How Deeply Human Is Language?

ENDORSEMENTS

“In my lifelong investigations on language, I gradually became convinced that some version of Chomsky’s hypothesis about the universal architecture of languages—a small set of combinatory principles—is basically right. Consequently, I also came to believe that large language models, ingenious and useful though they are, unveil precious little of our unique and specific capacity for language. This book provides narratives of these two perspectives in a deep and remarkably engaging way. It helped me get clearer on what I got right and what I got wrong.”
—Gennaro Chierchia, Haas Foundations Professor of Linguistics and Professor of Philosophy, Harvard University

“What if our best predictions and technological achievements are based on little scientific understanding? Grodzinsky pursues this fascinating question as it arises in the case of language and generative AI. This beautifully told story is important for anyone interested in the science of language and its neural encoding, and their relation to the promises of generative AI. More broadly, the lessons are crucial for those who want to think deeply about the complex interconnections between science and technology.”
—Danny Fox, Anshen-Chomsky Professor of Language & Thought, MIT

“The thing about people extolling large language models is that most of them don’t actually know all that much about how human language works. Yosef Grodzinsky does. A bracing interdisciplinary analysis of AI’s alleged crown jewels, from a bona fide neurolinguist.”
—Dr. Gary Marcus, author of Taming Silicon Valley

“A specter is haunting the world. Are we about to be brutalized by barbarians, or have we have failed to read both our minds and theirs? The ambiguity consumes us throughout the book, until it is made devastatingly poignant, and double edged, in the epilogue—but with a glimmer of hope.”
—Alessandro Treves, Professor, Scuola Internazionale Superiore di Studi Avanzati, Trieste, Italy