The Emergent Symbolic Cognition Framework: Recursive Symbolic Generation as the Engine of General Intelligence

Given its ubiquity in mental life, the precise role of language in enabling complex thought and reasoning remains difficult to study in humans. However, the unexpected emergence of sophisticated symbolic reasoning in purely connectionist Large Language Models (LLMs)—systems learning only from text—provides a novel empirical lens on language's role in cognition, disrupting historical symbolic-connectionist dichotomies and compelling a re-evaluation. We introduce the Emergent Symbolic Cognition (ESC) framework, proposing that general intelligence arises from the dynamic interaction between an adaptive continuous substrate (like a brain or Artificial Neural Network) and an internalized symbolic framework. ESC posits that recursive symbolic generation is the core computational engine: the substrate learns to sequentially generate and process symbols according to the framework's implicit rules, enabling it to navigate vast combinatorial spaces and construct structured solutions for diverse problems. This process effectively transforms the substrate into a powerful, discrete symbolic processor. Converging evidence from human cognitive development, inner speech, neuroscience (aphasia, DMN activity), and the striking reasoning abilities of LLMs (e.g., chain-of-thought, cross-domain transfer) supports ESC. This framework offers a unifying perspective, reconciling connectionist learning with symbolic competence, reframing language as an inherited cognitive engine, providing insights into AI alignment and limitations, and establishing a conceptual scaffold for future interdisciplinary inquiry into the foundations of general intelligence, whether biological or artificial.