in cascading down-stream neural networks, both direct as well as indirect paths affect admixed neuronal populations from multiple cortical areas. A tentative proposal may suggest that any notion of duality at the subcortical level may have the ability to simulate what we know of local vs distant binding dependencies as found in Dual Mechanism Model accounts of natural language. A theoretical (meta)-synthesis which seeks to connect what we know of Natural Language (NL) with current trends in AI/Transformers may offer us a potential merging of what has up until now been two quite disparate underlying systems. If we assume that NL systems mirror what we find in Parallel Distributed Processing (PDP) across neural networks—and via extension be applicable to any putative AI/Transformer-to-NL corollary—then, by definition, some component of the PDP would necessarily entail a capacity-state which corresponds to concepts, symbols and categorial rules—i.e., real recursive-based prerequisites for natural language which up until now have been sidelined in the implementation of AI modeling.
Format: [ pdf ]
Reference: lingbuzz/008165
(please use that when you cite this article)
Published in:
keywords: recursive-neurocircuitry, basal ganglia grammar, syntax, syntax
previous versions: v1 [May 2024]
Downloaded:319 times

 

[ edit this article | back to article list ]