THE LAWS OF THOUGHT

Digging deeply into the intellectual roots of cognitive science and artificial intelligence.

Griffiths, director of Princeton’s Computational Cognitive Science Lab, surveys centuries of efforts to formulate mathematical models of human thinking. Using diagrams and examples, he traces efforts to mathematically describe logical reasoning, from the 18th century philosopher and polymath Gottfried Leibniz to Geoffrey Hinton’s cutting-edge work on current AI systems. Along the way, the author introduces an eclectic cast of cognitive scientists, philosophers, logicians, engineers, and pioneers. Though not strictly academic, the book is far from a casual read; Griffiths invites us to enter the mental world of some of the most consequential thinkers in mathematics, logic, cognitive science, and psychology, even encouraging readers to test possible solutions for themselves. Figures including George Boole, Allen Newell, Herbert Simon, Alan Turing, and Noam Chomsky animate the narrative. Of Newell and Simon, who drove the Cognitive Revolution, Griffiths notes, “They recognized that people break challenging tasks up into goals and subgoals…expressed and navigated using simple sets of rules.” Formalizing those rules proved daunting, and researchers attempting to encode thinking with rule-based systems created an unwieldy 25 million rules. Current approaches combine rules and symbols, neural networks, and probability theory. Ultimately, humans possess broad intelligence ranging across many domains using “the same brains to put diapers on babies, play chess, prove theorems, cook dinner, write novels, and compose symphonies,” while current AI systems focus on optimally performing a single task. Readers engaging with Griffith’s account of intelligence research will be rewarded with a deep understanding of the future challenges.

A rigorous, lively guide for those seeking clarity about the science behind an increasingly AI-structured world.