This thesis presents a novel method of modelling metacognition computationally. Metacognition is commonly described as cognition acting on itself, and is correlated with enhanced performance in memory, reasoning, emotional regulation, and motor skills. How it attains these effects remains unclear. Understanding the mechanisms of metacognition requires surmounting two barrriers: the subject's highlevel abstraction and disputed terminology. To overcome these obstacle, and to clarify the workings of metacognition this thesis employs a computational cognitive architecture to define the base units of cognition, and how they come to act on themselves. Well-defined computational units are built upon to form increasing complex metacognitive processes. These computational forms of metacognition are then connected to the research literature. Finally, each form of metacognition is built into working models within the cognitive architecture ACT-R. These working models serve as an existence proof of the models' viability and functionality. The intention of this thesis is to help clarify the nature of metacognition, its underlying mechanisms, and its implications for advancing a unified theory of metacognition. CLARIFYING METACOGNITION of 3 78 Acknowledgements I owe a debt of gratitude to my advisor, Robert West for his depth of knowledge, discussion, and patience. I am thankful to have the opportunity to work with Myrto Mylopoulos, whose support and insight I greatly benefit from. I am also grateful to Andrew Brook for his wonderful philosophical guidance. Finally, my colleagues and lab mates have been a great source of knowledge and conversation. I am deeply thankful to have found such excellent support at Carleton University and the Cognitive Modelling Lab.