Brendan Conway-Smith's theory maps how metacognitive skills become automatic

Brendan Conway-Smith's doctoral dissertation at Carleton University offers the first formal computational account of how metacognitive skills—the ability to monitor and control one's own cognitive processes—are learned. Published on arXiv (2608.12393), the thesis addresses a long-standing gap: while metacognition has been widely studied, no theoretical framework explained its acquisition. Conway-Smith draws on classic models of skill acquisition from perceptual-motor and cognitive domains to propose that metacognitive skill develops primarily through proceduralization. In this process, explicit, effortful metacognitive strategies are transformed through repeated practice into implicit, automatic routines, much like learning to drive or type.

The theory's explanatory power is demonstrated by its unification of diverse phenomena, including attentional training, emotion regulation, the metacognitive threshold, and detached mindfulness—all shown to emerge naturally from the same proceduralization mechanism. This integrated account not only organizes existing empirical findings but also generates testable predictions for future research. Crucially, the framework extends beyond cognitive science into artificial intelligence: it suggests how AI systems could be designed to learn to monitor and regulate their own processes, moving toward more robust, self-aware architectures. For educators and therapists, the theory offers mechanistic guidance for strengthening metacognitive skills in everyday practice.