Name: How Language Was Born and Evolved
Author: Mutsumi Imai, Kimi Akita

📖 Book Summary

*The Essence of Language: How Language Was Born and Evolved* tackles four enormous questions: Why do only humans have language? How do babies learn to speak from scratch? Where did this massive language system originate? What separates humans from AI and animals? The author's key clue: onomatopoeia and mimetic words (Japanese: *onomatope*). The original Japanese title is *Gengo no Honshitsu: Kotoba wa Dō Umare, Shinka Shita ka* (Chūkō Shinsho, first edition 2023/5/24). The publisher's blurb cuts straight to the point: asking about the essence of language is asking "what is a human being," and the keys to the answer are **onomatopoeia** and the uniquely human capacity for abductive reasoning. This is no niche title in Japan either—it won the "Shinsho Grand Prize 2024" first place and received the "Asian Book Award 2024 Best Book Prize (General Books Division)," among other accolades.

✍️ Reading Notes

The real difficulty of language has never been "inventing a word" but what that word actually connects to. You can use one word to explain another, but eventually you're just spinning in a dictionary, forever pointing between symbols. This is the core perspective the book keeps returning to: language needs to be "grounded"—it needs to stick back to sensation and perception. Onomatopoeia is the earliest "grounding wire." These words have a cheat-like advantage: the sound itself carries feeling. Japanese examples like **WAKUWAKU (excitement)** and **DOKIDOKI (heart pounding)**—even if you don't know Japanese, it's easy to guess what state they describe. Even more amusingly, the author uses Pokémon naming as a massive database: the way a name "sounds like that creature" is actually language playing "sound-feeling matching." But the book aims to do more than rehabilitate onomatopoeia; it flips "onomatopoeia is childish" into "onomatopoeia is primitive and crucial." The path by which babies acquire language may hint at how language was generated in human history: starting from sounds close to bodily experience, then gradually moving toward abstract symbols—from expressions that "resemble something" to a vocabulary system based on "arbitrary convention." Then it throws the question back at our most pressing modern anxiety: AI is great at talking, but does it understand? The book's position isn't to bash AI but to point out a key capacity in human language: abductive reasoning / hypothesis formation (abduction). That is, you see fragmented clues, dare to guess a possible rule first, and then revise it with new experience. The publisher's blurb even calls it "a learning ability unique to humans." Think of it this way: children don't grow up by memorizing grammar; they grow up by "constantly guessing, getting it wrong, and correcting." By this point, I think the book's true ambition is to transform "linguistics" from a humanities trivia subject into a new lens for seeing the world. You start noticing: how much of what we say every day still uses the body as a foundation (gestures, onomatopoeia, tone, rhythm) rather than relying purely on abstract vocabulary. Language isn't just a tool—it's a technology for "how humans turn the world into shareable experience."

💬 Discussion Points

1If language originally grew from "bodily sensation," what word do you most often use to describe your feelings? Are you more of an "onomatopoeia type" or an "abstraction type"?
2Have you ever had a conversation where "you knew what you were saying, but the other person completely didn't get it"? What shared experience was usually missing?
3As AI gets better and better at talking, what do you think the minimum threshold for "understanding" is: being able to converse counts? Needing embodiment (body and experience) to count? Or needing to form new hypotheses rather than just rearranging sentences?

← Back to Books