Breaking names into etymologies divides one name into other names. But at some point you reach a name that cannot be divided further. Socrates calls these the “primary names” (τὰ πρῶτα ὀνόματα). A primary name has no other names for parts. So how does it stand for its object?
If we had no voice — the gesture argument
Socrates’s answer is mimēsis (imitation). He first has us imagine the case without voice. If we had no voice and no tongue, we would try to convey things with the body.
εἰ μέν γʼ οἶμαι τὸ ἄνω καὶ τὸ κοῦφον ἐβουλόμεθα δηλοῦν, ᾔρομεν ἂν πρὸς τὸν οὐρανὸν τὴν χεῖρα, μιμούμενοι αὐτὴν τὴν φύσιν τοῦ πράγματος.
To indicate “up” or “light,” I think, we would raise a hand toward the sky — imitating the very nature of the thing. (Greek quotations throughout; Japanese and English are L/LAB working translations.)
For “down” or “heavy,” toward the earth. To indicate a horse, we make the body like a horse. With the voice, the principle is the same: a name is just the imitation a gesture would make, done in sound.
ὄνομʼ ἄρʼ ἐστίν, ὡς ἔοικε, μίμημα φωνῇ ἐκείνου ὃ μιμεῖται
A name, it seems, is a vocal imitation of that which it imitates.
But it does not imitate the shape, as a gesture does. With letters (γράμματα) and syllables it imitates the being (οὐσία) of the thing. So the name-maker (ὀνομαστικός), Socrates says, is one of the craftsmen with an “art of likeness,” alongside the painter and the musician. As a painter likens with colour and a musician with pitch, the name-maker likens with the sound of letters.
The sound values of the letters, and the research programme
The letter discussed at greatest length is ρ (rho). Pronouncing it, the tongue vibrates hardest of all. So it suits the expression of motion.
τὸ δὲ οὖν ῥῶ τὸ στοιχεῖον, ὥσπερ λέγω, καλὸν ἔδοξεν ὄργανον εἶναι τῆς κινήσεως τῷ τὰ ὀνόματα τιθεμένῳ πρὸς τὸ ἀφομοιοῦν τῇ φορᾷ
The letter rho, as I say, seemed to the name-setter a fine instrument for motion, for likening names to movement.
λ (lambda) is the reverse: the tongue slips smoothly, so it enters words like “smooth (λεῖος)” and “slip (ὀλισθάνειν).” ι is for fine things that pass through, ν for the inside, α for size, ο for roundness.
Socrates also says that doing this properly needs a research programme — first sort the sounds by kind (vowels, semivowels, mutes), then sort the “things” by kind, then see which sounds fit which. As a painter mixes pigments to make a colour, the name-maker mixes letters to make a name. But Socrates only sketches the programme; he does not carry it out.
It collapses on its own terms
What is striking is that this sound-symbolism theory falls apart within the same dialogue. The Greek word for “hard (σκληρόν)” contains λ. On the sound-symbolism account, λ means “smoothness” — the opposite. And yet we hear “hard” and understand hardness. Why? Cratylus answers: “because of habit (ἔθος).” Socrates at once retorts, “how is habit different from convention?” (article 041). Likeness alone does not make a name; convention is at work — Socrates builds the sound-symbolism theory and then shows its limit himself.
Afterwards — sound symbolism and Saussure
This passage of the Cratylus is the first theory of sound symbolism in the West (the claim that sounds themselves carry meaning). It has been chased experimentally since the twentieth century.
- Jespersen noted the tendency of the front vowel i to mark “small” (English little, French petit, Italian piccolo).
- Sapir (1929) had subjects choose which of
milandmalfitted “a large table,” showing a vowel tendency. - Köhler’s shape experiment linked a spiky figure to “kiki (takete)” and a round one to “bouba (maluma)”; Ramachandran and others revisited it as the “bouba/kiki effect.”
- Blasi et al. (2016) surveyed basic vocabulary across thousands of languages and found sound–meaning links — n in “nose,” l in “tongue,” i in “small” — appearing statistically across families and regions.
- Japanese mimetics (kirakira “glittering,” nurunuru “slimy,” gotsugotsu “rugged”) are a frequent object of study as a case where sound symbolism is grammatically productive.
But modern linguistics took the opposite road. Saussure makes “the arbitrariness of the sign” the first principle of general linguistics: there is no natural link between sound and meaning. He allowed onomatopoeia and interjections as exceptions in a footnote, but treated them as marginal to language. This passage of the Cratylus is the very position Saussure set out to overcome — and Socrates himself, right after, begins to concede that imitation in sound is not enough to make a name.