The Cognitive Architecture of Linear Thinking
#Cognitive #Language
The tendency toward linear extrapolation is not a mere intellectual bad habit that education might correct. There is substantial evidence that it is deeply embedded in human cognitive architecture, the product es that optimized for rapid, locally valid judgments rather than globally accurate models of complex systems.
The psychological research of Daniel Kahneman and Amos Tversky identified a family of cognitive heuristics — mental shortcuts that are fast and often useful but systematically biased. Among the most relevant is anchoring, the tendency to make estimates by starting from an initial value and adjusting insufficiently. When people reason about future states of a system, they anchor on the current state and project a continuation of recent trends. The adjustment is almost always too small, and almost always linear. Similarly, availability bias causes people to weight recently experienced or easily imagined trajectories more heavily than statistically representative ones. Since most of lived experience consists of locally continuous periods — the chaos and discontinuities are, by definition, exceptional — the cognitive record is dominated by smooth progressions, reinforcing the linear projection habit.
Cognitive research on numeracy has repeatedly demonstrated that human intuitions are poorly calibrated for exponential processes. Experiments on compound interest, epidemic spread, and population growth consistently show that people dramatically underestimate the outcomes of exponential dynamics, mentally substituting a linear approximation. The Lego bricks to the moon experiment — how many times must you double a tower of Lego bricks reach the moon? — reliably produces wildly underestimated answers precisely because intuition linearizes the exponential. Even more fundamentally, human intuition has essentially no native capacity for reasoning about chaotic dynamics — systems in which small perturbations in initial conditions produce radically different long-run trajectories. The concept of sensitive dependence on initial conditions is intellectually accessible but cognitively almost inert: we can state it, but we cannot feel it, and our reasoning defaults to the assumption that similar inputs produce similar outputs.