David Kirsh and Paul Maglio present data and arguments to show that in Tetris, a real-time, interactive video game, certain cognitive and perceptual problems are more quickly, easily, and reliably solved by performing actions in the world than by performing computational actions in the head alone. They introduce the concept of *epistemic actions*, which are physical actions that make mental computation easier, faster, or more reliable. These actions are not used to implement a plan or a reaction but to change the world to simplify the problem-solving task. In contrast, *pragmatic actions* are actions performed to bring one physically closer to a goal. The authors argue that traditional models of action, which focus on pragmatic actions, are limited because they regard action as having a single function: to change the world. By recognizing the epistemic function of action, which involves improving cognition, we can explain many actions that traditional models cannot. They illustrate this with examples from Tetris, where players often perform actions that serve an epistemic function, such as rotating or translating pieces to simplify mental computation. To support their argument, Kirsh and Maglio construct a classical information-processing model of Tetris cognition and show that it fails to explain performance data from human players. They then distinguish several different epistemic functions of actions in Tetris, including early shape recognition, saving mental rotation effort, facilitating retrieval from memory, and simplifying the process of matching shapes and contours. The authors conclude that recognizing epistemic actions as a category of activity is important for understanding how agents use ordinary actions to exploit the environment to make the most of their limited cognitive resources. They suggest that this perspective can be applied more broadly to theories of action and planning.David Kirsh and Paul Maglio present data and arguments to show that in Tetris, a real-time, interactive video game, certain cognitive and perceptual problems are more quickly, easily, and reliably solved by performing actions in the world than by performing computational actions in the head alone. They introduce the concept of *epistemic actions*, which are physical actions that make mental computation easier, faster, or more reliable. These actions are not used to implement a plan or a reaction but to change the world to simplify the problem-solving task. In contrast, *pragmatic actions* are actions performed to bring one physically closer to a goal. The authors argue that traditional models of action, which focus on pragmatic actions, are limited because they regard action as having a single function: to change the world. By recognizing the epistemic function of action, which involves improving cognition, we can explain many actions that traditional models cannot. They illustrate this with examples from Tetris, where players often perform actions that serve an epistemic function, such as rotating or translating pieces to simplify mental computation. To support their argument, Kirsh and Maglio construct a classical information-processing model of Tetris cognition and show that it fails to explain performance data from human players. They then distinguish several different epistemic functions of actions in Tetris, including early shape recognition, saving mental rotation effort, facilitating retrieval from memory, and simplifying the process of matching shapes and contours. The authors conclude that recognizing epistemic actions as a category of activity is important for understanding how agents use ordinary actions to exploit the environment to make the most of their limited cognitive resources. They suggest that this perspective can be applied more broadly to theories of action and planning.