Part 1. Ancient, Medieval & Modern

1*. The Origin of Plato’s Theory of Forms

S. V. Keeling was a learned man who attracted attention in France as well as in the UK. His major distinction from almost all other philosophers was that he was a follower of the Cambridge philosopher McTaggert. McTaggert himself had also the rare distinction of inventing an argument for a paradoxical conclusion, an argument that has been discussed ever since. Like Zeno, who argued that there was no such thing as motion, he has inspired many to refute him, but there is little agreement between the refuters. He was an atheist and his general view of things was that we—human persons—are all enjoying an eternal life of love and mutual knowledge without realizing this. In short, he thought that we were all gods, though he would not have put it like that. He wrote in a very good prose style.

Keeling was a teacher at University College, London, and when he died he left money to a former student of his who decided to found the Keeling Memorial Lecture. To him, therefore, I express gratitude, as also to those who chose to invite me to give the lecture this year.

Now for my title: ‘The origin of Plato’s theory of forms’. I am going to argue, not for any source in earlier philosophers’ writings, but for a root in the doctrine that like-knows-like. If you mention this doctrine to a cultivated native German speaker, you are almost certain to be given a quotation from Goethe, the lines:

Waer’ nicht das Auge sonnenhaft

Wie koennt’ die Sonne es erblicken?

Waer’ nicht in uns des Gottes eigne Kraft

Wie koennt’ uns Goettliches entzuecken?

To translate:

If the eye were not sunny

How could it glimpse the sun?

If God’s own power were not in us

How could what is divine enchant us?

Goethe does not claim to be more than translating from someone he calls Johannes Secundus. So far as I have been able to find out, it is not known who this was. It sounds as if it might be a Pope, but I don’t know of any other reason to think it was, or how Goethe stumbled upon it.

The doctrine that like-knows-like was a frequent one among the early Greek philosophers. It is reported by Theophrastus that Anaxagoras strongly rejected it, and maintained that, rather, unlike-knows-like. He mentioned among other examples that sensation will show you that something is warm if your hand is cooler, and vice versa. ‘Like things cannot be affected by like. An image is cast upon what is not the same colour, but a different one.’ (This is about sight.) We do not apprehend sweet and sour by sweet and sour respectively but by contrasts.

When I first heard Anaxagoras’ objection, I was much impressed by it: it seemed to be a successful dismissal of the ‘like-knows-like’ doctrine. Now I am not so sure. You know that something is cool by way of the warmth of your hand; but coolness and warmth are different degrees of the same thing. (They are not, as the Greeks thought, different mixtures of contraries.) Sweet and sour make us recognise each in the presence of the other by contrast, but again it is contrast within a range of the same kind of properties. It is by difference of colour that we see the pattern of the image presented to the eye. This may be a sufficient reply to Anaxagoras. If so, it would mean that we had to give a more sophisticated version of ‘like-knows-like’ than he was rebutting.

There is a statement in Plato’s Phaedo which appears to be based upon some, perhaps restricted, form of the doctrine that like-knows-like. It is the statement that the soul (the human soul, he means) is akin to the forms. He says it there in the Phaedo without arguing for it. But why did he think it? The reason would seem to be that the soul must in its nature be something like the forms in order to know them. On the other hand, it is not suggested that the soul of a man is a Platonic form. That the human soul must have known the forms from forever we know partly from Plato’s way of having Socrates argue in the Meno for the eternal pre-existence of the soul, and partly from the theses of the Phaedrus and the Phaedo about the process of acquiring knowledge of a form F. In the Phaedrus this is described as a process of contracting into one unitary object what is got from a multitude of perceptions—for example, the knowledge of ‘the equal’ by seeing equal sticks and stones. The descriptions seem equivalent and make one think of the British Empiricists. But the process is described as a process of being reminded, by the many things that are F, of the form or type F itself, which the mind has so to speak seen in a previous existence. Thus there is no commitment to a doctrine of sense impressions as a source of all our ideas.

From the argument in the Meno we can derive that Plato’s Socrates thought—or ought to have thought—that no matter when one could manifest knowledge of mathematical facts, it would always turn out that the apparently new knowledge was really being reminded. Socrates indeed does not explicitly come to this conclusion, but produces a phoney argument which he then has the grace to say he ‘wouldn’t exactly insist on’. The argument is that, without having been taught it in this life, Meno’s slave has turned out to know that the square which is double the size of the given square is the square on the diagonal of the given square. So he was reminded (by Socrates’ questions) of something he already knew when he was not a man. But—and this is what Socrates is a bit shamefaced about—the time of his being a man and the time of his not being a man add up to the whole of time. So for the whole of time he has been a knower of the mathematical fact, and by the same methods Socrates will be able to prove the slave’s previous knowledge of the whole of mathematics.

Without Socrates explicitly saying so, the example does prove that in a certain sense there is no such thing as being taught mathematics. At any rate it is not like being taught history, where you know all you know in the first place by being told. The restriction of the term ‘knowledge’ to necessary and invariable truth was inherited from the Greeks even as late as the thirteenth century. It is to be found then in scholastic philosophy. In consequence of it there has been a known philosophical problem ‘whether one human being can teach another’ which is not part of the familiar repertory of problems in modern Anglo-American and related philosophy.

It is possible to continue the discussion in the Meno from the point where Socrates is a bit shamefaced about his argument for the eternal pre-existence of the soul. One can enlarge the subject matter so that it comprises understanding proofs, which need not be mathematical proofs. Socrates can be made to point out that there is no such thing as understanding valid reasoning and not accepting it as valid. If that is so then the understanding of the validity of valid arguments cannot be taught as something one essentially learns from being told it, like so much that we learn. This enlarges the interest of the argument in the Meno, for it covers much more than mathematics.

However, the argument in the Meno does just concern the slave’s knowledge of geometry, i.e. of mathematics, and we have to remember that Plato thought that mathematicians were trying to reach forms, the only really real things, but that they did not succeed, but only ‘dreamed of reality’. In short, their thoughts and arguments concerned not forms, of which there is only one for any type of thing, but ‘mathematicals’. In the understanding that is ‘revived’ in Meno’s slave, there were two squares, and any square has two diagonals. I don’t think Plato ever tells us how the geometrical knowledge (which is ‘revived’ in the slave’s coming to know the theorem) is to be expressed in terms of the forms the square, the diagonal and the triangle. But in the Phaedo and the Phaedrus he is definitely concerned with the forms, although his example of ‘the equal’ in the Phaedo has to introduce a designation of a form which ought to be necessarily plural: auta ta isa, meaning the equals. If, following a suggestion of Wittgenstein, one compares the notion of a form to that of the standard yard at Greenwich, this leads to one’s asking oneself: ‘Wouldn’t the standard equals have to be two equal lengths side by side, or something of that sort?’ Furthermore, if someone were willing to accept that, together perhaps with some more ‘standard sames’, the question would arise how the standards were to be used as standards.

Once more, so far as I know, though Plato does make Socrates speak of ‘the equals’, auta ta isa, which are plural, he does not offer us any solution of the problem here presented. It is perhaps only a slight problem. Can’t we say: ‘there can be only one form of the equals, even though you have to conceive it as consisting of two things which as being the form in question are essentially equal?’ It appears to me that we can, if we are at all justified in speaking of forms as Plato conceived them. We can just ascribe it to Plato’s perception of awkwardness that his Socrates sometimes says auto to ison—singular, the equal itself, and sometimes auta ta isa—plural, the equals themselves.

So much by way of introduction to the topic of my title: the source of Plato’s theory of forms. As I have said, I am interested not in a derivation from previous thinkers who influenced Socrates or Plato but rather in the philosophical thinking involved in believing in Platonic forms. This thinking is so connected with a doctrine that like-knows-like that I am inclined to look for a form of that doctrine that fits its use to postulate forms, which are not particular but universal, not variable but unchangeable, and each not multiple but single.

Before my account of this source of Platonic thinking, however, I will advert to remarks he makes about knowledge, ignorance and opinion in the Republic. He says that knowledge (episteme) is related to to on—i.e. to what is; non-knowing, or ignorance (agnosia) to to me on—i.e. to what is not; opinion (doxa) comes in between: it is brighter than non-knowing and darker than knowing. So there must be something in between what is and what is not, and that is what opinion is related to. (Republic V, 477–8.)

One cannot cite this as illustrating a belief that ‘like-knows-like’, as two of the states are not states of knowledge. But we may say: what a state (a state of the cognitive mind, to use later language) is related to is like the state; hence, as opinion is between knowing and ignorance, it must be related to something between what they are respectively related to. The word (epi) which I translate ‘related to’ is a preposition, but I cannot find an English preposition that will do the same job. Perhaps I could use an emphatic ‘of’: knowledge is of what is, ignorance of what is not, opinion of what is between. We may note that it is tempting to translate agnosia by ‘error’ rather than by ‘ignorance’ or ‘non-knowing’, but I think we must resist the temptation. Error would be an example of non-knowing, and the only example for which it is reasonable to say ‘here the object is what is not’. Nevertheless it would be a mistranslation, and the correct translation emphasises for us that we have here an example of how weird Plato’s thought sometimes is.

Leaving this, however, we can say that the basic thought displayed in this passage of the Republic is very much the same as that displayed by reasonings which would appeal to the principle ‘like-knows-like’. For example, it might be argued ‘there cannot be discursive thought concerning the deity, because discursive thought is essentially complex, and God is essentially simple’. Or again ‘you cannot intellectually grasp there being two men; because the proposition does not distinguish the two men, in order to think that proposition you would have to double your idea of a man, but that is impossible because the idea “man” is single in anybody who conceives it’. Similarly, if you think of such a proposition as ‘any boy will pick a fight with any boy’ you have to think the idea ‘any boy’ twice over, and yet that idea is single in your mind. In all such cases there is reasoning from the essential features of some thought to what would have to be a feature of the object of thought.

Now I come to the exposition of my title subject: the source of Plato’s theory of forms. A little history is relevant. The first Greek philosophers to enquire into the nature of things thought that there was nothing—no substance—in the world except bodies. Further, they observed that all bodies are movable, or capable of motion, whether this is being moved by others or is moving in the intransitive sense of the word. Indeed, it was thought that bodies were in continual flux, and in consequence some thought that we could not have any certainty about things: certain truth was not available to us. For what is in continual flux cannot be grasped with certainty, because it has slid on before the mind judges it. As Heraclitus said: ‘You cannot step in the same river twice’, and this was not just an observation about rivers with their obviously flowing water: everything is flowing, panta rei, and his utterance can be taken as symbolic of the whole non-state of things.

Upon this stage came Plato. In order to preserve the possibility of our having certain knowledge of truth through our intellects, he laid down that there was another genus of things besides those corporeal ones, a genus separated from matter and motion, which he called kinds or ideas, by ‘participation’ in which any of those particular and sensible things is called a man or a horse or anything else. In this way he declared that pieces of knowledge and definitions and whatever pertains to the action of the intellect do not relate to those sense-perceptible bodies, but rather to such immaterial and separate things. Thus the mind’s thinking is not thought of those corporeal things, but is of the separated ‘kinds’ of such things.

Now there are obviously two falsehoods here. First, with these kinds being immaterial and unmoving, the understanding of matter and motion would be excluded from natural science—whose proper preserve they are. So too would proof by material causes and by any ordinary causes which move things. Secondly, it seems ridiculous to bring in other beings, when we are looking for an account of our knowledge of the very objects that are clearly in our view; other beings, that is, which cannot be the substance of the familiar objects, since they have a logically quite different sort of existence. That being so, having acquired knowledge of these separate substances, we would not be able thereby to make inferences to, or judge well about, the sensible objects. We may remember that in the first part of the Parmenides, in which Parmenides wipes the floor with Socrates, this objection is made by the great man to Socrates’ theory of ideas and Socrates is not able to answer it, or indeed any other.

However, to return to our theme, it does appear that Plato has got misled precisely by an application of the like-knows-like principle. One can use this in either of two ways, take either of two directions with it. One may say: this is how what we are acquainted with is, therefore we have to describe our mode of acquaintance in such a way that it too has these features, or somehow accommodates these features even if it does not seem to have them. Or we may say: this is what our process or apparatus for knowing something is like, therefore what we know with it must be like that too. An example of a problem arising in the context of the first way is to be found in recent discoveries about colour vision: a ray of a certain green combined with a ray of a certain red yields sight of pure yellow. How is this to be explained? Well, why does one want it explained? Because it seems so odd that our power of sight should get such a result out of red and green. How can we accommodate that yellow in our account of colour vision? We’d not be puzzled like that about seeing green when something green was presented to our eyes.

So the most usual use of the like-knows-like principle would be to argue from the object to the kind of thing that being acquainted with it is. Now Plato seems to have moved in the opposite direction: to have thought that the way some known object is in the knowing mind shows us how the known object must be in itself. Now, we may say, he could tell in what way what is known is in the knowing mind when the knowing mind’s knowledge is what is expressed by propositions. Propositions almost always have general terms in them. A general term is expressive of some general feature of the things it is rightly used to describe. So we can say that at least the substance known is a universal—i.e. common to all the members of a given class. (Here I will remark on the fact that Aristotle’s term ousia, which we regularly translate ‘substance’, is a steal from Plato: an ousia is a being, and Plato thought the only ousiai were forms. Aristotle successfully impounded the word for his philosophical account of the most fundamental beings.)

A further property of the meaning of a general term is that it is not only general but it cannot change: if it changed, it would be a different term. This thought is to be found in William James’s Principles of Psychology:

Conceptions form the one class of entities that cannot under any circumstances change. They can cease to be altogether; or they can stay as what they severally are; but there is for them no middle way. They form an essentially discontinuous system and translate the process of our perceptual experience, which is naturally a flux, into a set of stagnant and petrified terms. The very conception of flux itself is an absolutely changeless meaning in the mind: it signifies just that one thing, flux, immovably.

Coming back to the existence of such-and-such a meaning in the mind, it is also immaterial. Making then the transition from the properties of an object of thought as it exists in the mind to what the mind is thinking of in using the term, there is a temptation, to which it seems Plato succumbed, to ascribe to the object of thought as it exists outside the mind the very properties which characterise it as a mental object: being a universal, being immaterial, being unchangeable. The form of the object of thought is in the intellect in these ways, with these characteristics: so much is obvious if you simply consider meanings of a vast number of words—words that stand for objects in the widest sense of ‘object’. And so Plato thought that the things which were objects of thought must have all these properties as they existed apart from being thought about: they must be general, immaterial and unchangeable. This of course puts the understanding at a great distance from material things.

There is however no need for this to be the case. It is true that the understanding gets hold of the material and moveable kinds of objects in an immaterial, and in a sense changeless, kind of way, according thus with its own way of getting hold of anything. Thus it can and ought to be said that the mind can and does know corporeal objects with a knowledge-of that is immaterial, ‘universal’ (i.e. by way of general concepts) and in a way unchangeable, without introducing intervening objects with those properties, and without supposing that the corporeal objects themselves have those properties.

 

* The Stanley Victor Keeling Memorial Lecture delivered at University College, London, on 8 March 1990, and published in Robert W Sharples (ed) Modern Thinkers and Ancient Thinkers. The Stanley Victor Keeling Memorial Lectures at University College, London, 1981–1991. (London: UCL Press, 1993).

† Anscombe seems to have been (mis)quoting from memory. Goethe’s text reads: Wär’ nicht das Auge sonnenhaft/ Die Sonne könnt’ es nie erblicken;/ Läg’ nicht in uns des Gottes eigne Kraft/ Wie könnt’ uns Göttliches entzücken?/

‡ William James, Principles of Psychology. Authorised edition in two unabridged volumes bound as one (Dover Publications, 1950), Volume 1, pp. 467–8.