In “Leviathan” (1651), in the chapter “Of Reason and Science,” Thomas Hobbes made a claim that reads almost like a manifesto for artificial intelligence written three centuries early. When a man reasons, he wrote, he does nothing else but conceive a sum total from addition of parcels, or a remainder from subtraction; reasoning, in Hobbes’s phrase, is reckoning. To reason is to compute: to add and subtract not just numbers but names, ideas, and propositions, combining and separating them to reach conclusions.
The force of the idea is its deflation of mystery. If reasoning is a kind of calculation, then there is in principle nothing about it that requires a special, non-physical spark; it is an operation that could be specified and, by implication, carried out by anything able to perform the right additions and subtractions. Hobbes did not build a machine, and his own theory of mind was more complicated than the slogan suggests, but the slogan itself, that thinking is computing, is the seed of the entire computational approach to intelligence.
The line runs forward through Leibniz’s dream of a calculus of reasoning that could settle disputes by computation, through Boole’s algebra of logic, and ultimately to the idea that a computer manipulating symbols could be said to reason at all. The primary source used here is the Gutenberg edition of Leviathan.
For a general reader, Hobbes is where a now-familiar and still-controversial assumption first appears in plain English: that what the mind does when it thinks might be, at bottom, the same kind of thing a calculating machine does.