Cratylus

Author: Plato · Year: c. 4th c. BCE (trans. C.D.C. Reeve, Hackett 1997) · Type: dialogue

Plato's dialogue on the correctness of names (orthotēs onomatōn) — "a topic of great interest to Plato's contemporaries that figures little in our own discussions in philosophy of language" (Cooper). It stages two positions: Hermogenes' conventionalism (a name is correct by agreement) and Cratylus' naturalism (each name names by nature). Socrates problematizes both, develops a theory of names as tools made by an expert "rule-setter," fills the dialogue's vast middle with etymologies showing the Greek lexicon encodes a doctrine of universal flux, and then — in the close — argues that knowledge cannot be gotten through names at all but only by reaching "the things themselves," and that the flux-doctrine the names encode may be false. Cooper explains the etymological middle: etymology was a prestigious contemporary "science," so Plato makes Socrates "a first-rate practitioner of the art… as then practiced," speaking against it "as an insider, not an outsider."

Core Arguments

  1. Claim: Things have a fixed being or essence of their own, by nature — not relative to us. Because: if Protagoras' "man is the measure" held, no one could be wiser than another; so "things… are by themselves, in relation to their own being or essence" (386d–e). Against: Protagorean relativism; Heraclitean/Euthydeman flux. Location: 385e–386e.

  2. Claim: A name is a tool (organon) — "for giving instruction, that is to say, for dividing being" (ousia) — not an arbitrary tag. Because: speaking is an action with a natural way of being done; as a shuttle is the tool that divides warp and woof, a name is the tool that instructs by dividing reality (388a–c). Against: Hermogenes' "any name you give a thing is its correct name" (384d). Location: 386e–388c.

  3. Claim: Names are the products of a rare expert craftsman — the rule-setter (nomothetēs) — who gives them correctly by looking to "the form of the name," "what a name itself is." Because: a carpenter whose shuttle breaks looks to "the very form to which he looked," not the broken one; so the rule-setter "embod[ies] in sounds and syllables the name naturally suited to each thing," and different rule-setters may use different syllables "as long as they give it the same form" — reconciling natural correctness with cross-linguistic variety (389a–390a). Against: naïve naturalism's demand that the correct name be literally identical for everyone (383b); and conventionalism's "anyone names correctly." Location: 388d–390a.

  4. Claim: An image must not duplicate what it images — some difference is constitutive of being an image. Because: if a god made a perfect duplicate of Cratylus (color, shape, soul, motion), "there would be two Cratyluses," not Cratylus and an image (432b–c); so correctness is not exact resemblance, and a name may lack some appropriate letters and still be a name. Against: Cratylus' claim that adding/subtracting one letter makes "a different name" outright (431e–432a). Location: 431e–433a.

  5. Claim: Convention must, after all, contribute to meaning — even the naturalist must concede this. Because: 'sklēron' ("hard") contains 'l' (unlike hardness) yet we understand it "because of usage"; and numbers cannot be named by likeness at all — so "we have to make use of this worthless thing, convention" (435a–c). Against: Cratylus' pure naturalism (likeness alone). Location: 433b–435c.

  6. Claim: Knowledge cannot be gotten from names; one must learn "the things that are… through themselves." Because: the giver of the first names had no prior names to learn from, so he knew the things directly, independently of names (438a–b); and the names are not even self-consistent — a "civil war among names" between motion-words and rest-words (438d). Against: Cratylus' "anyone who knows a thing's name also knows the thing" (435d). Location: 436b–439b.

  7. Claim (anti-flux close): If everything flows, there can be no knowledge and nothing to be known — so there must be stable realities (the beautiful itself, the good itself). Because: knowledge requires a stable object; "if it never stays the same, how can it be something?" (439e); "no kind of knowledge is knowledge of what isn't." Against: Cratylus is left unconverted ("things seem… very much more as Heraclitus says," 440d); Socrates leaves the metaphysics formally open. Location: 439c–440e.

Key Findings

  • The dialogue is structured as a double reductio: naturalism and conventionalism are each pushed to absurdity, and Socrates' "third way" — natural correctness instituted by an expert and supervised by the dialectician (390b–d) — collapses into neither pure phusis nor pure nomos.
  • The non-duplication-of-the-image principle (432b–d) is the lever that defeats pure naturalism and re-admits convention: a sign that fully coincided with its object would cease to be a sign.
  • The etymological middle (391–427) is the evidentiary body for the thesis that the lexicon encodes Heraclitean flux — then the close throws that doctrine over, warning that "no one with any understanding will commit… the cultivation of his soul to names" (440c).
  • The Forms enter twice: as "the form of the name" grounding naming (389–390) and as "the beautiful itself / good itself" grounding knowledge against flux (439c).

Concepts Developed

  • correctness-of-names — the Cratylus is the locus classicus: orthotēs onomatōn, the naturalism/conventionalism debate, the organon and nomothetēs theory, the mimetic theory of primary names, and the non-duplication principle.
  • mimesis — a primary name as a "vocal imitation" of a thing's being/essence (not its sound, shape, or color), 423b–424b; refined sharply against crude onomatopoeia.
  • theory-of-forms — "the form of the name" (389a) and "the beautiful itself" (439c) anchor naming and knowledge respectively.

Concepts Referenced

  • flux / Heraclitus — "everything gives way and nothing stands fast" (402a); the doctrine the names encode and the close contests.
  • Protagorean relativism — "man the measure," the foil for stable natures (385e–386c).
  • sōma/sēma — the Orphic aside "the body (sōma) is the tomb (sēma) of the soul" (400c), a bonus sign-theory locus.

Key Passages

"there is a natural correctness of names, which is the same for everyone" (383a–b) — Cratylus' naturalism "any name you give a thing is its correct name" (384d) — Hermogenes' conventionalism "a tool for giving instruction, that is to say, for dividing being" (388b–c) — the organon "the kind of craftsman most rarely found among human beings" (389a) — the rule-setter "embody in sounds and syllables the name naturally suited to each thing" (389d–390a) "if names are to be given well, a dialectician must supervise him" (390d) "would there be two Cratyluses or Cratylus and an image" (432c) — non-duplication "we have to make use of this worthless thing, convention" (435c) "far better to investigate them through themselves than… through their names" (439b) "if it never stays the same, how can it be something?" (439e) — anti-flux

What's Not Obvious

  • The serious-or-parodic status of the etymological middle is the interpretive hinge. The argument structurally requires the etymologies be both provisionally serious (they are the evidence that the lexicon encodes flux) and ultimately discardable (the close throws them over). Socrates' repeated oracular disclaimers — blaming his "inspiration" on Euthyphro (396d–e), "nothing I've said is set in stone" (428c) — are the textual marker of this double status. Reading them as flatly serious or flatly parodic both miss the dialogue.
  • The conventionalist's own word institutes the natural. Nomos (convention/custom) is the root of nomothetēs, the figure who lays down natural correctness by looking to the form. So a "natural" rightness has to be instituted by an expert and then sediments and disguises over time (names "covered over by those who wanted to dress them up," 414c). This is a strong seam with the wiki's institution/sedimentation cluster.
  • Conventionalism is not arbitrariness. Hermogenes is a nomenclaturist — names as agreed labels for pre-given things — not a proto-Saussurean. Saussurean arbitrariness is diacritical (value from differences within the system); the Cratylus's two options are both pre-diacritical, lacking the relational conception entirely. The real seam is "conventionalism ≠ arbitrariness," which blocks the tempting false-friend equation Hermogenes = Saussure.

Critique / Limitations

The "third way" (instituted natural correctness supervised by dialectic) is asserted more than secured; the mimetic theory of primary names strains against the sklēron counterexample the dialogue itself raises; and the metaphysics is left formally open — Socrates argues that knowledge needs stable objects without establishing that there are any, and Cratylus walks away a Heraclitean.

Connections

  • contrasts with plato-meno regarding learning — both ask how the new can be learned; the Cratylus answers "through the things themselves, not through names" (438e), the Meno "by recovery from within."
  • contrasts with innere-sprachform — the etymological-revelatory thesis (a language encodes a worldview) anticipates Humboldt's inner-form/Weltansicht, but imposes a critical limit Humboldt's thesis lacks: the encoded worldview may be false and must be checked against the things themselves.
  • has cross-tradition cousin indirect-language — the non-duplication principle (a coinciding image ceases to be an image) is a structural cousin of MP's thesis that exhaustive/frontal expression would cease to express; the grounds differ (Plato: ontological gap; MP: diacritical/lateral signification).
  • develops the Forms for language and epistemology (the "form of the name"; the "beautiful itself" against flux).

Sources

  • Cratylus, trans. C.D.C. Reeve, in Plato: Complete Works (Hackett, 1997), Stephanus 383a–440e; raw file lines 2351–3993. (Reeve follows Schofield's transposition of 385b2–d1 to after 387c5.)
  • Depth layer: wiki/sources/.extraction-plato-cratylus.md.