Does Plato offer a satisfactory account of naming in the Cratylus?


As a preliminary, it is important to be clear about what is meant by "naming" in this title question.   The Greek word that Plato uses in the Cratylus is "onomata" (plural of "onoma").   Although usually translated as "names", it

"in fact varies between being (a) a general term for 'words', (b) more narrowly, nouns, or perhaps nouns and adjectives, and (c) in certain contexts, proper names alone.   In (a), the most generic use, it comes to designate language as such."(1)  

Plato is generally considered to be a "naturalist".   At Cratylus 388b, Socrates attributes two functions for names: (a) conveying information to each other; and (b) distinguishing things in nature as they are.   A name is "correct" Socrates claims "so long as the essence of the thing remains in possession of the name and appears in it"(393d).   The correctness of names (words) consists in their descriptive adequacy.

So the title question is asking whether Plato, in the Cratylus dialogue, has presented a satisfactory justification and rationale for the naturalistic relation between words, and the referents of those words.

The answer has to be "No!" -- for three reasons.  


Like many of Plato's dialogues, the Cratylus dialogue does not end with a clear-cut answer.   Plato examines both the "conventionalism" of Hermogenes and the "naturalism" of Cratylus.   He begins by persuading Hermogenes that, contrary to his initial conventionalism, there must be some degree of naturalism in words.   By focusing Hermogenes'attention on the pure subjective relativism of Protagoras, Plato persuades Hermogenes that, contra Protagoras, things must have an objective nature.   Since things have an objective nature independent of us, there are necessarily objectively determined skills for dealing with them.   With this reasoning, Hermogenes agrees.   Plato then slides quickly into dealing with "naming"as one of those objectively determined skills.   Thus supposedly defeating Hermogenes'conventionalism.  

One might think that this leaves the field to naturalism, and Plato does proceed to present a long list of etymologies that appear to be based on a theory of words as natural descriptions.   But at the close of the dialogue, Plato employs an analogy with paintings to demonstrate to Cratylus that there can be a marked difference between the name of something and the thing for which it is the name.   Combining the analogy that Plato employs of the picture and his later discussion of the words "σκηρότης"and "σκηρότηρ", one can imagine Socrates arguing that a portrait of Hermogenes might be so badly drawn that it looks more like Cratylus than it does like Hermogenes, leading other people, who do not know the intended subject of the portrait, to incorrectly believe that it is a portrait of Cratylus.   Hence there is the possibility of "incorrectly/wrongly"assigning a name to a thing, yet still have it function by custom or convention as an identifier of that thing.   Plato further weakens the naturalist argument by pointing out that the names of numbers would be impossible to explain without including the element of convention.

"How can you ever imagine, my good friend, that you will find names resembling every individual number, unless you allow that which you term convention and agreement to have authority in determining the correctness of names?"(435b-c)

So the dialogue ends without a clear statement of either a naturalist theory or a conventionalist theory.   Plato seems to leave matters as a poorly delineated mixture of both.   Hardly a "satisfactory account".


The etymologies that Socrates offers in support of his theory of names as natural descriptions are hilariously off the mark.   Certainly, the Greeks of Plato's time had little historical awareness of the origins of the Greek language.   So it is entirely likely that Plato was in earnest with his suggested etymologies.   Aristotle, for example, gives no indication that Plato's etymologies were in jest.   They may have been quite persuasive at the time, when there was no information to the contrary.   However, given the availability of modern etymological knowledge of the origins of words (especially the Greek words that are the basis of Plato's reasoning), the entire argument for names as natural descriptions dissolves into an amusing entertainment, rather than any solid support for the argument.

While it is certainly acceptable to present an explanation of the meaning of words "sui generis", absent the etymologies, there is nothing in Plato's naturalistic analysis of words that can function as a logical rationale (a "logos") for the theory.   With the bulk of the Cratylus dialogue being dismissed as hopelessly fanciful, it leaves Plato's statements about the functions and correctness of "names"as having to be taken as given.   His theory is left as a simple assertion, with no supporting rationale.   Again, hardly an approach that could be called a "satisfactory account".


In his attack on the conventionalism of Hermogenes, Plato confuses the divisions of nature with the divisions of language.   He properly points out to Hermogenes that reality has divisions that we ignore at our peril.   For example there is a "right"(meaning efficacious) way of cutting something, a "right"way of separating the threads while weaving, and so forth.   In dealing with men and horses, one treats them the same only with a cost.   It is more efficacious to treat them differently.   Therefore, it is more efficacious to think of them separately.   Therefore, it is more efficacious if our names of things reflect the natural divisions of nature.   Plato makes this point to Hermogenes rather briefly as part of his critique of conventionalism.   But then Plato slides into treating the names of things in the same manner as he has been treating the natural divisions of nature.   He conflates the differences between names with the differences between the various natural divisions of nature.   And this is a move that Hermogenes, being the representative of conventionalism, should have objected to immediately.

There is a considerable difference between saying that there is a natural difference between horses and men that should be reflected in our having different names for them, and saying that there is a natural difference between our name for horses and our name for men that should be reflected in our naming of them.   Plato does not acknowledge this.   He does not allow that Hermogenes can readily admit to the former, without being committed to the latter.   As long as we recognize the natural division between horses and men by using different names, it is purely a matter of convention (and history) that we call horses "horses"and men "men"rather than the other way around.  

Considering that Plato's attack on the conventionalism of Hermogenes is fundamentally flawed in this manner, and the etymologies that form the bulk of the Cratylus dialogue are dismissible as failing to offer any support for a theory of words as descriptions, the only remaining element of the dialogue is a critique of the naturalist position of Craytlus.   Hence the question must now be raised -- "just what is Plato's account of naming in the Cratylus?"We are left with nothing but the conventionalism of Hermogenes.   While conventionalism may be the proper answer, reaching this solution by default hardly amounts to a "satisfactory account".

Plato may have been led astray by the seeming reasonableness of his etymologies.   Even today, with our more historically accurate picture of the true etymologies of words in various languages, we might also be led similarly astray.   Considering such words as "microcomputer", "craftsman", or even "lymphoma", we might erroneously conclude that the descriptive nature of these words was a necessary feature of all words.   They are, after all, compound words made up of descriptive sub-units.   Just like the more fanciful etymologies that Plato offered.   But if one considers instead a wider selection of some of the "adopted"words that a polyglot language like English has incorporated, the picture becomes a bit more doubtful.  

One does not have to know ancient Greek to understand the meaning of "polyglot"(from the Greek "poluglottos"meaning "many-tongued").   One looks in the dictionary to find out what the conventional meaning is of the word (adjective - knowing, using, or written in several languages; noun - a person who knows or uses several languages).   A similar result is obtained from considering closely related dialects.   If I told you to "put the bread in the boot", would you put the groceries in the car, or would you put the money in the footwear?  

Perhaps Plato did not have readily available either a number of closely related dialects, or a lot of borrowed words.   He does mention a couple of borrowed words during his critique of naturalism, however, and does maintain that their Greek meaning is adopted by convention.   So perhaps it is possible that Plato is a conventionalist after all, despite his flawed attack on the position of Hermogenes.   In any event, for the reasons outlined here, in presenting an "account of naming", the Cratylus dialogue proves quite unsatisfactory and leaves the reader to ponder just what Plato's account really is.



(1)     Sedley, David, "Plato's Cratylus", The Stanford Encyclopedia of Philosophy (Winter 2006 Edition), Edward N. Zalta (ed.), URL = <>.

[Up] [Home] [Next]