Noam Chomsky’s “Syntactic Structures” was published in 1957 by Mouton in The Hague, a slim volume of about 117 pages drawn from his lecture notes. It argued that the grammar of a language is a formal system of rules that can generate an infinite set of sentences from finite means, and that this syntactic machinery operates independently of meaning. The Internet Archive holds a digitized copy of the original Mouton edition.
The book’s most quoted line is the sentence “Colorless green ideas sleep furiously,” which Chomsky offered as an example of a string that is grammatically well-formed yet semantically nonsensical. He used it to argue that grammaticality cannot be reduced to whether a sequence of words is statistically probable, since neither that sentence nor its scrambled reversal had ever appeared in English text, yet only one of them is grammatical.
That argument carried directly into computing. Chomsky’s claim that statistical models of word sequences could not capture syntax shaped a generation of rule-based, symbolic approaches to natural language and put generative grammar at the center of the new fields of computational and mathematical linguistics. His formal hierarchy of grammars also became standard material in computer science, underpinning the theory of programming-language parsing.
The book set up a decades-long tension in language technology. The statistical methods Chomsky dismissed - n-gram models, then neural language models - eventually proved enormously effective at exactly the tasks his framework had ruled out of bounds, a reversal that the modern large language model embodies. Even Frederick Jelinek’s obituary recounts how Chomsky’s lecture notes for this very book undermined the early case for statistical language modeling.
Why business readers should care: the fifty-year argument over whether language is best handled by hand-built rules or learned from data is not academic trivia - it is the same choice every team faces today when deciding between curated logic and statistical models trained on large corpora.