Plato's dialogue on whether names fit things by nature or only by convention, staged as a debate between Socrates, Cratylus, and Hermogenes.
Cratylus is Plato's dialogue on language: does a name belong to its object by nature, the same for Greeks and foreigners alike, or is a name correct only because a community has agreed to use it? Hermogenes, unable to get a straight answer out of his friend Cratylus, brings the question to Socrates, and the two positions -- naturalism and conventionalism -- are tested against each other for most of the work.
Socrates first pushes back on pure conventionalism by comparing a name to a tool, like a shuttle or an auger: not every piece of wood makes a good shuttle, and not every craftsman who makes tools is equally skilled, so there ought to be a 'correctness' that a name-maker (the lawgiver) can get right or wrong, even if different languages render that correctness in different syllables. He then spends a long central stretch doing etymology -- working through the names of gods, heroes, virtues, and abstract terms (including a famous run on Homeric names like Hector and Astyanax) to show how a name can be an 'imitation' of the thing's nature spelled out through sounds and syllables. But the etymologies grow increasingly playful and strained, and Socrates himself keeps signaling that he doesn't fully trust the method.
In the second half, addressed to Cratylus, Socrates turns the argument against pure naturalism too: if names were simply correct or incorrect images of things, a bad name-maker could never mislead us, yet clearly people do use wrong or contradictory names. He also confronts Cratylus's Heraclitean view that everything is always in flux, arguing that if nothing ever stays the same, nothing could be known or even named, since the object of knowledge would keep changing out from under the word for it. The dialogue ends without a settled verdict -- Socrates urges Cratylus to keep investigating rather than accept either extreme -- making it one of Plato's most open-ended and most-cited works on the philosophy of language.
“Socrates, Cratylus here says that there is a correctness of names belonging to each thing by nature, and that a name is not whatever people agree to call a thing”
— Cratylus · read this chapter →
“So a name too is a kind of instrument.”
— Cratylus · read this chapter →
“So it isn't every man, Hermogenes, who can assign a name, but only a certain craftsman of names. And this, it seems, is the lawgiver, who is the rarest of craftsmen among men.”
— Cratylus · read this chapter →
“Haven't we agreed, more than once, that names correctly given are like the things they name, and are images of those things?”
— Cratylus · read this chapter →
“But then it isn't even reasonable to say that knowledge exists at all, Cratylus, if all things are changing and nothing stays put.”
— Cratylus · read this chapter →
It is a dialogue on whether the correctness of names is natural (the same everywhere, fitting the thing itself) or merely conventional (whatever a community agrees to call something). Socrates mediates between Cratylus, who defends naturalism, and Hermogenes, who starts out as a conventionalist.
Not fully. Socrates uses a long etymological survey to make naturalism plausible, but he explicitly distrusts the method he's using, and in the second half of the dialogue he argues against Cratylus that a name could be poorly made and still mislead people -- something impossible if names were perfect natural images.
Cratylus holds the Heraclitean view that everything is constantly changing. Socrates argues against this near the end of the dialogue: if the objects of knowledge never stay the same even for an instant, then knowledge itself could never take hold of anything, since the moment someone approached to know a thing it would already have become something else.
As part of his etymological method, Socrates analyzes Homeric names to test whether their sounds reflect the nature of the things they name -- arguing, for instance, that 'Hector' and 'anax' (lord) share a root sense of holding or possessing, since a king's son would naturally inherit a name related to rule.
No. The dialogue ends without settling whether names are natural or conventional; Socrates tells Cratylus that the question deserves further investigation rather than a fixed answer taken on faith.
Alcibiades 1 · Alcibiades 2 · Apology · Charmides · Cleitophon · Critias · Crito · Epinomis · Euthydemus · Euthyphro · Gorgias · Hipparchus