Reviewed work(s): Source: The Philosophical Quarterly, Vol. 28, No. 112 (Jul., 1978), pp. 189-207 Published by: Wiley for The Philosophical Quarterly Stable URL: http://www.jstor.org/stable/2218840 . Accessed: 10/03/2013 10:11 Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp . JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org. . Wiley and The Philosophical Quarterly are collaborating with JSTOR to digitize, preserve and extend access to The Philosophical Quarterly. http://www.jstor.org This content downloaded on Sun, 10 Mar 2013 10:11:22 AM All use subject to JSTOR Terms and Conditions THE PHILOSOPHICAL QUARTERLY VOL. 28 No. 112 JULY 1978 COMMON PROPERTIES AND EPONYMY IN PLATO BY THOMAS W. BESTOR Dating right back to Proclus, there has long existed a minority opinion that Plato was aware that his notorious "Third Man" argument was not an effective objection to his theory of Forms after all. In recent times1 this opinion has been based on a distinctive insight: since for Plato words have both direct and also strictly second-hand referents, it is plain that though the (second-hand) application of a word 'F' to a particular inevitably in- volves regressing to a Form, the (direct) application of the same word to the Form itself does not. Usually this minority opinion is dismissed very casually. But recently Roger Shiner and Frank White2 have subjected it to searching criticism: no matter in what other ways it might be F, Plato's theory of predication has to presuppose that the Form b is also F in the ruinous sense of possessing the property F-ness, or else it remains utterly mysterious why it is specifically the Form 0b which is "responsible", as re- quired, for the particulars possessing that property rather than some other. I am sure that this is the most damaging criticism which the minority opinion has to face. In what follows I want to do what I can to weaken it, concentrating primarily on White's recent article in this journal. 1See for example G. E. L. Owen, "Logic and Metaphysics in Some Earlier Works of Aristotle", in I. During and G. E. L. Owen (edd.), Aristotle and Plato in the Mid-Fourth Century (Goteborg, 1960), pp. 181-2; R. E. Alien, "Participation and Predication in Plato's Middle Dialogues", in R. E. Allen (ed.), Studies in Plato's Metaphysics (London, 1965), p. 59; Arthur L. Peck, "Plato versus Parmenides", Philosophical Review, 81 (1962), pp. 160-61, 164-9; Max Cresswell, "Participation in Plato's Parmenides", Southern Journal of Philosophy, 13 (1975), pp. 164-5. 2Shiner, "Self-Predication and the "Third Man" Argument", Journal of the History of Philosophy, 8 (1970), pp. 373-5, 384-6; White, "Plato's Middle Dialogues and the Independence of Particulars", Philosophical Quarterly, 27 (1977), pp. 206-11. This content downloaded on Sun, 10 Mar 2013 10:11:22 AM All use subject to JSTOR Terms and Conditions 190 THOMAS W. BESTOR I. THE DISPUTE Essentially the Third Man argument (specifically the "first Third Man" at Parm. 132a-b) goes like this: (A) If any set of x1, x2 . . . n are all F, they must be F because there is some Form b separate from them all in which they all participate. (B) The Form c is F as well. (C) Therefore, since x1, x. .. xn and the Form f are all F, they must be F because there is some Form 02 separate from them all in which they all participate. Step (A) here is the famous "One over Many" principle. Step (B) is the equally famous "Self-predication" thesis. The conclusion (C) results from simply re-applying the One over Many principle to the One itself. And the entire controversy boils down to whether Plato's theory of predication rules out or backs up such a re-application. According to the minority opinion, the central moves of Plato's semantics are these. That words have primary and also purely secondary referents. That the manner in which one and the same word hooks up to its primary referent is quite different from the way in which it hooks up to its secondary referents. That the primary hook is direct and the derivative hook only roundabout. Thus the single Form b is what the general word 'F' names, precisely as a proper name names. The many different F particulars, how- ever, are only "named after" the Form; to them 'F' applies not as a proper name but eponymously, indirectly, at second-hand as it were. This gives us not one but two ways of applying the same word 'F' to different pieces of the world-'F' names the eponym (the Form b) and eponymously desig- nates the named-after things (the xl, x2 . . . n particulars which participate in that Form). This simple semantic point breaks the back of Parmenides' Third Man completely. In that argument step (A) tells us what it is like with the Many which are named after a One; we can indeed rightly call several things by the single word 'F', but only insofar as those things are appropriately related to the One named by the word. Step (B), on the other hand, is not about any named-after thing at all, but simply repeats the point that it is the One itself, the Form b, which the word 'F' properly stands for. Of course the conclusion (C) does not then follow. It makes the Form into a named-after thing like all the F particulars it has just given its name to and then posits another Form which it must have been named after in turn. But the Form f is not named F after anything else. It is the thing which 'F' names. Consequently, when we apply 'F' to a pedestrian particular we refer to-and hence regress to-something else, to the Form f, but when we apply that same term to the Form c itself we bring in nothing further and hence can never regress past it. According to the majority opinion, however, eponymous predication cannot stand on its own like this. Plato's theory will not do the work it is This content downloaded on Sun, 10 Mar 2013 10:11:22 AM All use subject to JSTOR Terms and Conditions COMMON PROPERTIES AND EPONYMY 191 supposed to if particulars are simply named after some eponym and have no other relation to it. Eponymous predication must always be backed up by some explanation of why the named-after things are named after what they are. By way of contrast, think of the widespread application of such an eponym as 'George Washington' in the United States. Countless high- schools are named "George Washington High School" after George Washing- ton, but there are no nagging questions about why these but not other high schools take George Washington's name. Various school boards could, and probably did, just adopt the name out of a hat. With general words, how- ever, the puzzle has always been to find some account that will explain why it is in some sense "correct" or "fitting" or "appropriate", rather than merely arbitrary, to group some individuals together under one word and not other individuals. So to insist that the particulars we call F are named- after things is never more than the start of the explanation we want. We are always bound to go one step further and ask "Why are these specific partic- ulars (but not others) justifiably to be called after this specific Form (and not after some other)?". Plato's answer, according to the majority opinion, is that a particular F thing is justifiably called F after the Form b because and only because it possesses the property F-ness in common with it: Beauty itself and beautiful particulars are both beautiful univocally. They differ because when beautiful particulars are beautiful, they are not so without limit or qualification. They are only beautiful in part, or at one time and not another, and so on. By contrast, Beauty itself is beautiful but without limit or qualification, without admixture of opposites (White, op. cit., p. 210). In this way it makes perfectly good sense to say of a given item, x, that x is a secondary designate of a predicate, F, because although x has uni- vocally the same property as y which is the primary designate of 'F', x owes its possession of F to y (p. 208). Now if this is how eponymy goes for Plato, the majority opinion continues, then it is perfectly legitimate to apply step (A) to step (B) and so derive the conclusion (C) after all. When all is said and done, the One over Many principle (step (A)) is what gives Plato's fundamental analysis of the ex- pression 'possesses a property': to say that some thing possesses the property F-ness is ultimately to say nothing more, and nothing less, than that it stands in a special relation (of participation) to some additional philosophical entity (the Form q). That is what 'possesses the property F-ness' means. And hence, no matter whether the Form qb is also F in the sense that it is the eponym which the word 'F' names, as long as the Form itself does possess the property F-ness, there is no way to avoid saying that it too participates in some additional entity, the further Form 2--again, that is just what 'possesses the property' means. Hence the regress which Plato deliberately allows the Third Man to bring out is fundamental and inevitable. This content downloaded on Sun, 10 Mar 2013 10:11:22 AM All use subject to JSTOR Terms and Conditions 192 THOMAS W. BESTOR It should be obvious, then, that it really is crucial to decide whether or not a theory of eponymous predication builds on an assumption of common properties-either in point of fact or in Plato's own mind. II. ARE COMMON PROPERTIES AN INTEGRAL PART OF THE SEMANTICS OF EPONYMY? To decide the point of fact I think we have to look at some semantics, because, though Plato did not see himself primarily as a philosopher of language, his theory of eponymous predication is indeed a distinctly semantic theory. (1) Eponymous predication provides a deliberately one-many hook as opposed to a one-one hook between words and the world. In saying that words must first of all name Forms, Plato gives in to our natural inclination to assume that the case of one single name for each single item is the paradigm case of naming. Not so with eponymous predication, however. To explain the point of letting a word do this extra job and so become general is to explain what Eucleides and Antisthenes, for instance, found so unexplain- able: how one word can apply to more than one item. It looks as if it is not possible at all if the only job a word can do is be a proper name as 'Thomas Wheaton Bestor' is a proper name. This is where eponymy steps in. Perhaps there can rightly be only one piece of the world named F, but there can easily be many more than one piece of the world rightly called F after it. Thus there may be only one The Beautiful which the word 'beautiful' names -but lots of things can bear some relation to that Beautiful and so be called beautiful after it. (2) Furthermore, eponymous predication is a distinctly indirect relation between our words and the world. Take a simple parallel. With 'Fido' and Fido we have an immediate, direct relation. Here we have the word, there the thing. We use the word in place of the thing for purposes of communica- tion. Other words are not like this, however. 'Sandwich', for example, is the proper name of John Montagu, fourth earl of Sandwich, and applies to that man as directly as 'Fido' ever does to Fido. But the word also applies to my favourite snacks, and it does so only by picking out the man who invented them on the way: here we have the word, there the man named by the word, and there again the things in my lunch-bag named after the man. The application of the word to my snacks is thus a roundabout affair, not a direct one. Indeed, this is the whole point of calling sandwiches "named- after" things and Sandwich a "named'" thing in the first place. No word can possibly attach to any named-after thing as directly as it attaches to the original named thing, simply because it attaches to the named-after thing only via attaching to the named thing. And so it is with Forms and particulars for Plato. It is the Forms which are the named items. Partic- ulars are only named-after things. A general word attaches to each Many only indirectly, only by attaching to the One on the way. This content downloaded on Sun, 10 Mar 2013 10:11:22 AM All use subject to JSTOR Terms and Conditions COMMON PROPERTIES AND EPONYMY 193 (3) Though eponymous predication vastly extends the range of a term, it does not do so arbitrarily. Not just anything can be named after some named item, only those things "appropriately" related to it. Thus with 'sandwich', the word is the proper name of the man Sandwich and applies eponymously to certain things in my lunch-bag as well-but not to every- thing in my lunch-bag; the man has given his name only to those snacks he is famous for inventing. So too for particulars and Forms in Plato's theory. Not every particular can be named F after the Form k, only those partic- ulars which are specially related to it. The pen on my desk cannot possibly be called pious, even after the Form of Piety; it is not close enough to the Form for that. But the hymnal on my desk can, and is. What all this means is that the relation of being named after is always parasitic on an additional relation obtaining between the named item and the named-after things, a relation in virtue of which the eponym relation also holds. For Plato this additional relation, however it is cashed in, is often dubbed "participation": F particulars are rightly named F after the Form < because they (and not others) "participate in" that Form (and not in the Form 0 say). The point of bringing out these special features of eponymy is simple. According to the majority opinion a particular can be called after a Form if and only if the particular and the Form share a common property in some way. Semantically, this amounts to the view that the only relation which will support the indirect extension of a One's name to some Many is sharing a property. The eponym possesses a distinctive property (in a quintessential way). The named-after things possess that property too (in a rough and ready way). And the named-after things take the eponym's name just be- cause they share the eponym's property. The majority opinion here is by no means eccentric (it is the majority opinion after all). Nonetheless it is entirely false. As a semantic thesis the theory of eponymous predication says nothing at all about what the named things must be like, about what the named-after things must be like, or about what the undergirding relation between named thing and named-after things must be like. Consider just a few extra possibilities in the rest of this section, as a sort of insurance against the usual one-sided diet. (The groupings are naturally fairly arbitrary and fluid. And, of course, the eponymous Forms are not themselves people, nor even always objects, as most of the following eponyms are.) Class One: There are cases where a named-after thing is so named because it does share a common property with the named thing-indeed, because it shares precisely that property which the eponym itself is distinctive and memorable for possessing. White's favourite example is a case in point: a girl may be called "a Helen" because she is beautiful, as Helen wife of Menelaus was famed for being (pp. 206-7). We seldom talk this way in real life, of course, but there are lots of more ordinary cases to be going on with: a feeble-minded person is called a moron after the memorably feeble-minded character Moron in Moliere's play La Princesse d'Elide; a shady lawyer is This content downloaded on Sun, 10 Mar 2013 10:11:22 AM All use subject to JSTOR Terms and Conditions 194 THOMAS W. BESTOR called a shyster after a memorably shady New York lawyer named Scheuster. Calling something "an X" after X is not the only pattern; 'X' undergoes a variety of grammatical transformations as it moves from one function (noun, adjective, verb) to another in a sentence-thus 'a Cartesian', 'Byronic', 'to dun'. Class Two: Many other named-after things are so called simply as a means of honouring those eponyms. The normal procedure with flower species, for instance, is to name a plant after someone not memorably plant- like himself but at least memorable for plant-centred activities. Thus my lobelias are called after Malthias de l'Obel, a distinguished Flemish botanist (who never saw a lobelia in his life, much less was one); my magnolias are called after Pierre Magnol, who wrote a book classifying plants. The names of the units of measurement are much the same. In none of these cases are the named-after things named what they are because they share a property with their eponyms. The very idea is preposterous. But still it is by no means capricious to honour one specific person rather than another by making him the eponym for the specific kind of thing at hand. It would have been decidedly odd to name the unit of brightness the volt to honour Volta, for instance, since Volta never worked on light; Lambert did, though, and so the unit is quite justifiably named the lambert after him instead. Class Three: In a vast number of cases, though, a named-after thing is so called because it belongs to a class originated by the person named. Thus we already have the case of the sandwich; so too braille is the special system of writing for the blind devised by Louis Braille; sideburns are the special kind of facial fungus first sported by General Ambrose Burnside, and so on. Discoverers also are frequently eponyms: thus salmonella is named after the pathologist Daniel Salmon who first identified it and guppies take their name from R. J. L. Guppy who presented the British Museum with its first speci- mens. In each of these cases the specific named-after things may all possess a common property, but it certainly is not a property possessed in common with the eponym, nor is it because the items possess that property that they take the eponym's name. Indeed, the class of sandwiches, for example, can be individuated by the common properties of its members, making no refer- ence at all to their eponym, yet things falling in that class are rightly called "sandwiches" and not something else because Sandwich was responsible for the first of them. The class is formed on the basis of properties shared by each of its members. The name, however, comes from outside and applies for another reason entirely. Class Four: In another kind of case the undergirding association between eponym and named-after thing is much less immediate than this. Thus the nuts I occasionally buy are called filberts, after St Philbert, because that saint's feast day falls on August 22, the height of the season for harvesting them. The type of shoes called plimsolls are so named because the line be- tween cloth and rubber on them resembles the horizontal load line painted This content downloaded on Sun, 10 Mar 2013 10:11:22 AM All use subject to JSTOR Terms and Conditions COMMON PROPERTIES AND EPONYMY 195 on merchant ships and named after one Samuel Plimsoll who led the parlia- mentary crusade for its requirement. And so on ad nauseam.3 Here again we have properties shared by all the named-after things, properties which individuate those things, though not properties shared with their eponym and not properties in virtue of which they take the eponym's name. But what exactly the undergirding relation is varies almost without limit depend- ing on the specific case, and may not even require the eponym ever to have been immediately involved with any of the named-after things at all. Class Five: On the other hand, there is another important type of case where the involvement of the eponym is so immediate that there is no other common property shared by the named-after things except their being so related to it. For example, a vehicle may come as close as it likes in appear- ance or performance to a Honda car yet still cannot rightly be called a Honda unless it was actually manufactured by the company founded by Soichiro Honda. It is not because they share some common property with the epony- mous Mr Honda himself that millions of Hondas are rightly called what they are (as with Class One)-for the man and the vehicles have themselves nothing in common (Mr Honda is not finely engineered). It is not because they share properties with a thing originated by Mr Honda that they are called Hondas either (as with Class Three)-for Honda did not invent a kind of vehicle which other people now can make (no one can make Hondas but Honda's company or those he may license). Indeed, it is not because they share any independently characterizable common properties among themselves that they take the same name (as with Class Four)-for there could well be an infinitely various range of entirely different Hondas (cars, trucks, motorbikes, of all sizes and shapes). If we want to talk about "com- mon properties" at all here, we have to say that the property common to all Hondas is solely their having been caused to come into being by the action of Soichiro Honda. (Note how nicely this parallels the peculiarly dependent status of particulars on Forms: the eponym itself does not possess the relevant property at all, and yet it is a property which none of the named-after things could possibly possess were it not for the eponym.) In precisely the same way you cannot eat off Wedgwood not made by Josiah Wedgwood; you cannot wear a Givenchy outfit not designed by Hubert de Givenchy. Class Six: Other created things are named not after who or what made them but who or what they are about. Every cartoonist in the United States once had his Nixon caricature and the reason why some particular display of pictures in a museum, say, is rightly called a display of Nixons would therefore be because that man is their subject (rather than their author). In like manner heaps of books and shaving mugs are rightly called Lincolnalia after the man Lincoln because they are objects pertaining to him (in fact the suffix 'iana' can be attached to almost any name to turn it into an eponym in this way). 3There are some 3500 (!) more, covering every part of speech, in Robert Hendrickson, Human Words (Radnor, 1972). This content downloaded on Sun, 10 Mar 2013 10:11:22 AM All use subject to JSTOR Terms and Conditions 196 THOMAS W. BESTOR Class Seven: Still other named-after things are named after their source rather than after their creator or his subject (and it is this provenance which individuates as well as names them, as with Class Five). For example, the many different bottles of cognac scattered around the world are all rightly called what they are for no other reason than that they contain brandy distilled from grapes grown in the area around the town named Cognac in France. (At least in the beginning it was so. We are now less pernickety.) There, then, are just a few of the manifold relations which a named after thing can have to a named thing in virtue of which it is quite rightly called after that thing. Though not all are equally likely possibilities for Plato, and though he seems to have other ones in mind too, these classes are suf- ficient to make the point I want to make. Taken as a semantic thesis, Plato's theory of eponymous predication requires that there be some extra relation between a Form and its participants in virtue of which one is named after the other. But it does not by itself require any specific relation. An eponym can be "responsible for" a thing being what it is and being called what it is for all manner of reasons. We have to take the matter case by case. III. DID PLATO HIMSELF BELIEVE EPONYMY INVOLVED COMMON PROPERTIES? Did Plato see any of this, though? The majority opinion is that he did not. The minority opinion is that he did. Plato was no more single-minded in this matter than in any other, certainly, but the balance of the evidence, I think, is in favour of crediting Plato with understanding fairly well that many relationships can and do underwrite the eponym relation. First Fact: As a way into the matter it is useful to notice how unordinary all this talk about eponyms is in English. The word alone makes the blood run cold. Not so in Greek, however. Eponymy was a very familiar pheno- menon. The term was not at all uncommon or technical or off-putting. For example, every Athenian was himself a many times named-after thing and knew it full well. Thus (i) when Kleisthenes re-organized the constitution, one of his major innovations was to institute ten artificial tribes, to which every citizen belonged, as a replacement for the old blood clans. Each man's tribe was named after a tutelary hero of Attica who was known as its "eponymous hero". Indeed, during Plato's prime a monument called "The Statues of the Eponymous Heroes" was erected in the market place for all to see. (ii) Another of Kleisthenes' innovations was to fix by official sanction the names of the individual demes (parishes) that were now the basic social group (usually after a prominent natural feature or the name of their supposed founder), so that each inhabitant could become officially named after his deme. Before, citizens were picked out by their own (first) name plus the (first) name of their father, but now a person's full name consisted of three parts-as in the charge against Socrates: "this indictment is sworn by Meletus son of Meletus of the deme of Pitthos against This content downloaded on Sun, 10 Mar 2013 10:11:22 AM All use subject to JSTOR Terms and Conditions COMMON PROPERTIES AND EPONYMY 197 Socrates son of Sophroniscus of Alopece".4 (iii) Simultaneously every 18-year- old was also enrolled in the official lists of citizens under the name of the current Archon Eponymos, the head magistrate who gave his name to the year in which he held office. The point is simple. In not one of these cases do the named-after things share any common property with their eponym. Yet in every one of these cases there is a perfectly good reason why these specific things (and not others) should be named after what they are (and not after something else). Other cases display other just as "appropriate" relations ('Hecademus'-'Academy' -'academic'). Insofar as Plato was himself an ancient Greek and worked against an ancient Greek background, therefore, he certainly was under no presumption that eponymous predication required common properties. The most familiar facts would have suggested Classes Four and Five as models rather than Class One. Second Fact: Plato's own use of the term 'eponymy', in non-Form-ist contexts at least, demonstrates just as wanton disregard for the majority restriction. The Greek for 'is named after', commonly rendered by three cognates, the passive of the verb cErouvotajco, the noun 'rrcovvlita and the adjec- tive errVcvvvLos, is used to refer to everyday eponym relations some 45 times in the course of the dialogues. Only about half the time, however, do these terms work in the way they are supposed always to work. Start with the ones where common properties do seem to figure. Thus a man is said to be called a glutton after the gluttony in him (Phdr. 238a); crazy after the craziness of his conversations (Symp. 173d); wise after the wisdom present in his writings, lover of wisdom after the serious pursuit those writings show (Phdr. 278d-e). The appetitive component of the soul takes its name from the intensity of its bodily appetites (Rep. 580e). A certain life style is called Pythagorean after the man who was distinguished among his contemporaries for it (Rep. 600b). In the Cratylus etymologies section we have a great list of similar epony- mous predications. There Plato says personal names tell a story or predict a destiny and so "truly" apply to a thing only insofar as the thing actually has that property; otherwise they are "lying" names. "Archepolis" means ruler of a city, for instance, "Theophilus" means beloved of God, "Sosias" means saviour; these names should not be applied to people they don't fit (394c, e, 397b). The same for the names of heroes and gods: "Tantalus" means weighed down by fortune, which he was (395d-e); "Artemis" means hates intercourse, which she did (406b); and so on. The term E'rcovvptla is used in seven of these specific derivations and also twice so as to make the more general point. When Socrates turns to explaining the correctness of general words he employs much the same mechanism: a general word is often phonetically similar to a phrase or compressed statement which speci- fies some particular property and so rightly applies to the things it does only 4John Edwin Sandys, Aristotle's Constitution of Athens (London, 1912), p. 85. This content downloaded on Sun, 10 Mar 2013 10:11:22 AM All use subject to JSTOR Terms and Conditions 198 THOMAS W. BESTOR when and because they have the property so specified. Thus the word for 'god' sounds like the word for 'run' and so was truly applied by the aboriginal Hellenes to their deities because they were always running about in their starry courses (397d). The word for 'wisdom' is phonetically similar to the phrase for 'perception of motion and flux' and applies because the Name- giver, turning around from one thing to the next, got dizzy and thought that the things so perceived were really in motion rather than realizing that the motion was an internal condition in himself (411b-e). Socrates goes through some 60 examples in this manner, ending in a veritable barrage of eponymous one-liners. As before, the term ETrowvvtla frequently carries the weight here (some 14 times). The general pattern so far seems clear enough. Both in the Cratylus and in other dialogues, a thing is named F after its distinctive F-ness. Yet we must not go overboard. Socrates is quite facetious about the Cratylus theory of the "natural fitness" of words from start to finish, and besides, in several of the crucial occurrences of ErrowvvJla the undergirding relation is not actually one of sharing a property after all. Thus, for example, 'good' (like 'wisdom') comes from an expression referring to motion not because it is itself in motion but because the Namegiver (mistakenly) thought it was about motions (412c)-Class Six, that is, rather than Class One; longing too is named after its subject. Passion, on the other hand, is named after its special effects, and mind is called beautiful for the same reason (416d)-Class Five, I sup- pose. Love is named after its source-Class Seven; similarly men are named after their ancestors (397b). All sorts of relations carry the weight, then. Nor are things much better in the non-Cratylus cases. Is it really perfectly obvious, say, that Socrates shares a common property with his eponym craziness and so can be called after it? Are there three things here: Socrates, craziness, and the distinctive property which the eponym craziness possesses and that Socrates also possesses? The real equalizer, however, is the fact that these are by no means the only cases of legitimate eponymy that Plato entertains. Thus he uses exactly the same set of expressions to repeat the traditional view that Athenians take their name from the goddess Athena whom they venerate (Laws 626d). Likewise the ideal city Plato plans will be called after some god, probably the local deity of the area in which it will be constructed (Laws 704a, 969a), and each of the 12 tribes that make up its population are to be named after the god to whom it has been allocated and consecrated (Laws 745d-e, 828c). The ideal city apart, Plato says that the name of a religious temple usually comes from the name of the god or goddess to whom it is dedicated (Laws 738b). Stars have been named after the various gods and goddesses too, as a means of honouring them (Epin. 987b). The art of harping, singing and dancing (i.e., music) is named after the Muses because they are the specific goddesses that foster that specific art (Alc. I 108c-d); alternatively, Plato suggests the Muses themselves might be so called because they are the main This content downloaded on Sun, 10 Mar 2013 10:11:22 AM All use subject to JSTOR Terms and Conditions COMMON PROPERTIES AND EPONYMY 199 goddesses of the race of Ligyans renowned for their musicianship (Phdr. 237a). The island of Atlantis and also the ocean it is in, the Atlantic, are both so called after the name of the first king of that island, Atlas (Critias 114a). A child is usually named after his kin (Lys. 204e; cf. Rep. 330b, Pol. 257d-258a), but in certain circumstances can rightly be named after someone of another blood line (Laws 878a). Since it is a general principle that a city's constitution takes its name from the name of the ruling class in it, the ideal constitution should properly take its name from that of the god who is master of ideal (i.e., rational) men (Laws 713a). By contrast with the other eponymous predications, in these it makes no sense to look for some property the named thing shares with the named- after things in virtue of which it rightly gives them its name. The island of Atlantis, for instance, possesses no distinctive property in common with the delni-god Atlas; yet that does not somehow make it inappropriate for the one to be named after the other-being the king of is a perfectly good relation for undergirding eponymy. So is being an object of veneration. So is fostering or nourishing. So is being the ancestor of. All in all, then, we have some 20 or so occurrences of commonplace "eponymy" in the dialogues which trail no clouds of common properties, 25 or so which may. This is hardly convincing evidence that Plato himself felt especially constrained to explain why X is named after Y in terms of X possessing that distinctive property which Y possesses. Third Fact: When we turn to distinctly Form-ist contexts I think we have to come to the same conclusion. The textual proof of this can be put in a surprisingly simple argument: (1) Throughout the course of his dialogues Plato seriously enter- tains only one basic theory of predication, eponymy: from start to finish it is by being somehow or other related to the Form that a particular comes rightly to take what is in fact the Form's name. (2) At the same time, throughout the course of his dialogues Plato seriously entertains a great many different accounts of the basic relation holding between Forms and particulars, some in no way reducible to sharing a common property. (3) Therefore, Plato must have realized for himself (what is in fact true) that the theory of eponymous predication is compatible with several ontological models and does not require the backing of the common properties model. Let me try to unpack this a bit. Undoubtedly the clearest expressions of Plato's basic account of predica- tion occur in the Phaedo, Parmenides and Timaeus (Loeb translations: the italicized phrases translate variations on the term e'rcowvla): As I remember it, after all this had been admitted [that nothing can come into existence other than by participating in the proper essence of each thing, beautiful things in Beauty and two things in Duality, for example], and they had agreed that each of the abstract qualities This content downloaded on Sun, 10 Mar 2013 10:11:22 AM All use subject to JSTOR Terms and Conditions 200 THOMAS W. BESTOR exists and that other things which participate in these get their names from them, then Socrates asked . . . (Phd. 102a-b). We said before that in the case of concrete things opposites are gener- ated from opposites; whereas now we say that the abstract concept of an opposite can never become its own opposite. . . . Then we were talking about things which possess opposite qualities and are called after them, but now about those very opposites the immanence of which gives the things their names (Phd. 103b). Well, tell me: do you think that, as you say, there are ideas, and that these other things which partake of them are named from them, as, for instance, those that partake of likeness become like, those that partake of greatness great, those that partake of beauty and justice just and beautiful? Certainly (Parm. 130e-131a). We must agree that One Kind is the self-identical Form, ungenerated and indestructible, . . . invisible and in all ways imperceptible by sense, it being the object which is the province of Reason to contem- plate; and a Second Kind is that which is named after ((ojLcvvlzov) the former and similar thereto, an object perceptible by sense, generated, ever carried about, . . . apprehensible by Opinion with the aid of Sensation (Tim. 52a). In fact, however, the theory is given expression as early as the Lysis and as late as the Laws and changes not one whit from first to last.5 Later authorities are also unanimous in reporting the same single theory of predication.6 In sharp contrast to this constancy in his basic account of predication, Plato is notoriously fickle about just what specific relation it is that partic- ulars must have to Forms in order to be named after them. He rings the changes on some 15 technical terms; no single model is specially consecrated either. (i) Thus in the Hippias Major, much as the ingredient alcohol is in gin and whisky and thereby makes them "alcoholic", so beauty is said to be something that can be "added to" particular things (289d, 292c-d), almost drop by drop; when it is, that explains why the examples are entitled to the name (288a, 303d-e). (ii) In the Euthyphro the Form of F particulars has become the character common and peculiar to a set of examples whose exemplification puts them all in the same pot (5d, 6d-e). There is only one such real Form: it is the defining character of the F things, the word 'F' is its proper name, but it has no independent existence separate from the particulars that exemplify it (unlike an ingredient then-were there no alcoholic drinks there might still be some distilled alcohol somewhere; were there no pious deeds, however, there simply would not be that character common and peculiar to pious deeds). (iii) In the Republic a Form is con- sidered to be itself a discrete particular, of a kind; it is unlike the more 5Lys. 219c-220b, Laws 963a-964 with 965c-d. Also see Phd. 78e, 102c, Phdr. 249b, 250e, Parm. 133d, Soph. 234b, 252b, and Tim. 41c. 6See, for example, Aristotle (Metaph. 987a 31-b 15, Eud. Eth. 1217b 2-14); Alexander of Aphrodisias (Comm. in Arist. Metaph. 82. 11-83, 16); Proclus (Comm. in Plato Parm. 125); Asclepius of Tralles (Schol. in Metaph. 567a 41); and Diogenes Laertius (Lives III 13). This content downloaded on Sun, 10 Mar 2013 10:11:22 AM All use subject to JSTOR Terms and Conditions COMMON PROPERTIES AND EPONYMY 201 mundane particulars we meet with every day, to be sure (it is non-sensible, not in physical space, eternal, unchanging, divine, and so on), but still it is one specific 'entity. Such Forms are the inhabitants of a special world which we can gaze upon with our souls and which the inhabitants of this world can be compared against and named after. They are prototypes which the particulars try to pattern themselves on (472b-c, 484c-d, 500d, 501b), para- digmatic objects which particulars copy more or less accurately (519c-520c, 597b-c), perfect entities that particulars are mere shadows of (510b-d, 515c). (On such models a Form is nothing at all like the common character Bed- steadity which my bed has, say, but more like the First Bed which my bed yearningly likens itself to.) (iv) The Cratylus makes a Form not some kind of perfect particular but rather an abstract rationally intelligible design or pattern, more on the lines of a schematic diagram, a chemical formula, a list of instructions or architectural specifications. Thus if we want to make a new shuttle to replace a broken one we have naturally to look to the real shuttle, the Form; but this "looking" is not gazing at a perfectly shaped piece of something but attending to the functions of shuttles and to the specific requirements a shuttle must embody to perform that function best (389a-390e). (v) In the Timaeus appears a model of the Form as an ancestor. Things of this world in a process of generation are like children; the receiving matrix in which all this generation takes place is like the mother; the source of the generated things' being what they are is like the father. In the rather special sense that kin are "resemblances" of their primogenitor, the "races" of particulars are thus "resemblances" of Forms and named after them accordingly (50c-e). Contrary to common opinion, none of these various models wholly displaces any other in the dialogues. For example, the in- gredient model is present not only in Hippias but also in Laws 963a-965d. Common and peculiar characteristics show up at Meno 72c but also at Sophist 247a-b. The copying account appears in Phaedo 74b-75b and again in Politicus 285e-286a and Epistle VII 343b-c. The functional specifications view occurs in the Cratylus but also at Republic 601c-e and Timaeus 28a-29b. The primogenitor model occurs at Timaeus 36e and also at Hippias 297a-b and Politicus 258a. This gets us to the heart of the problem. Plato takes seriously several ontological models. But he takes seriously only one semantic theory. The obvious conclusion is that Plato felt no more committed to a single relation undergirding eponymy than he felt committed to a single characterization of "participation"-which is to say the same thing of course. And lest it be said that all these models "really" boil down to the single sharing-a- common property model nonetheless, just pause a bit. What property does a shadow, on rocky ground say, have in common with the object that pro- jects it in virtue of which we say that it is this particular shadow which is the shadow of this particular man?-clearly it is who projects it that matters. Again, what property does a functional specification necessarily have in This content downloaded on Sun, 10 Mar 2013 10:11:22 AM All use subject to JSTOR Terms and Conditions 202 THOMAS W. BESTOR common with the thing that embodies it? What property does a primo- genitor necessarily have in common with his offspring? The truth is that for Plato it seems not to matter what precisely the appropriate relation is, just that it is because and only because sensible particulars are somehow "appropriately" related to Forms that they have the character and name they do: I cling simply and straight-forwardly and no doubt foolishly to the explanation that the one thing that makes that object beautiful is the presence in it or association with it, in whatever way the relation comes about, of absolute beauty. I do not go so far as to insist on the precise details-only upon the fact that it is by beauty that beautiful things are beautiful (Phd. 100d, Tredennick translation; cf. Tim. 50c, 51b). If Plato does not insist on the precise details, neither I think should we. IV. THE CASE FOR A COMMON PROPERTY RESTRICTION ON EPONYMY At least this should be our conclusion in the absence of any convincing evidence to the contrary. The majority opinion believes that there is evi- dence to the contrary, however, and White spends much of his recent article marshalling it. So this side of things has to be looked at too. (1) White's first point is the expression of a widespread incredulity: if the implications of the [unrestricted] image analogy and the epony- my theory of predication had been clear to Plato, it is not plausible to suppose that he would have hesitated to spell out their relevance to the Third Man and other regress arguments. It is too high a compliment to the general run of his readers to argue that Plato would have taken for granted that they would easily spot what he took to be an obvious fallacy. It is too little complimentary to Aristotle and others unable to see in what way Plato could have dealt with the arguments (p. 208). We cannot, I think, put any weight on the argument from Plato's silence here. The whole of the Parmenides is a hotch-potch of good and bad argu- ments from beginning to end and Plato never offers any commentary on any argument or any antinomy. Socrates is the official yes-man in this dialogue and he does his duty devotedly and uncritically. He nods at superb pieces of reasoning. He nods at outrageous fallacies. That he nods at the Third Man as well therefore tells us nothing. Actually, the conclusion that Plato himself thought the Third Man a devastating objection to the theory of Forms runs into even worse difficulties than the conclusion that he thought it no objection at all, for Plato continues quite openly to use both of the main accounts of participation supposedly devastated by it.7 Moreover, in fact Plato does offer some hints that the Third Man is specious. He prefaces it and the rest of the arguments in the first section with the clearest ever 7For example, the sharing account at Soph. 240a, 247a-b, 248c, Tim. 65e-66a, 77a-b, 90c, Pol. 275c, Laws 965c-d; and the copying account at Soph. 234b, 235d-e, Tim. 31a, 50c-d, 51a, Pol. 285e-286a, Laws 962a-b, Epist. VII 343b-c. This content downloaded on Sun, 10 Mar 2013 10:11:22 AM All use subject to JSTOR Terms and Conditions COMMON PROPERTIES AND EPONYMY 203 statement of his theory of eponymous predication (130e-131a)-which theory does in fact break the back of each and every one of those arguments. And then he goes on to insist that Socrates is a "quite young" man (127c, 130e, 135d) and for that reason cannot refute his objections as a sufficiently "gifted man" with the "proper training" (in logic) easily could (133b, 135a-d). Curiously, the argument from Aristotle's loquacity here is in even worse case. White believes that Aristotle offered his famous battery of Third Man arguments because he saw (as eventually did Plato) that naming-after is univocal (we are indicating the possession of the same property both times). The astonishing fact, however, is that Aristotle himself thought precisely the opposite! Every case where we apply a single predicate to both a copy and its original is a case of equivocation pure and simple, he insists. Thus, for example, a saw is defined by what it can actually do. But a saw in a painting cannot actually do anything. Hence it is really no more a saw in the sense of a cutting instrument than a wise saying is a saw in the sense of a cutting instrument. Accordingly, if Plato is serious in believing that particulars in this world are copies of Forms in another world, as his "intriguing meta- phors" (Metaph. 991a 20-23) suggest, then, Aristotle says, when we apply the same word 'F' both to a particular and to the Form f, we have committed a simple 'saw'/'saw'-type equivocation.8 This is so even when we name the particular F "after" the Form S. (2) The most important of White's arguments is that Plato's own reflec- tions or shadows analogy itself shows that eponymy must be backed up by the Form's possessing properties in common with their participants. For several reasons. Because it is implausible to suppose we are using a different language when we come to speak of mirror images and the like: we might imagine two persons being taught the meaning of the word 'red', the one being shown red scarves, the other being shown mirror- images of red scarves. Is it not implausible to argue [as Aristotle did] that the two are being taught quite different meanings of the word 'red' (p. 208)? Because otherwise how could we possibly confuse reflections with their originals? Surely, in the case of images, what makes something an image is precisely its possession of so many properties in common with its original that, given the right circumstances, a man might even mis- take one for the other (pp. 209-10). Because particulars serve to help us to recollect the Forms: suppose we ask: how do like things remind us of like things? How do images remind us of their originals? Is it not by presenting us with a selection of the very properties that the originals themselves possess (p. 209)? 8Compare Meteor. 389b 30-390a 14, Pts. of Animals 640b 30-641a 6, On the Soul 412b 30-22, Categ. la 1-5, Pol. 1253a 20-25, Gen. of Animals 726b 22-24. This point is Owen's discovery in "Logic and Metaphysics in Some Earlier Works of Aristotle", op. cit., pp. 187-8. This content downloaded on Sun, 10 Mar 2013 10:11:22 AM All use subject to JSTOR Terms and Conditions 204 THOMAS W. BESTOR These points are not, I think, as strong as they might seem at first glance. The reflection view which underlies all of them, and which is certainly the Official Doctrine, is what might be called a "static" one. We start with the familiar mundane objects of our ordinary world, sensible particulars. In addition, Plato has discovered, there are also certain other non-sensible Forms, belonging to a different, higher world. Each has a determinate nature of its own. It turns out, however, that the two are also related to each other-the familiar objects of the sensible world turn out actually to be dummies, simulacra, imitations of the other really real things. This is a sensible, coherent view, but it is not Plato's view-at least not in the main passages (Phd. 74d-75b, Parm. 132d-133a, Tim. 48e-52d). Plato's own imitation model is a decidedly "dynamic" one: copying a Form is much more like copying a master's tennis technique than like copying by Xerography. The crucial feature of particulars is not that they have a (static) nature like the nature the Form has, but instead that they are in a (dynamic) process of constantly likening themselves to the Form. Particulars ceaselessly attempt to imitate the Forms but just as ceaselessly fail to do so successfully -and so ceaselessly fail to achieve a plateau where they actually possess what the Form possesses. For example, a pair of particulars we call equal are so-called because what they "aim at being like", what they "strive after", is the Form of Equality; being particulars, however, they inevitably "fall short"; and so, naturally, they never come at last to share any supposed equal-making characteristic with the Form (Phd. 74d, 75a). What White has done to Plato here is what Parmenides did to Socrates long ago (in the "second" Third Man at Parm. 132d-133a). Both replace a process with a state: "Is there an activity of striving to be like a Form going on? Yes. Well, consider that activity as having reached just the state it has reached in relation to its target. Freeze it there. Now insofar as the particular has reached just that state, it must bear some achieved resem- blance to the Form. That is, it must possess in some degree the property which the Form possesses to an ultimate degree." A clever ploy. A typically Eleatic ploy.9 But in other contexts, like Zeno's arrow, it is an admittedly fraudulent one. Why then should it be particularly acceptable here? If we really do freeze a copy, refuse to consider it as something in process but consider it just on its own, clearly it is no more to be compared with one specific original than with any other-frozen, everything in the universe is like everything else. The alternative, dynamic, view usually makes at least as much sense: what determines that some particular is a copy of this rather than that Form is not that it possesses something which the Form is distinctive for possessing, but rather that its activity is directed to this rather than to some other target. (Think of Class Six eponyms here.) (3) White too shares the Official Doctrine that Plato's general theory of 9As Edward N. Lee shows in "The Second 'Third Man': An Interpretation", in J. M. E. Moravcsik (ed.), Patterns in Plato's Thought (Dordrecht, 1973), pp. 102-6. This content downloaded on Sun, 10 Mar 2013 10:11:22 AM All use subject to JSTOR Terms and Conditions COMMON PROPERTIES AND EPONYMY 205 causation depends on shared properties: whatever is the "cause" in another thing of a property, F, must itself possess F-it must be true of it that it is F. Clearly if this interpretation is correct it follows at once that in Plato's mind the Forms are satisfactory as "causes" of properties precisely and only because the Forms themselves first possess those properties. But let us suppose for the moment that the narrower interpretation is accept- ed, namely, that if something is to be the "cause" of F it must not be Opposite-F, even if it does not itself positively possess F. On this interpretation too it is plain that if Plato rejects other sorts of things as "causes" on the grounds that they possess the wrong kind of properties (e.g., he argues that a head cannot be the "cause" of a man's being taller than another because a head is something small, Phd. 101a), then in some sense he means that the Forms are acceptable as "causes" because they have the right properties (p. 209). This is so. Of course. But consider the variety of Class Three and Five eponyms a moment, for instance. Hondas certainly cannot be caused to be by just anything. Flatworms cannot sire them. They require the "right" sort of cause, with the "right" properties for causing such a thing. Yet whoever would think that the "right" properties for producing Hondas include being a car-or even not being an opposite-of-car? The whole idea is utterly ridiculous. Why must we pin such idiocy on Plato, even in the apparently idiotic passages of the Phaedo? Having bones rigid and separated at the joints is the wrong sort of property for causing Socrates to be sitting in a bent position in prison (98c-d). Conducing to the overall goodness of the universe is the right sort of property for causing that (98b). So is being condemned by the Athenians (98e). Here, in Plato's own examples, "right- ness" and "wrongness" in a cause have nothing at all to do with sharing or not sharing a property with the effect. Nor probably have they anything to do with most cases where an agent brings about something (you cry because I hit you) or where one thing grows into another (an acorn into an oak)-and these, not the billiard-ball conservation-of-momentum sort of case, were the model cases for causation in Plato's time. (4) The fact that Plato often says the Forms can never bear their opposites once again shows that the Form 0b must itself have the property F-ness, White argues: If a particular, x, is F it is also Opposite-F. But if a Form, b, is F it is not also Opposite-F. If a particular man or act is just, the man or act is also in some respects unjust. By contrast, Justice itself is just without being in any respect unjust. Now on the supposition that by 'b is F' Plato means 'b is identical with b' [rather than 'd possesses the property F'], the contrast he draws between Forms and particulars turns out to be curiously lacking in symmetry. For in place of a readily intelligible contrast (a, b below), we have one that reads very awkwardly (a', b'): (a) Particulars possess their properties in some respects only; in other respects they possess the opposites of those properties. This content downloaded on Sun, 10 Mar 2013 10:11:22 AM All use subject to JSTOR Terms and Conditions 206 THOMAS W. BESTOR (b) Forms possess their properties without qualification; they ex- clude the possession of the opposites of these properties. (a') Particulars possess their properties in some respects only; in other respects they possess the opposites of those properties. (b') Forms are identical with themselves and are never non-identical with themselves . . . (p. 211). It is true that there is a fundamental "lack of symmetry" here: 'is F' said of a particular attributes the property F-ness to it, 'is F' said of the Form b does not. However, this asymmetry is hardly "curious" since it is precisely what we expect of a semantic theory designed to allow a predicate (be it 'F' or 'Opposite-F') to apply for either one of two quite different reasons-because x just is itself the relevant Form or else because x is ap- propriately related to the relevant Form. Nor is the asymmetry in the least bit "awkward". White, like so many others, assumes that because 'b is F' expresses a "proposition of identity", as he calls it (p. 206), it expresses the proposition 'the Form b is identical to the Form s'. This is what makes '< can never be Opposite-F' into the admittedly awkward (b') after all. But the assumption is really very strange. There are all kinds of propositions of identity. So why choose the one that is patently empty, tautologous, hardly ever worth saying? Why not choose one that is at least significant and synthetic and gives us something we should like to know? In 'o is F' Plato is in point of fact telling us what manner of beastie F is: the Form + just is itself the property F-ness. What we think is commonplace property talk, that is, is actually, upon proper analysis, talk about rather ethereal and sophisticated philosophical animals (Forms).l0 Accordingly, the most reasonable translation of the so-called "proposition of identity" is not 'The Form b is identical to the Form <' but 'It is the Form b which the word 'F' ultimately hooks up to, not anything less (like a transient sensible, for instance)'. This is exactly the analysis which eponymy also suggests: 'The Form b is the nominatum of 'F' '. And on this translation the explanation of 'b can never be Opposite-F' is not awkward at all. Any particular can be both F and Opposite-F because such terms apply to it only at second- hand and the particular can readily participate in both named Forms, the Form b and the Form Opposite-of-+ (e.g. Parm. 129a-b). However, if the Form S itself were to be both F and Opposite-F, since such terms apply there directly, as proper names, this would mean that the different words 'F' and 'Opposite-F' would between them name only one thing-i.e., they would have the same meaning. In fact, of course, they plainly have quite different meanings-i.e., they name quite different things (e.g. Phd. 103b-c, Parm. 129c-130a).1 10As H. F. Cherniss, "The Relation of the Timaeus to Plato's Later Dialogues", in Allen, Studies, pp. 370-71, and Allen, "Participation and Predication in Plato's Middle Dialogues", op. cit., pp. 44-6, have insisted all along in their "identity proposal". 1This is K. W. Mills' point in "Plato's Phaedo 74b 7-c 6", Phronesis, II (1957), pp. 145-147. This content downloaded on Sun, 10 Mar 2013 10:11:22 AM All use subject to JSTOR Terms and Conditions COMMON PROPERTIES AND EPONYMY 207 V. CONCLUSIONS I dare say the truth is probably that Plato sometimes did think that the Form b and its F participants shared the common property F-ness and some- times did think that this was the reason why the particulars were rightly called after that Form and not some other. But this is not an unsettling conclusion. It would be downright silly to insist that Plato never backed up eponymy with common properties. What I have tried to show instead is (1) that there is nothing about eponymy taken as a semantic theory which requires such a backing up, (2) that there is no evidence Plato himself ever believed eponymous predication was limited to shared properties cases, and hence (3) that there is no reason to believe Plato's Third Man expresses his final realization that limiting it thus proves fatal. Massey University, New Zealand This content downloaded on Sun, 10 Mar 2013 10:11:22 AM All use subject to JSTOR Terms and Conditions