Pretend Play

In the AI Story

To understand causes, you must be able to reason about what would have happened if things had been different — if the ball had been heavier, if the slope had been steeper, if the lever had been longer. This counterfactual reasoning is the foundation of scientific thinking. And it is the same cognitive operation that drives pretend play. The child who imagines what would happen if bears could fly is exercising the same cognitive machinery as the scientist who imagines what would happen if the speed of light were different. The content differs. The cognitive process is the same.

Large language models do not construct counterfactual worlds. They produce text statistically consistent with their training data — text that sounds like what someone might say about counterfactual possibilities, without the underlying cognitive operation of actually generating and reasoning about those possibilities. The distinction is subtle but consequential. An LLM asked 'What would happen if gravity were twice as strong?' can generate a fluent, plausible response, because its training data contains many discussions of hypothetical physical scenarios. But the model is not reasoning about a counterfactual world. It is producing text that resembles the text that people reasoning about a counterfactual world have produced.

The features of pretend play that developmental research has identified as essential — voluntariness, intrinsic motivation, process orientation, positive affect — map precisely onto the conditions Csikszentmihalyi identified as producing flow states in adults. Flow is, in developmental terms, adult play. The threat AI poses to play is not that it replaces play with something worse but that it converts play into performance. The developer who begins an evening exploring possibilities with Claude — following tangents, trying unlikely combinations, failing and laughing — may gradually shift into compulsive production that has lost every feature of play except the superficial appearance of engagement. The voluntariness drains away. The process orientation is replaced by the pressure to produce. Play becomes productive addiction.

Play has no metrics. This is not a deficiency; it is the feature that makes play cognitively productive. The moment you introduce metrics — the moment the child starts stacking blocks to beat a record rather than to see what happens — the cognitive mode shifts from exploration to exploitation. The block tower may get taller. The cognitive construction stops. The developmental prescription that follows is institutional: protect the conditions that allow play to occur, because those conditions are under relentless pressure in an AI-saturated environment where every playful exploration can be converted into measurable production.

Origin

Gopnik has written about pretend play across her career, with the most sustained treatment in The Philosophical Baby (2009) and further development in The Gardener and the Carpenter (2016). The linkage between pretend play and counterfactual/causal reasoning was formalized in papers with Caren Walker and Paul Harris in the 2010s. The framework builds on earlier work by Jerome Bruner, Lev Vygotsky, and Paul Harris, but Gopnik's contribution was to connect pretend play empirically to the same Bayesian causal-inference machinery that the theory theory describes.

Key Ideas

Dual representation. Children hold reality and pretense simultaneously, reasoning rigorously within both.

Counterfactual machinery. Pretend play exercises the same cognitive operations that scientific reasoning and causal inference require.

The engine of innovation. The capacity to imagine what does not exist and reason about its implications is the developmental foundation of genuine novelty.

Play is defined by intrinsic features. Voluntariness, process orientation, tolerance for failure — these cannot be reverse-engineered through extrinsic incentives.

AI converts play into performance. The pressure to make every exploration measurable destroys the cognitive mode that makes exploration generative.