(Compare Phaedrus.)

When the fervour of his etymological enthusiasm has abated, Socrates ends,
as he has begun, with a rational explanation of language. Still he
preserves his 'know nothing' disguise, and himself declares his first
notions about names to be reckless and ridiculous. Having explained
compound words by resolving them into their original elements, he now
proceeds to analyse simple words into the letters of which they are
composed. The Socrates who 'knows nothing,' here passes into the teacher,
the dialectician, the arranger of species. There is nothing in this part
of the dialogue which is either weak or extravagant. Plato is a supporter
of the Onomatopoetic theory of language; that is to say, he supposes words
to be formed by the imitation of ideas in sounds; he also recognises the
effect of time, the influence of foreign languages, the desire of euphony,
to be formative principles; and he admits a certain element of chance. But
he gives no imitation in all this that he is preparing the way for the
construction of an ideal language. Or that he has any Eleatic speculation
to oppose to the Heracleiteanism of Cratylus.

The theory of language which is propounded in the Cratylus is in accordance
with the later phase of the philosophy of Plato, and would have been
regarded by him as in the main true. The dialogue is also a satire on the
philological fancies of the day. Socrates in pursuit of his vocation as a
detector of false knowledge, lights by accident on the truth. He is
guessing, he is dreaming; he has heard, as he says in the Phaedrus, from
another: no one is more surprised than himself at his own discoveries.
And yet some of his best remarks, as for example his view of the derivation
of Greek words from other languages, or of the permutations of letters, or
again, his observation that in speaking of the Gods we are only speaking of
our names of them, occur among these flights of humour.