The central section of the Cratylus (391–427) is a catalogue of etymologies by Socrates. Some 140 names are taken apart into pieces, one after another. By the standards of modern linguistics almost none of them hold up as etymologies. But the point lies elsewhere: to prise a name open and show what theory it carries.

Three worked examples

Ζεύς (Zeus) = one name, but two forms Ζῆνα (Zêna, acc.) → ζῆν "to live" Δία (Dia, acc.) → διά "through" "that through which all living things live" = king of all Read the name as "a kind of sentence (λόγος)," split it in two, and draw out the theme (396a–b).
The Zeus etymology. Socrates treats a name as a compressed little sentence, splits it, and reads off the meaning. Diagram: L/LAB

The worldview built into the names

first name-giver holds the worldview "all things flow and move" names "flow" and "motion" as roots = a fossil of the theory later speakers inherit the worldview with the names Note: if the first name-giver got the world wrong — the error too is inherited, intact, through the names (436b). That the etymologies agree is no evidence that the worldview is correct.
The point of the etymological section. A name carries the theory of the people who made it. But the theory it carries has not been tested. Diagram: L/LAB

Across the scattered etymologies there is one tendency: motion and flow. On Socrates’s analysis, many words for virtue and knowledge have “go,” “flow,” “turn” at their root. He sums it up:

ὡς ἰόντων ἁπάντων ἀεὶ καὶ ῥεόντων

As though all things are always going and flowing. (Greek quotations throughout; Japanese and English are L/LAB working translations.)

So the first name-givers, on this reading, held a Heraclitean worldview of universal flux and stamped it into the names. A name is a fossil of the theory current when it was made.

There is a device here. Cratylus, the interlocutor, was historically a Heraclitean (according to Aristotle’s Metaphysics, he held that you cannot step into the same river even once, and in old age gave up speech for pointing). So Socrates reads Cratylus’s own worldview out of the names — and then, in article 041, turns and refutes that worldview. The etymological section is a long run-up to the second half’s rebuttal.

The names of the gods — Socrates’s escape route

Before taking the gods’ names apart, Socrates enters a double reservation. First: “we know nothing about the gods — neither about them nor about the names they call themselves”; only the gods know their true names. Second: so we can only follow “the custom in prayers” and call them by the names they are pleased to be called. On that basis he restricts the inquiry to what beliefs led humans to give the gods their names. “That, at least, will draw no blame (ἀνεμέσητον).” He dodges the theology and reads the names as “a record of human opinion.”

Earnest, or a joke?

How you read this long performance sets the tone of the whole dialogue.

Whichever reading you take, the dialogue’s next move is the same. What if the first name-giver got the world wrong?

ὁ θέμενος πρῶτος τὰ ὀνόματα, οἷα ἡγεῖτο εἶναι τὰ πράγματα, τοιαῦτα ἐτίθετο καὶ τὰ ὀνόματα

The one who first set the names set them to match what he believed the things to be. (436b)

If that belief is mistaken, we who follow the names are led into the same mistake. That the etymologies encode a worldview guarantees nothing about whether the worldview is right — and this realisation leads into the critique in article 041.