Yes, I remember.

Very well, then. So long as this intrinsic quality is present, even though the name have not all the proper letters, the thing will still be named; well, when it has all the proper letters; badly, when it has only a few of them. Let us, then, grant this, my friend, or we shall get into trouble, like the belated night wanderers in the road at Aegina,This seems to refer to some story unknown to us. and in very truth we shall be found to have arrived too late; otherwise you must look for some other principle of correctness in names, and must not admit that a name is the representation of a thing in syllables and letters. For if you maintain both positions, you cannot help contradicting yourself.

Well, Socrates, I think what you say is reasonable, and I accept it.

Then since we are agreed about this, let us consider the next point. If a name, we say, is to be a good one, it must have the proper letters?

Yes.

And the proper letters are those which are like the things named?

Yes, certainly.

That is, then, the method by which wellgiven names are given. But if any name is not well given, the greater part of it may perhaps, if it is to be an image at all, be made up of proper and like letters, but it may contain some inappropriate element, and is on that account not good or well made. Is that our view?

I suppose, Socrates, there is no use in keeping up my contention; but I am not satisfied that it can be a name and not be well given.

Are you not satisfied that the name is the representation of a thing?

Yes.

And do you not think it is true that some names are composed of earlier ones and others are primary?

Yes.

But if the primary names are to be representations of any things, can you suggest any better way of making them representations than by making them as much as possible like the things which they are to represent? Or do you prefer the theory advanced by Hermogenes and many others, who claim that names are conventional and represent things to those who established the convention and knew the things beforehand, and that convention is the sole principle of correctness in names, and it makes no difference whether we accept the existing convention or adopt an opposite one according to which small would be called great and great small? Which of these two theories do you prefer?

Representing by likeness the thing represented is absolutely and entirely superior to representation by chance signs.

You are right. Then if the name is like the thing, the letters of which the primary names are to be formed must be by their very nature like the things, must they not? Let me explain. Could a painting, to revert to our previous comparison, ever be made like any real thing, if there were no pigments out of which the painting is composed, which were by their nature like the objects which the painter’s art imitates? Is not that impossible?

Yes, it is impossible.

In the same way, names can never be like anything unless those elements of which the names are composed exist in the first place and possess some kind of likeness to the things which the names imitate; and the elements of which they are composed are the letters, are they not?

Yes.

Then I must now ask you to consider with me the subject which Hermogenes and I discussed a while ago. Do you think I am right in saying that rho is expressive of speed, motion, and hardness, or not?

You are right.

And lambda is like smoothness, softness, and the other qualities we mentioned?

Yes.

You know, of course, that we call the same thing σκληρότης (hardness) which the Eretrians call σκληρότηρ?

Certainly.

Have rho and sigma both a likeness to the same thing, and does the final rho mean to them just what the sigma means to us, or is there to one of us no meaning?

They mean the same to both.

In so far as rho and sigma are alike, or in so far as they are not?

In so far as they are alike.

And are they alike in all respects?

Yes; at least for the purpose of expressing motion equally.

But how about the lambda in σκληρότης? Does it not express the opposite of hardness?

Well, perhaps it has no right to be there, Socrates; it may be like the cases that came up in your talk with Hermogenes, when you removed or inserted letters where that was necessary. I think you did right; and in this case perhaps we ought to put a rho in place of the lambda.

Excellent. However, do we not understand one another when anyone says σκληρόν, using the present pronunciation, and do you not now know what I mean?

Yes, but that is by custom, my friend.

In saying custom do you think you are saying anything different from convention? Do you not mean by convention that when I speak I have a definite meaning and you recognize that I have that meaning? Is not that what you mean?

Yes.

Then if you recognize my meaning when I speak, that is an indication given to you by me.

Yes.

The indication comes from something which is unlike my meaning when I speak, if in your example σκληρότης the lambda is unlike hardness; and if this is true, did you not make a convention with yourself, since both like and unlike letters, by the influence of custom and convention, produce indication? And even if custom is entirely distinct from convention, we should henceforth be obliged to say that custom, not likeness, is the principle of indication, since custom, it appears, indicates both by the like and by the unlike. And since we grant this, Cratylus—for I take it that your silence gives consent—both convention and custom must contribute something towards the indication of our meaning when we speak. For, my friend, if you will just turn your attention to numbers, where do you think you can possibly get names to apply to each individual number on the principle of likeness, unless you allow agreement and convention on your part to control the correctness of names? I myself prefer the theory that names are, so far as is possible, like the things named; but really this attractive force of likeness is, as Hermogenes says, a poor thing, and we are compelled to employ in addition this commonplace expedient, convention, to establish the correctness of names. Probably language would be, within the bounds of possibility, most excellent when all its terms, or as many as possible, were based on likeness, that is to say, were appropriate, and most deficient under opposite conditions. But now answer the next question. What is the function of names, and what good do they accomplish?

I think, Socrates, their function is to instruct, and this is the simple truth, that he who knows the names knows also the things named.

I suppose, Cratylus, you mean that when anyone knows the nature of the name—and its nature is that of the thing—he will know the thing also, since it is like the name, and the science of all things which are like each other is one and the same. It is, I fancy, on this ground that you say whoever knows names will know things also.

You are perfectly right.

Now let us see what this manner of giving instruction is, to which you refer, and whether there is another method, but inferior to this, or there is no other at all. What do you think?

I think there is no other at all; this is both the best and the only method.

Do you think this is also the method of discovering realities, and that he who has discovered the names has discovered also the things named; or do you think inquiry and discovery demand another method, and this belongs to instruction?

I most certainly think inquiry and discovery follow this same method and in the same way.

Let us consider the matter, Cratylus. Do you not see that he who in his inquiry after things follows names and examines into the meaning of each one runs great risks of being deceived?

How so?

Clearly he who first gave names, gave such names as agreed with his conception of the nature of things. That is our view, is it not?

Yes.

Then if his conception was incorrect, and he gave the names according to his conception, what do you suppose will happen to us who follow him? Can we help being deceived?

But, Socrates, surely that is not the case. He who gave the names must necessarily have known; otherwise, as I have been saying all along, they would not be names at all. And there is a decisive proof that the name-giver did not miss the truth, one which you must accept; for otherwise his names would not be so universally consistent. Have you not yourself noticed in speaking that all names were formed by the same method and with the same end in view?

But that, Cratylus, is no counter argument. For if the giver of names erred in the beginning and thenceforth forced all other names into agreement with his own initial error, there is nothing strange about that. It is just so sometimes in geometrical diagrams; the initial error is small and unnoticed, but all the numerous deductions are wrong, though consistent. Every one must therefore give great care and great attention to the beginning of any undertaking, to see whether his foundation is right or not. If that has been considered with proper care, everything else will follow. However, I should be surprised if names are really consistent. Let us review our previous discussion. Names, we said, indicate nature to us, assuming that all things are in motion and flux. Do you not think they do so?

Yes, and they indicate it correctly.

Let us first take up again the word ἐπιστήμη (knowledge) and see how ambiguous it is, seeming to indicate that it makes our soul stand still (ἵστησιν) at things, rather than that it is carried round with them, so it is better to speak the beginning of it as we now do than to insert the epsilon and say ἐπεϊστήμ; we should insert an iota rather than an epsilon. Then take βέβαιον (firm), which expresses position and rest, not motion. And ἱστορία (inquiry) means much the same, that it stops (ἵστησιν) the flow. And πιστόν (faithful) most certainly means that which stops (ἱστόν) motion. Then again, anyone can see that μνήμη (memory) expresses rest (μονή) in the soul, not motion. On the other hand, ἁμαρτία (error) and ξυμφορά (misfortune), if you consider merely the form of the names, will appear to be the same as σύνεσις (intellect) and ἐπιστήμη and all the other names of good significance. Moreover, ἀμαθία (ignorance) and ἀκολασία (unrestraint) also appear to be like them; for the former, ἀμαθία, seems to be τοῦ ἅμα θεῷ ἰόντος πορεία (the progress of one who goes with God), and ἀκολασία seems to be exactly ἀκολουθία τοῖς πράγμασιν (movement in company with things). And so names which we believe have the very worst meanings appear to be very like those which have the best. And I think we could, if we took pains, find many other words which would lead us to reverse our judgement and believe that the giver of names meant that things were not in progress or in motion, but were at rest.

But, Socrates, you see that most of the names indicate motion.

What of that, Cratylus? Are we to count names like votes, and shall correctness rest with the majority? Are those to be the true names which are found to have that one of the two meanings which is expressed by the greater number?

That is not reasonable.