By the same author




to Wiltgenstein 5 Tractatus

Three l~hilosophers (wit11Pe~cr Ceach)

From Parmenides to Wittgenstein


Basil Blackwell

. Oxford

Q in this collection G . E. M. Anscornbe 1981


First published ill 198I b) Basil Blac.kwel1 Publishe~. 108 Cowley Road 0xli)l.d 0 x 4 1.11. England All tights reservrd. N o par1 ol'this publication may I)e rcprotluccd, stored ill a retrieval sysrc-111, or transi~~itted, ill arty forln o r by arty Inrans, c:lrctronic, mechanical, pllotr>copying,~.ecordir~/: o r otherwise, without the privr perlllisrion ol' Basil Blackwell Publisher Linlitctl.

Introduction PART 0NE: The Ancient Greeks

Parmenides, Mystery and Contradiction The Early Theory of Forms


3 The New Theory of Forms
4 Understanding Proofs: Meno, 85dg - 86c2, Continued

British Library Cataloguing in Pi~bliration Data Anscotrlbe, Gertrude Elirabeth Margarel The collected philosopl~ical papers of G . 1:. M. A ~ i s c o ~ n l ~ e . Vol. I : Fro111Par~rier~itles lo M'it~genstcill I . Philosophy. English - Adrlrcsses, essays, 1cctnrc.r I. 'Title lgnf.08 81618
ISBN o-631-1ngq2-7 Typesrt

5 Aristotle and the Sea Battle: De Interpretatione, Chapter IX Appendix: A Note on Diodorus Cronus 6 The Principle of Individuation
7 Thought and Action in Aristotle: What is 'Practical Truth?'

Photon Baskervillr

PART Two: Medieval and Modern Philosophers 8 Necessity and Truth g Hume and Julius Caesar

"Whatever has a Beginning of Existence must have a Cause": Hume's Argument Exposed Will and Emotion Retractation



13 The Question of Linguistic Idealism Index

I am very grateful to Professor C. H . von Wright for permission to include in this volume the essay 'The Question of Linguistic deali ism' which 1 contributed to his Festschrift published in Acla Philosophica Fennica 28, 1-3

When I wrote the first essay here reprinted, on Parmenides, I said that no one (so far as I knew) had reasoned thus: Only what can be thought can be, What is not cannot be thought, Therefore what is not cannot be. 1 was forgetting the celebrated argument of the Sophists, recorded for us by Plato: He who thinks, thinks something, He who thinks what is not, thinks nothing, Therefore he who thinks what is not is not thinking. which is certainly an argument for What is not cannot be thought understood in the sense: Whatever is going on, it can't be thinking if, if it is thinking, it is thinking what is not. However, the ground for this is already that what is not is nothing, and this would lead to "What is not cannot be", unless we think that what is nothing can be. It is more natural to hold that what is nothing cannot be than to go via "Only what can be thought can be" in order to reach the conclusion that what is not cannot be. Parmenides himself argues: What can be thought can be, What is nothing cannot be, Therefore whatever can be actually is. Therefore whatever can be thought actually is. All these arguments, except the first one, use as a premise: What is not is nothing and hence do not derive the nothingness of what-is-not from its unthinkability, but rather its unthinkability either from its nothingness or from its impossibility. The impossibility of what is not isn't just the impossibility of the proposition "What is not, is" - i.e. the truth of "What is not cannot be", taken in sensu composilo. That could be swept aside as irrelevant. What is not can't be indeed, but it may come to be, and in this sense what is not is possible. When it h a come to be, of course it no longer is what is not, so in calling it possible we aren't claiming that "What is not is" is possible. So it can't be shown to be impossible that it should come to bejust by pointing to the impossibility that

arltl so thrre'd be all exaiiiple of "Wliat is not is" that was true. does not exist.for that is riot tcasible --nos could you express it". being false. when not p. and something besides the one. All names which are names of the same mean the same. not a difference in what is thought of. partook of not being.e.ject. Knowledge 01' what is not is not possible. The fact that being is called by other narnes than "being". but only of what involvrs change and non-being. But some such distinction is necessary. of 'the other of being'.. all of whose properties are identical with his being. the only beings. So Parmer~ides is willing to call his b e i q "vne" and "changeless" and "contiriuous" ant1 "like the mass of'a well-tllrned ball". nut l i ~ thought r the impossibility is based on the impossibility of the suggestetl object. and so on. if the name of the oiie is not thc same as the one itself'. being light not being heavy. Therc is solnething besides being. For this was not the same as what absolutely is not. will lead Plato to argue that one and is must be difl'erent: otlieiwisr \ve coultl as well say "one one" as "one is". way round. i. We may count this just as a rejection of "more here. O r again like Spinoza's one and only substance. bccausc knowledge is only of' what is. If we have two names. on . Thus the form of the one is different - . the ar~cients rlever argued from constraints on what ( oulcl be a thought to restrictions on what could be. but not tor the saine thing. because together they will divide being from itself. Parmenides' behalf. It is true that Parmenides tends to avoid it. unless we may not make the distinction. So Plato's argument fails. Thus [lie iaipossibility ofkriowledge hrre does not have to he based on tlic in~possibility of the suggested object of knowledge.Here the condeinnatiotl is not of just ariytl~ingthat may be said about what is. corrie to be and pass away. Differenceof sense tbr these two narnes would suffice for non-substitutability in "The name is the name of the one. and so he thinks thc non-identity o f one and being ant1 the name of the one is proved. brings us into the domain of negative predicates actually holding and so of there being such a thing as not-being such and to. then according to him it is not being. i. Ol'course this sentcnce like "one one" is nonsense. or what is.antl that is nothirig but being. and this. But its nonsensicality doesn't show that. He regarded fbrms as beings.e. If it were about soii~ething. that car1 be" is about nothing at all. Plato's escape comes in his Parmenides and Sophist. then at least one of them cannot be a name of being. Parmenides' one being is like the one God. but only thc oillei. We might get out ofthe difficulty by saying that "one" and "bring" have different senses but still mean the same ot~.But this can't be the whole story.e. Sensibly apparent differences of names will be illusory: this is perhaps the hardest consequence. "So anything rnortals sel up for themselves will be a nairle . He also took it at this late stage that he had to give an account of not being and of the possibility that a thought or proposition. S o "What is not can be" taken tn \en\u dlvtso. In particular.g. then it would be about sorncthinp: that is lot. if "one" and "bei~~g" hoth stand for somrthirlg. less there. but he certainly thought there were many of them. in the first sentence and the same sentence with '>anew replaced by "his sister". between sense and object meant. but he cannot d o without it altogether as when he says that his being is not more here. "Jane" and "his sister" don't mean the same object.vl11 . and so the escape was not a presupposition of his start in philosophy or of his earliest theory of forms. and the difference of sense can be taken as a difference in ways of thinking of the one object. "What can be thought and thought's result are the samc. I think I made a false deduction that negative predication was altogether excluded by the objection to not being. Only in one place Parnlenides states the simple impossibility otthe n~eiital condition: "You could not know what is not . Also." That this could make sense if. If 1 am right. and came to think they differentiated being into parts. Just as a product of substitution the sentence is nonsense. These are supposed to be late dialogues. the name of'thc one has not been proved to be diBerent f r o n ~ the one by lion-substitutal)ility of tllr narne "name" and the name "one". That is to say.except that one nume is not another. in the sense that one cannot be replaced by thc other. But there can be nothing against the one being having a multiplicity of names . such that being phi excludes being xi. preferring privative prefixes as in "unchanging". One platonic form was always different from another precisely as an object named by the name of the form. sucll as "011c". to cllange place and to alter briglit colour. i. it seems. That being fire involves not being earth. except indeed that Parmenides has no room for different modes and affections of the one being. less there"." Similarly. there isn't any such thing. i. Silnilarly the {act that the onr i~vhich is the name Parrnenides' being) has a narne gives Plato his chance to argue ~liat is not the saine as the one: otherwise we could substitute in "The name is the name ofthe one" and get "The one is the one of the name". is impossible tjecause if A knows that p. e. "his sister" meant "thut one's sister" is irrelevant. Parmenides really did obiect " If there is any thing that is not Parmenides' being. Introduction Introduction ix it is." he says . or else Parmenides would hardly be able to say anything about his one being. it itself has being 'running through' it. phi and xi. . as everybody will agree. namely as: "Concerning what is not. knowledge that p. But that can be taken as a difference of sense. ifwe have: "The law forbids Jane to marry John" we might replace either "Jane" by "his sister'' or '30kln" by "her brother" but if we replare both at once we get the non-sense "The law forbids his sister to marry her brother. if so there are not many beings after all. the11 p. such." . This suggests that all the different names of things that there are arejust different names of that one being. That what is not i5 nothirlg iiriplirs that there isn't anything to collie to be.e... Plato's ad hontinem arguments rnade a breach in the wall. t o Le and rlot.

He would have done better to built into the verb. and needs must not be'. saying 'is. as well as being other . something being or some being thing. equivalent to the infinitive "to be". . das kann man nicht denken" : an impossible thought is an impossible thought. Wittgenstein himself in the Tractatus has language pinned to reality by its (postulated) simple names. and must be'. This needs argument. we should not move slickly here: "being" might be an abstract noun. "being" is the only term that expresses being. In English the lack of a subject may be fbund disturbing. It is thus possible for the same not to be other.x Introduction Introduction xi from the form of'beirig. and the answer to this is thought to be an explanation of meaning. We have seen what the argument is: what is not is nothing. The subject. for it is the same as what can be. It appears again in Aquinas' discussion of divine knowledge of the future. In Descartes it is reflected in ' the assertion that the descriptive terms which we use to construct even false pictures of the world must themselves stand for realities-even if the pictures are of nothing. with Wittgenstein in the Tractatus Logico-Philosophicur. For Plato significant terms other than "being" are names of beings which are other than being itself but not the~riselves therefore non-existent. arguing only that a thought was impossible because the thing was impossible. the problem of negation does not seem to have been so prominent. Therefbre it is often translated "It is". However. and can't be'. The assumption common to Plato and Parmenides is an ancestor of much philosophical theorizing and perplexity. and says that that is how they are (whether it is asserted or not). When we say "one" we are speaking of something other than when we say "is" or "same" or indeed "other". But the identification of what is truewith what must be true remained a focus of discussion. For being 'runs through' them all. as Parrnenides would have it. After Plato. Brentano thinks that the mere predicative connection of terms is an 'acknowledgement' (Anerhennung):he apparently forgets at this point both that predication need not be assertion and that assertions may not be true. we get his wilder results. or what something's being thus or so consists in. we are apparently committed to a kind of idealism. but tor Parmenides such other terms as are not names of nothing are other names of being. but rather being. even such a term as "a cause" as it occurs in "A beginning of existence must have a cause. . which mean simple objects. In Aristotle its descendant is the theory of substance and the inherence in substances of individualized forms of properties and relations of various kinds. she. In Hume there is the asSumption that 'an object' corresponds to a term. and it is not possible for what is nothing to be. But Parmenides does not treat to be as an object. as we see Hume to have done. and cannot not be".he. There is a siniilar difficulty about Parmenides' description of the two paths tbr thought: "is. It was left to the moderns to deduce what could be from what could hold of thought. This trend is still strong. whitehead's remark about Plato might. But when we cornbine this with the idea that being is an object. so they characterize it as well as naming it. i.tbr it is other than other itself Parmenides' complaint is verified: being and not being are thought to he the sarnt: and not the same. or they. it. But there is no indication in the' Greek that "it" is the right subject. and what can be thought of must be. At the present day we are often perplexed with enquiries about what makes true. "Was man nicht denken kann. perhaps Plato's rebuttal of Parmenides was accounted sufficient. and we might get closer to the sense by saying "what is". then he seems no more than good sense. that the proposition shows how things are ifit is true. Therefbre I would rather not give a subject word.e. . as the combination of"is" with "cannot not be" and of "is not" with "cannot be". What they express is what is true of being. as the Tractatus puts it. the other 'is not. "These are the only ways of enquiry for thought: one 'is and cannot not be'. one's thoughts can only go two ways. which therefbre does not seem incornplete without a separate word for a subject. . This can be seen from Aristotle's discussion of future contingents and the Master Argument of Diodorus Cronus. But the Greek does not need a subject-expression. or. The noteworthy thing abodt this is not so much the ungiven subject. the colours in which they are painted are real. or 'is not. What Plato has in common with Parmenides is the assumption that a significant tern1 is a name of an object which is expressed or characterized by the term." And he is also convinced that what can be lacking in a thought can be lacking to the reality that the thought is of. For Pannenides. But the ancients had the better approach. be applied to his great predecessor: Subsequent philosophy is footnotes on Parmenides. If we take Parmenides as simply warning us OKthe path of thinking there are things that do not exist. and "is not and needs must not be"." That is: Whatever enquiry one is making. as for Plato. It is difficult to use the participle in English in the required way. Aristotle and his greatest scholastic follower are marked by an indeterminism about such future events as do not have necessitating causes already in the present: this stance involves them in logical problems. somewhat narrowly. and so both whatever can be must be. If there is no external answer.

Part One The Ancient Greeks .

Mystery and Contradiction . it holds that that cannot be thought. Besides. Where the picture can be grasped without oscillation of attention that is difficult to substantiate. saying: What have you mentioned? Nothing at all. From Proceedings ojthe Aristotelian Society (1969). though only the former is relevant to Parmenides' argument. Pictures of impossible states of affairs seem to favour the former. and so.1 Parmenides. Descartes would say these don't give us clear and distinct ideas. The impossibility of thinking that there is such a thing as the picture presents cannot. rejecting. an essentially incoherent thought. . He has to formulate it somehow in order to show it up as incoherent. But this is the difficulty any philosopher is in. But it has no credibility except in sensu composite. (2) He might support the second premise taken in sensu diviso: Concerning that which is not. The first premise is false if either that can be thought which cannot be or that can be which cannot be thought. be offered as proof that the idea is not clear and distinct. The conclusion is also incredible. From this conclusion or just from the second premise he is able further to infer: ( 1) A false thought is impossible A negative thought is impossible (3) Change is impossible (4) Past and future. we may add. past and future tensed thoughts. Parmenides' argument runs: It is the same thing that can be thought and can be What is not can't be :. are impossible ( 5 ) Differentiation is impossible. Parmenides might be taken to mean: the impossible can't be clearly conceived to be. by reasoning. of course. And that's my point! Admittedly the same holds of the conclusion: Concerning that which is not.What is not can't be thought This is valid only if the second premise is taken in sensu diviso. it holds that that cannot be. He reaches a conclusion as to what cannot be thought.But it must be granted that these are out-ofthe-way cases. .

So farewell Parmenides. But here we badly need to distlnguish the different things that may be meant by the proposition "It is the same thing that can be thought and can be. Therefore your argument commits you to the belief. as the class with no metnbers. in the sentence saying that nothing has a selfcbntradictory predicate true of it. can there be a property which nothing has. and you pretended that 1 cotrid do that without accepting the conclusion "What is not cannot be thought. That other arm of his first premise.rtlo~lis 1. o r can be the case. we can always fbrrn a negative predicate. it seems it might mean: Only what can exist can be thought of Only what can exist can be.F x and that is not credible. in fancying you could use these tecliniques of a later time to show that rny thought was fallacious? For 1 prrccive that your thinkers introduce as existent a null class. the class of things with such and such contradictory properties. i.. you didn't purport to mentlon anything? Take the first synibolic for.g. though it has to be granted that some authors. that though yo11 can speak of non-existent objects. such as Quine.he might continue .101.F x - Nec . So pay that price.g. it will follow frorn this that no properties are universal. remains tantalizing. i. whether we interpret the premise as saying: What doesn't exist can't exist or as: What isn't the case can't be the case the proposition is not credible. then.he rnight go on . can be thought of" seems refuted by the argument from the impossible pictures. [hat therr are things that aren't identical with themselves. not even srll'identity.^. use them to determine a class. in which it isn't credible. which can be thought'." But your insistence that what is not can be has landed you in srlf-contradic-tiorl after all. which can be'. are accustomed to speak of the reference of variables. Would you rather say. thought to exist (3) Only what can be the case can be thought to be the case (4) Only what can be the case can be thought of. any rnorc tlla11 you1 variable x to range over non-existellt as well as existent ob. i. if negation is admissible. the other arm of his premise is 'Only that can be. So we have Nec -(3F) (x) .jectr:' I t . i." To take first the arm Parmenides actually used. failing to produce real thoughts.that the existence ofpropcrties i s something other than their holding of' something? Only so. h)r that rc!jc. What he used was 'Only that can be thought.e. nor ran you appeal to your rejectiorl of'~lcgativcs to protest against this way of putting thr prenlisc. but you mentioned properties arid object$.e. as appears to be done iri: Concerning the non-existent it holds that that can't he.F x e. or go along with me. But if this is given up.ntionccl non-existence. If some way of characterizing what can be thought could be found. "Onlywhat can exist.e. Parmenides is deprived of his claim that we are committed to self-contradiction in existence just because we are willing to use a selfcontradictory predicate .aseci 011 the conclusion ofjust this argument. We might call this arm of the premise the 'No Mystery' arm. tllcrl . it seems. which nevertheless they know cannot be avoided as a possibility in natural languages.adiction in what exists is just what I set out to avoid. He might reply in turn: What d o you mean. as it ought to be. Or.were you not labouring under an illusion. and stop short if anywhere only at paradoxical ones. except that it might make a difference to our estimate of the first that our property-variable is admitted to range over self-contradictory properties.e.g. to put the sanir poirlt without mentioning negation. So they admit contradictory properties. (9) On any of these interpretations Parmenides' own argument is vitiated by the requirement of the argument that the second premise be taken in sensu diuiso. Will you say. Carl you allow your varial)lc F to range over non-existent as well as existent proper tit. o t there nut being any ot~~jerts of' certain kinds. Mystery and Contradiction 5 But we may reply: We didn't purport to mention anything when we wid "Concerning what is not. (1 Nec (F) ( 3 x ) F x Now since. there's a quick way of excluding mysteries. there are things that are diflkrenr li-or11 themselves. e. you cannot speak of non-existent propc:rties? Then all PI-operties exist. including the most self-contradictory ones? Self-cont~. is impossible.e. And .nul la ti on: you may not have rnt. To this we can reply: it is false that one mentions either properties o r objects when one uses the quantifiers binding property variables and object variables. which he does not in fact use." Your premise comes to this: (F)(x) .4 The Ancient Greeks Parmenides. Thus it wasn't necessary to distinguish the various things Parmenides' "be" and "think" might mean in considering his actual argument. or the non-exister~t as such. then if this proposition is true. .

So if the premise were true. True: but wouldn't it be satisfactory if we could refute There may be what can't be thought or: Something may be which can't be grasped in thought? The idea has at any rate had a strong appeal. However ttie inipossible pictures show that something can be conceived which. the 'No Mystrry' a m : Only what can be thought can be. However.. about the contradictory. . as subject for "can't be thought". can't be conceived to exist. wc shall have a second prrrnise (what is in fact Parmrnides' c. then its content is contradicted by the fact that it is a thought. can be the case.) Only what can exist car]. The diflerence hetwecn the various interpretations becomes important: Only what can be thought of can exist Only what can br thought to exist can exist ( 3 ) Only what can be thought to be the case can be thc case (4) Only what can be thought of can be the rase. . but what it says is irreducibly enigmatic. think that Parmenides did argue like that. bc thougl~t to exist. Of course if the sentence is mere abracadabra n o Only what can be thought of can be A seventh regular solid can't be thought of A seventh regular solid can't be ( 2 ) Only what can be thought to be the case can be the case The contradictory can't be thought to he the case . And be true is that: what r~iay Orily what can be ttie case can.and if we put in "wittingly" the first premise with "wittingly" inserted becomes totally doubtful. For presumably it means: Only what can be thought to be the case by someone who has in mind all the implications of what he thinks. I can't think of anything I can't think of. not "What is not". But now let's try the othcr arm.onclusion): What is not or is not tile case thought { can't 1be or be thought to be the case of' the Parrnenidean premise "What is not (or is not the yielding as ~onclusion case) can't be (or be the case)". If the thought "Something that can't be thought may be" isn't yet the thought of. without misunderstanding.e. but what eludes explanation. i. (Not: what one can't invest with the feeling of having thought it. then perhaps it's just that we can't manage the thought. one is purporting to think of something one isn't thinking of. Antl this proposition is pcshaps acceptable. it is impossible for someone to have in mind all the But sincep---L implications of anything at all. what remains enigmatic. just as it is conceiveti and understood.Parmenides.and who knows what they are? If I say I can think something. what of it? Let us try the negation of the proposition: (A) There may be what can't be thought. And in a way. it can be thought of. anything in particular. reading inattentively. we've no need to worry. logical error or contusioii. if. "Only what can exist ran be thought to exist" is false for the same reason. For in thinking (A) one is not purporting to think the unthinkable. meets the difficulty that the contradictory can be thought to be the case. Presented with the first argument. Mystery and Contradiction 6 The Ancient Greeks The unverifiable-in-principle can't be thought to be the case The unverifiable-in-principle can't be the case "Only what can be the case can be thought to be the case" is evidently false .: The contradictory can't be the case (3) Only what can bc thought to be the case can he the rase . what of it? If I say I can't. without misunderstanding. But this is wrong.ifwe aren't very restrictive on what we call thinking something. What people have done is to try some other subject in the second premise. though one might. nothing could be the case. or confusion. For then no one could have any right to produce aparticular sentence and say: this is true. (1) (2) 11' the thesis is going to be used in an argument something like Parnlrnides'. The second argument. 'was ich nicht weil3. This however is not a path anyone has taken so far as I know. as we say. any more than. does that mean I can't manage to do what I d o in the other case? Again. be thought to be the case : . macht mich nich heiR'. If the only reason for saying a seventh regular solid can't be is that it can't be thought of. logical error. ofcourse. in thinking that there is something one isn't thinking of. But what are we to make of this first premise anyway? I t appears to draw attention to the possibilities for thought . pvq. one's hackles rise. Here are some rrsulting arguments if we try this task: (I) At first sight one might think one could disprove this by arguing that if it is a thought.

of equal by seeing equal sticks and equal stones (Phaedo. (6)The mind is reminded of the type F by perceiving Fs because of their resemblance to the type F itself (Phocdo. in previous existence (Phaedrw.But he woultln't be exercised by any definite claimant to be that which can't be grasped in thought. So much is regarded as obvious. It is a sort of prejudice.540. 74a. F itself. "can be presented in a sentence which can be seen to have an unexceptionable noncot~tradictory sense". And that would be very agreeable. Someone who thought this might think "There may be the inexpressible. ( g ) The many Fs resemble. That gives one some conception of a Platonic form. 74-5). there may be loo tokens of the type letter e on a page. 951). 597c). ~ 4 9Phaedo. we might say: Since the sentence cannot be taken as expressing a clear thought . however. This is the theory of forms or types (see gnd and iooc). nqgb). was in that sense irreducibly enigmatic . 596a). F itself.510. The trouble is. but e itself is just one letter. This suggests as the sense of "can be grasped in thought".484. with which the mind has been acquainted ~ . c). nqgb-c).reek one will take any notice. A form of: whatever can he said at all can be said clearly. But suppose the sentence is not abracadabra but yet there are difficulties about claiming an unexceptionable sense for it? If that is the situation. . ( 3 ) This single type F. 1 There is a philosophic theory which is accepted as certain by the participants in Nowadays when talking about words o r letters people distinguish between 'types' and 'tokens'. or are expressly argued for in the Phocdo: (4) It is characteristically human to understand according to such single types (Phaedrus. 2 The Early Theory of Forms . (3) from other dialogues groupable with the Phaedo.therefbre it doesn't say anything. and therefore not anything true. (91. . is quite obscure and naming it as having or participalion is not meant to be informative.g." And so in that sense think "There may be what can't be thought".8 The Ancient (.is a process of being reminded by the many Fs of the type.590. use the word "type" on purpose. is different from the many Fs. or rather the arm of it that he did not use.409. cannot possibly fail to be F (Phacdo. unfamiliar as it is in this context.i. To this we may add the following. nor are any of those things to be identified with it (Republic. This affords a further proof that no particular F is to be identified with the type F: for that. each is a unitary object derived by contraction to a single thing from a multitude of perceptions by the power of reason (Phaedrus. We could perhaps become quite satistied that a sentence. or the impossibility of clearing it up would show it was really a non-thought. also somehow not F. to which the common name F is applied (Phaedo.e. 14-51.501.e. and then no mystery. Mystery would be illusion-either the thought expressing something mysterious could be clarified. IOIC). 74d-75a). (8) The relation of other Fs to the kind or type. a thought which is clearly fi-ee tiom contradiction or other conceptual disorder . 7qa-b) .e. that they h u e it (Republic. The theory is as follows: ( 1 ) We ordinarily assume a single kind or type in connection with any ~lurality to which we apply a common name F (Republic.and so we could convince ourselves we had the right to dismiss it. or are copies of the single type F (Republic. which are either supplementations to (11. It is not to be identified as those things. the discussion in the Phacdo.500. thing we call F . can we distniss the possibility that this enigmatic sense is a trutli i1 If we coi~ld prove Parmenides' principle. or ~articipate in it (Phaedo. there docsn't seem to be any ground for holding this position. Plato in the Parmcnides raised the question of how this 'participation' or 'sharing' worked does something sharing in F have the whole of F or part of it? Is participation - - This paper has not previously been published. (7) This resemblance is imperfect inasmuch as it often happens that a particular F thing i. (5) This process of contracting into one (Phaedrus) or of acquiring knowledge of F itself by seeing F sticks and F stones . 479. 476~). Phaedru~.

It is clear that in order to get out of this difficulty it would be necessary to develop something already hinted at in the Republic. then all particular Fs plus the type F must be o f a hrther distinct type . as opposed to participatirlg in. This does not mean a retreat from SelfPredication. Since he thinks that our sozils resemble the divine and the types to a certain extent. just because it has no members. the type or form F. such that the many Fs participate in both. 78d). ( I 1 ) The type F is eternal. 5970). if all resembling things are of a separately existing type. O r (as he says earlier ( I sgdg)). . But for purposes of logic the null class is indispensable. namely that the ways in which the type F is F and in which a particular F is Fare too different to allow us to put F itself and the many Fs together into a plurality. In the Phaedrw Plato locates these objects beyond the I~eavens: but this is in a passage declarerlly figurative as far as concerns the charactrr ot'ttic soul.. ( I 2 ) The type F is thus non-temporal and not to be Lbund in any place in heaven or earth. . if this helps. ( 10)The type F. Not that in this way he can avoid all contradictions. so we can't say that an ordinary bed and the bed itself make two beds. or type as 1 have called it. it rather develops the suggestion of the Republic that the bed itself alone is what i s a bed (59713. Socrates has just suggested that participation may be resemblance. that there is a form wherever there is a set of resembling objects. 132 d-e). "A place beyorid the heavens" rriight havc the same ring as "Somewhere outside space" would tor us (f'haedrus. 132-9) The question is not solved by Socrates in Parmenider. So far as I know Plato does not use this conception. It is mentioned in the Parmenides. The present-day theoretical concept most like a Platonic form. one will grasp a second type in connection with the many Fs taken together with F itself: And so on ad inznitum. It is this comforting picture that gives the concept of 'class' as used among philosophers the deceptive air of being both an impeccable logical notion and a down-to-earth sort of conception with no 'metaphysics' about it. One should realize that this is a total misconception about classes when one encounters the null class. or. a class may be thought of as an aggregate or collection or combination of objectsjust so long as 'aggregate' or 'collection' or 'combination' is understood strictly in the sense of 'class'. This will surprise only those who suppose that a class is its members. any combination of objects of any sort. for in the sense in which a class o r set simply is that lot of objects. an empty class does not exist. forms are not elsewhere introduced as different similarities . which comes to this place beyond the heavens to see the rypcs: and so the localization itself may well be figurative. . If "F" should actually be "similar" the argument will of course apply to it as to any other form: but Socrates need not have admitted attack by two separate infinite regress arguments. qua similar. saying "Isn't it very necessary that resembling things should participate in a single type?" and Socrates acquiesces. Thus it looks as if Socrates ought t o have repudiated the 'great necessity' of this principle. as Plato would say. then. and certainly in the Phaedo and elsewhere he regarded the many Fs as resemblances of the F itself. 5 9 7 ~ ) . or either (Hppublic. Some people would adtl to tl~is list of conceptions determining what Plato meant by a form. The class of people in this room at present is simply . any collection. well and good. where Fs are a plurality all called by the same name F. which he certainly does not foist on Socrates as I suggest he foists on him the necessity fbr sets of resembling objects to be of one type. In short.10 The Ancient Greeks The Early Theory of Form 11 resemblance? (Parmenides. echoing the P h e d o doctrine. 247c-c). O r again. besides the things to which we apply the same name. that a class is composed of its members as a wall is composed of bricks. and now I give a list. except in the sense that similar things share in the form 'similarity'. (9)?'here are 110 two distinct types F. and further that the aggregation or collection or combination of say seven given pairs of shoes is not to be identified with the aggregation or collection or combination of those fourteen shoes. It might seem that Plato's forms differed from classes by the Law of Extensionality which applies to classes: classes which have the same members are Set r h r o t y rmd if1 Logic (Cambridge. O r take the opening of ~ u i n e ' dSet Theory and its Logic: We can say that a class is any aggregate. If one grasps a type F in connection with the many Fs. 1963).students of modern philosophy in the British Empiricist traditiorl need warning against this approach. F itsell'or F-ness (the abstract fbrnl occurs. For there is another such infinite regress of types previously extracted by Parmenides from it seems to be in this place in the Parmenides . though not as olteri as the concrete "F itself") is F. is the concept of a class. F (Republic. it is not a very probablr foundation for the theory. This is the doctrine of separate existence which is the great rock of offence in the Platonic Theory of Forms. But I do not know of any other place where it is suggested . that there is nothing to it but those objects which are its members.that some similar things. and all particular Fs resemble the type F.and so we are in for a vicious infinite regress. must all be of one type F and other similar things of one type G and so on. similars are similar by similarity and not by anything else a t ail.p. simple and changeless (Phaedo. Indeed it is the only thing that is. Just as we should not say that there were three beds in a room if there were an ordinary bed there and alsd a picture showing two beds. Mass. A class in modern logic is thus something btsides the things of which a certain predicate holds. But wen this will be less help than hindrance unless we keep clearly in mind that the aggregating or collecting or combining here is to connote no actual displacement of the objects. His acquiescence in Parnienides' suggestion is therefore disastrous. 1. nor with that of the twenty-eight soles and uppers. That is. if F itself is F. Parmenides introduces it (Pamrenides.

Relations are indeed a difficulty for the theory of forms. However. type or lbrm. that is the same as the doctrine of 'separate existence'. which no one has been able to explicate and which ~ e o p l e take as a rock-bottom notion . In Anders Wedberg's wonderfully . For one thing Plato does not speak. but with further aspects of the theory of forms which are foreign to the theory of classes. but "either a man or a horse" would not be a designation of a distinct form. it does not matter how you characterize a class so 101% as your clmaracterizations apply to just the same things as mcml>ers o l the class. Surely. it's either a horse or a man! This attempt to say something breaks down -which has made one think it unsayable . the theory of forms is a theory of characteristically human understanding. Now Plato explairled a type as what we custolnarily assume for each plurality to which we apply the sarrlr nanie. is fairly striking. Stockholm.give a form by enumerating the things that participated in it. it is a theory of what are the true objects of knowledge. predicates and being.g. who dwelt obsessively on the theory of forms and is apt to be found having a bang at it in every context: we may say the theory of forms was to him what private ostensive definition was to Wittgenstein. and lie argued that there were never two types for such a plurality t o br of'. not thereby the same? But we must scrutinize the Law 01' Extensionality a little Inore closrly.So here again liis concept ofa type or form is noticeably similar to tllr rmodern logician's notion of a class. valuable book Plato's Philosophy oj' Mathematic. in general. He says. he said. will be ofthe slave as such. There alr indeed several diHerent class-concepts for any not just those obJertr but sornetlming besicit%s its rrier~lbers . but the meet of the class of red things and of'squares is itself a class. many class-concepts are concepts of relation to individuals. for example. 1 9 5 5 class of things that are either men or horses. it does not look as il' Plato's conceptiorm of definition was otlmcl. Again. lbrms participated in by thc same objects arc. . or the type: master. or that this is not inpart connected with the extensional conception ofclasses. Plato would not have forms red square or white man. of the type: slave. so. There is this hrther point in favour of the feeling that the extmsionaPty of class concepts makes them contrast with Platonic forms. Again. The rock of offence is that a form is something 'besides' the particular things that have it. the forms are supposed to have a pure and divine sort of existence. (Here I anticipate a doctrine of the Phocdo which will get closer discussion.) Finally. Plato would never we may surely say with safety . This sl~oe is not a mrrnber of the class one of wtlostr ~nenibers is the pair to which this shoe belongs. it either is a horse or is a m. This relation is dark because as we have seen the members do not con.r2 it is rca~narked upon. the descendants of Asclepius. a difficulty with which Parmenides presents Socrates:-the master as such. "you undmtal~d 'collectic)n' as equivalent to 'class'". This resemblance of the notion o f a class to that of a Platonic Fornm is recognized by those who have studied the matter. no analogue of the null class occurs in his ohilosophy.e. Assuming Asclepius to have existed. The class of the fourteen shoes and the class of the seven pairs arc dilkrent classes with dilltrent members. vl'a The puule about participation in a form recun about nlel~ibership class. This is a rather greater degree of what some call "rnetaphysicality" on the part of the moderns. i. Second. however. than extensional . 'Membership' of a class and 'participation' in a Flatonic kind might be called the same sotion. This is one standard way of giving a class. in just as special a way. The similarity. First. nothing like that holds of classes. of fornis in which nothing participates.i t call matter howyou take those objects. as Quine the Meno Socrates simply looks around for a definition of colour that will apply to all rolour m d in the Sophi~tPlato propounds various detinitions ofthe Sophist which do not s u ~ p l a rone ~ t another. would form a class. is something separate. members ofclasses resemble themlet alone that the class of Fs is a kind of cause of an F's being F. Still it would be wrong to think that there is no difference between Platonic forms and classes. "Plato is our friend but truth is dearer". . This leads to - ."We don't like Plato but falsehood is worse": an acknowledgement of the common rock of offence. either-a-horse-or-a-man.rtitute a class. as the theory of classes is not.12 7'he Ancient Greek3 The Early Theory of Forms 1s the same. Of course it is difficult to see why this plurality with a common name should nof involve us in the assumption of a form: but it seems clear that this connection with an individual would be a contamination from Plato's point of view. it is not professed in set-theory that sets are the only true and truly knowable objects. e. incompatible with that 'pure and divine' existence appropriate to the forms. But the dislike of Plato is probably connected not just with the rock of offence which the theory of classes avoids as little as the theory of forms. A thing isn't either-a-horseor-a-man. Echoing this there is a famous remark of Tarski's "Inimicus Plato sed magis inimica falsitas" . but it would probably be quite foreign to Plato to admit a form in connection with Asclepiads. it is not heard of that. we may say. but. this is a failure to notice that classes are only collections "if".The contrast will not impress someone used to doing things as they are done in logic nowadays. so far as I know. It was this doctrine that offended Aristotle. Similarly for the join of two classes: we have the 1 . What classes do have the sanie mernbcrs? Since a class is not the same thing as takrn collectively . that classes are collections and forms are not.but perhaps we could say it like this: A thing isn't. and not of any individual slave.. especially in the great rock of offence. it has about it a strong flavour of Greek philosophy about forms.but which surely does neetl explication if the class. Wedberg notices that the Law of Extensionality may well apply to hrrns so that that does not differentiate thrm fronl classes. 1 . And it does matter. whether there is anything in it or not. Third. just as the number o does not occur in his mathematics. however. he suggests.

the participants are very sure of this doctrine and its corollary. Turns into Miss T. In contrasti~lg fi)rrrisand classes I mentioned one point in the doctrine of Ibrms which I did not put into niy list ( 1 I-( 11): that a form F is a cause of F things being or becoming. (5) The division of one makes two.- .whether by presence of the form or communion in it but the safest thing to say about causes is: - What makes beautiful is the beautiful What makes bigger is bigness What makes smaller is smallness What makes ten more than eight is multitude What makes two when one is added to one is the Dyad . if something other than the form F is or comes to be F. and it shows Plato up as having a very screwy idea of reasoning.e. addition and division might be efficient causes. the most powerful explanatory assumption that he can make is that. as the best possible assumption one can make. 'By a head' and 'by two' in "taller by a head" and "greater by two" do not have an analogue that I know of in any causal concept. then it is beautiful through the beautiful itself or because it participates in that.b. e one can put a. can neither know The conclusion is not satisfactory to Socrates but he nor govern our alfsi~s. telling how as a young man hr was occupied with natural-srientific questions about the causes of things. This "strongest available hypothesis" is that. This quite valid procedure Plato had apparently decided was what was wrong with the arguments of some of the Sophists. eating. and he calls explanations like 'food'. (3) Ten exceeds eight by two. - - - - . and simply answer like that if asked about causes.i. one must make no answer till one has tested it for consistency. Using his assumption. If however one is asked to give an account of that hypothesis itself one must assume higher hypotheses until one comes to something adequate. I left this out because it is not part ofthe doctrine ol. an equivalent formulation. stand certain see~nirlgly He gives as examples that lie did not understand how it is that: Flesh accrues to Llesh out of food through eating and default of an explanation in terms of good and mind. ( 4 ) The addition of'onc to one makes two. It is intmdacrd in a rather special way. Socrates next says that if one is to follow him one must hang on to that safe point in the hypothesis. F itself. if a. However. it is or comes to be F by the form F.the existence of the forms and the hypothesis of the forms. he heard that the philosopher Anaxagoras held that mind was the cause ofeverything. and the gods. but not that this was a cause of their being F. in defgult of supplying this kind of explanation. We come now to the most obscure passage in the Phaedo. Illat whatever Miss T. nor is i t a11 a~n~lilication that can readily be supplied. division. c and c.d. without argument but with a good deal of firr: Socrates says that in raisinga certain doubt about the immortality of the him in a general enquiry about generation and corsoul Cebes is involvir~g rt~ptiori.c e. eats.the Greek word is the verb from which 'hypothesis' comes . the pre-existence of the soul. or participated in.b c and a. and he thought that sounded splendid and would mean that Anaxagoras explained evetything in terms of what was best. But if anyone hangs on to the hypothesis itself. For there is nothing wrong with mixing up prernisses and conclusions derived from them so as to derive fresh conclusions. Consider. digestion. As queer as can be. Anaxagoras' book did nothing of the sort. This last remark is the one thing that is clear about this difficult passage. or. He ther~ gives 11sa bit ofautobiography. ~ov where d = a: we have a. This remained Socrates' desire. Socrates does not intend this to exclude other types of causality. ( 2 ) One man is taller than another by a head. which everybody grants him. That said that particular Fs had.e. He still wants an explanation in terms of 'the best'.The Early Theory $Forms 14 The Ancient Greeks 15 the conclusiorl that the world of forms is quite dissociated fi-om our world. Now however Socrates speaks of assuming the strongest account available to him . drinking. such as how flesh grows: ~t's a very queer thing.Ii. e. So we can put a. just too difficult for him-but he does not imply that they are false. Two itself Whatever is to be one will be one by the Monad . as Mrs Kneale put it in the While in this perplexity.b e. and the sort of cause which Socrates offers as the best he can do for the time being is what Aristotle called theformal cause. the addition of one to one. by participation in the form F.d. and in relation to this hypothesis the word begins to get a more technical sense. The dialogue has regularly spoken of assuming . This is certainly an addition to the theory as I have so far sketched it. but.i.b. that the forms exist. he must he says for the present adopt another.rnls that is accepted as obviously right. These people constructed arguments which. One itself. (1) With hindsight we can see here the first adunibrations of the various kinds of cause described by Aristotle. food and the two ones in two are material causes. cannot solve the problerrl. Socrates says he doesn't know how this works . This however does not imply any uncertainty. if something other than the beautiful itself is beautiful. Above all one must not confuse principles and their consequences like the disputatious Sophists. in which an hypothesis needs to be based on something else before it can be finally accepted. After a lot of thought he came to the conclusion that he did not at all underobvious statements about how things come about. Take a particular case of this o v v ~ c r ~ Bccupwa. The explanation in terms of the best that Socrates hankers after is something like an explanation in terms of final causality. who have knowledge which is knowledge itself.

Combinability and uncombinability are readily explained. in a basket. or more precisely that one being added to one. p. Then what is the hypothesis as opposed to this 'safe thing ahjut it' . These would be 'the monad.e. Here we have Plato making a very general boss shot at characterizing sophists' fallacit:~. but it seems to cast a great deal of light on these ifwe see in them the germ of what ~ristotle gives an account of. each of yhich had many instances and which were 'combinable' and composed of exactly similar units. That is the result Socrates arrived at when he puzzled about how the addition ol'orie to one should bring it about that there are two.other .for only that has so far been introduced expressly as an hypothesis? I can see no answer but that the hypothesis is: that lhtre is a result two. I appeal to two sources of help. o The null class 1 The class whose only member is the null class 2 The class whose only members are: the null class and the class whose only member is the null class o The null class 1 The class whose only member is o 3 The class whose only members are o and I . if in a more modern vein we explain the number two as the class of couples. that they are speaking of things which can only be grasped in thought.Kneale. of course! But it is one thing to see that an argument rnust be wrong. weren't they? No rleed to add them or bring them together. The Dcvclopmcnt oflogic (Oxford. or far as believing the argument goes. Of course you can add one egg to another egg in a basket and then ~v!~rre you had one egg in a basket.. not one qanything. The rest of the passage is so obscure that many translators render the tirst occurrence of one and the same verb &xeoflat "hang on to" and the next mean the salrty of'the hypothesis. the safe hypothesis itself: However. the reading "what is safe about the hypothesis" appears to me a very possible way of taking the Creek as Greek. One is the Republic which has a passage so closely similar to the Phaedo passage in certain points that it seems certain they relate to the same matter. The other is some very pertinent information given by Aristotle in Melaphysicr XI11 about Plato's teachings which may of course be rather developed in comparison with what we have in the Phacdo and the Republic.' "would not deceive a child" . Nobody makes such shifis to some quite contrary sense a verb may be capable any rate you . But if you just say one and one are two and mean the number one. 536: "If someone were to ask them: what sort of numbers are you talking about. 'mathematical' numbers. the dyad' of which Plato speaks in the Phacdo. Does "one" mean the same both times? If so why not say the moon and the moon make two? Does "one" mean something different in its two occurrences? If w. or ifanything is to be one. then isn't it nonsense? O r if you mean one and one of anything you had. it comes about that there are two. containing no parts? how would they answer? . First. 1s. Plato says mathematicians are concerned with such numbers as these at Republic. & M. etc. in the next line. "attack" -which is intolerably harsh. they were already two before." The other sort were the numbers which were forms: there was only one of each of these and they were uncombinable and the units of which any one was composed were all unlike one another and uncornbinable. And this is a genuine difbcultv. another to say accurately what was wrong with it. The puzzle is that evidently the person who is hanging on to the hypothesis itself is doing something different from the person who is hanging on to the safe point about it. every single one equal to every one. Now you cannot combine the members ofthese classes in the following sense. what? And how is the opposite process of separation also supposed to make two? Readers of Frege's Foundations o j Arithmetic will recognia the source of my duplication of Socrates' difficulty. i. which contain rme as you conceive it. and the only way of making sense of the passage. It rnust mean "hang on to" t~oth times. you get: . the text does not say this doesn't happen. not differing in the least. In further defence of my interpretation.C. For example. by the approach of one . But if you follow someone like von Neumann and define o as the null class and each number as the class whose members are its predecessors. And that is like combining (A and 8) with (A and B and C ). 196s). but just in the particular instance that occurs in this paragraph: if there is a resull two when one is added to one. If you try thinking of the addition of two and three as a matter of combining the members of the classes two and three you will be trying to combine: 2 The null class The class whose only member is the null class with 3 The null class The class whose only member is the null class The class whose only members are the null class and the class whose only member is the null class - W.egg you have two p g ~ . in every case this is through the form. then any member of any couple will combine with any other member of any couple to make a couple and a set which is thejoin of one couple and another non-overlapping couple will be a set of four. but that the ollly cause one will be confident in giving will not be addition. or divided. Most translators rake "the safe thing about the hypothesis" -literally "the safe of the hypothesis" .16 The Ancient Greeks The Early Theory of F o m 17 Development o f Logic. This safe thing was of course the thing mentioned as "safe" before. Aristotle says that Plato distinguished between two sorts of numbers. unless in a very pointed manner.

Mathematicals are not visible objects but not forms either. Since he thinks. will 'postulate another hypothesis' from among higher ones. Putting the two passages together we get as Plato's doctrine: (1) Ortlinary mathematics postulates numbers (e. loidg).e. the result will only be ABC still. the mathematician. It seems safe to equate this with the account in the Republic: he will make the hypotheses not into principles but really into hypotheses. is like the seeing of shadows and images in relation to the true objects of sense (534a5). etc.g. whichever seems best. angles and so on (Phaedo and Republic). if anyone holds fast to it. let us say . for there is only one form of the circle. For exarriple. however. he will lay hold of that. etc. 5 1ob-I I c we get a strong impression that they relate to tlie same matters. he does good by drawing the soul away from preoccupation with the sensible world. The mathematician "dreams of reality" but does not really get a vision of it (533b8). but one should not pile up too many discrepancies. however we learn in the next book of the Republic can only really be seen by pure intelligence engaged in dialectic (Hepubhc. but only that the mathematician's reasonings arefor the sake of these and that he is trying to see forms. even if they would not. But it would in a way be wrong to say that Plato believed in the real existence of mathematicals. I say "if they exist".. and if he did think mathematicians grasped the forms. (5)So the philosopher who follows Socrates when asked to give an account of the hypothesis. but only thinks it certain that if these postulated nurnbers. but in relation to the true objects of knowledge his thinking. says indignantly that he is not talking about visible bodies that have numbers (Republic. the monad and the dyad of the Phaedo.. but firstly the al~cients wc.533~-(1). I f'we read Phaedo. he is talking about numbers. by seeing if they are consistent notions. that is not the only question at issue. i. this monad is not the actual . (3)The tbrms are not the postulated numbers. Perhaps we may say it postulates that two is generated (Phaedo). The development of the propositions of arithmetic and the theorenis of geornetry goes forwards from these postulates or hypotheses. namely that something is one. wtlich. to reach forms themselves. e. subject of the mathematical reasoning.5 1 la) surely contains an allusion to the theorem that the square on the diagonal of a square is twice that square in area.At any rate it postulates being one and becoming two (Phaedoo)and geometrical shapes (Republic). o r that there is a result 2 ( 2 comes about o r something becomes two). etc. and also speaks of testing the hypothesis. Again. Plato might not have noticed this. I suggest. Some have thought. Since. etc. angles. not using the sensibly perceived at all. it postulates odd and that the dyad is the cause of there comirig to be two. indeed Plato might not have noticed this.There are thus several ones in any of his numbers. And in the passage about 'the square as such and the diagonal as such' (5iod7) he speaks rather ambiguously: he does not actually say that these are what the mathematician is reasoning about. angles. figures.18 The Ancient Greeks The Eatly Theory OfForms 19 won't get five that way! W e indeed shouldn't much object to speaking of corribiriing AB arid ABC. by seeing if it is consistent: thus it appears that he does not regard the Iiypothesis itself as certain. and Plato says that the mathematician's reasonings are "for the sake of the square as such and the diagonal as such". they are subjects of mathematical reasoning. (4) The philosopher who follows Socrates will react to the postulation of numbers.e. Similarly the square as such is only one. figures. Against the view that Plato introduced mathematicals it is objected that nothing like the foregoing point is made in the Republic. but arc the causes of these (Phaedo)if these exist. There seem to be strong hints that there are frightful difficulties about the postulated numbers and figures. who is trying to glimpse the square as such (Republic. figures. and secondly.. Plato says. and he says that the safe thing about the hypothesis. ( 2 ) I t is csserltial to mathematics to use drawillgs and physical objects. He only believed that mathematicians did. the forms are their causes. he contradicted himself flatly in the next book of the Republic. and others have denied. but the reference to 'the square and the diagonal' (Republic.but a mathematician has to speak of intersecting circles. one and two). Since there can be only one monad in the sense of the Phaedo. exist. at any rate rornbirling here is not that join of non-overlapping sets which yields a set of the appropriate number. but using forms themselves. because they are only the hypotheses that Plato is speaking of. but it is corroborated by (3)above. yet the circles that he is speaking of are not the marks on a particular piece of paper or the indentations on a particular patch of sand. up to where he reaches what is unhypothetical. till he comes to 'something sufficient' (Phaedo. i. Ordinary mathematical reasonings to propositions of arithmetic and theorerns of geometry seem to be admitted as consistent (53. 5 I l a i ) and so presumably is also trying to glimpse. i o I c-e and Republic. o r two.g. 5n5d8): no.5~5) but. This theorem mentions two squares. by seeing it'he can derive a contradiction from these hypotheses (Phaedo). in which each one is exactly the same as any other one (526an). that Plato introduced 'mathematicals' into the Republic. Plato scerns to leave it an open question whether such a contradiction can bc derived or not. what he actually grasps in thought is different.uld probal~lyhave objected. through forms themselves. Nevertheless as we have seen he expressly speaks of several identical ones which mathematicians claim are in the numbers they are talking about. and fasten on to the things that are made fast by it. but only by means of dialectic. the principle of everything. stepping stones and jumping-off points.and herice presumably the only safe thing about it. he later reveals that the mathematician is trying to see what cannot be seen his way. and so he will descend to the end. these are mere irriages tbr what the mathematician is after: he uses them to try and see the forms.

and good is From The Monist 50. Thus the postulated mathematical objects are themselves mere shadows and images of' the forms . ( 1) (9) Plato's writing touching the forms may be divided into the naive and the sophisticated or reflective. indeed. and the numberless parts of the other. the form of the good. that belong to it. Here is an analogy for the new theory: a laminated sheet divided into two halves which can be folded one upon the other so as to coincide in their exterior boundaries.except. or being is the forms. The bottom three laminations have no holes in them. intelligible shadows of the real seen on the upward path from to reveal things of sense to the fonns. Any pattern is stamped right through all three layers. o r being dflerentfrom. And naturally mathematicians remain in their half-way house. of the one. We now suppose the diptych shut. of the same. good itself is one and simple and that the forms are being. L 3 The New Theory of Forms . he does not come to them. whole and part in the later dialogues The conception of certain forms as 'parcelled out' (3) The problem of negation. The whole represents the totality of beings. These laminations are common to both sides. But of course this only means that Plato disbelieved in mathernaticals as Parrnenides disbelieved in the various and changeable world. and beautijul. dialogues. "We say that p t itself. 1 believe. In knowledge. The Republic reveals that this unhypothetical thing is the h r m of the good . Each such pattern has a leaf of the right side which corresponds to it by having its shape cut out. for example he applies his theory to what they introduce. which otherwise differ. Wheri the hypotheses are done away by dialectic proceeding fiom that sufficient unhypothetical tint principle. in the Sophist. They remain only the hypotheses of people who d o mathematics and d o not do dialectic. But there are very many further leaves on this side. as many as there are distinct patterns stamped on the left side.g. The bottom layer corresponds to being. Or perhaps I might rather say that he constructed a new under- 1 want to suggest that Plato arrived at a revised theory of forms in the later pinning for the theory. So no wonder we get the Academy developing mathematicals as Aristotle describes. . we have the unitary forms alone and through these discern that the mathematician in his reasonings was producing a mere conjectural picture of what we now possess. They are an unreal halfway house. In offering this analogy I am treating as key themes: The study of one. so it is intelligible that belief in the mathematicals should be ascribed to Plato. When he becomes sophisticated he reflects immediately on his own formulations. there. The objects of mathenlatics and geometry which are not yet forms aren't obtrusive er~ough to qualify as a whole intermediary world. so that the shape of the pattern is there a hole instead of a forms-tion of the material. In the left half of the diptych patterns are stamped. but it takes mathematical enq~liry them. the next to unity.they do not get established by dialectic. although the pure form doctrine is the only theory of the really real. But patterns may divide into parts which are also patterns. Must not there be being itself? This question would perhaps have greatly startled and perplexed him when he wrote the Phaedo. This can be discerned. His attemnt to derive numbers as forrns in the Parmenides is a mere exercise. For every pattern the perpendicular geometrical projection through the whole thickness gives us the part ofbeing. and the philosopher advances from it in the establishment of real knowledge. g (1966).Po The Ancient Greeks arid so conclude with forms. the next to identity.thus claiming to sketch the fulfilment of Socrates' desire in the Phaedo. and a 'two worlds' dichotomy is constantly taught in the dialogues. They are what mathematicians use drawings and other physiral objects as images of (51od7) but are not the same as the b r m s the mathematician is pursuing his reasonings for the sake of. 'Naively' he says that e. The right half of the diptych represents the other. Let us suppose this diptych to be lying open. according to Plato. and so that this hole coincides like a mirror image in position and shape with the stamped pattern on the left side. taken together with certain parts of the dialectic of the Pannenides which use the same language as the Sophist. The patterns correspond to the forms of the early theory. sketchy and unconvincing.

"The first is whether we ought to accept some such monads as truly existent. all these are fires I~ecausethey have the characteristics of' lire. which may be fbllowed by a predicate-expression. I t is what fire is. 1965). (This separateness. 304). See 'Plato and the Copula'. Next how these . Greek itself'worked with Plato to produce the early theory of fc)rnls.5d). I aln talking about them all. q-itself. Sophist. The transition from "something" to "one thing" is easily made to seem compulsory by Greek.g. Firc.The New Theory o f Forms 22 The Ancient Creeks 23 sornething . (See Parntenide. on being. In the Philebur we find this sort of sophisticated reflection on calling the forms single. Studies. one. But I cannot find that the word for it is used until Plato has begun to write critically about the theory of forms. unitary changeless natures. Being. man. ed R. . The argument seems to run somewhat as follows: The many men. same and other 'run through' all things. We can infer from Parmenides. could gv naturally into sentences like the last four of thc: Ian paragraph. we might say. and substantial kinds generally (of which fire is one for an ancient). In Greek the form "the p itself" --functioning like "fire itsell'" .ill has adequatelyrefuted Cornford's view that the intercommunionof the forms is symmetrical. E. and had realized that if 'thejust itself' was what he wanted it to be in speaking of it .therefore something with a numerical identity which one has before one's mind' may seem inescapable (cf.5.Greek naturally fbrms "the many q's" for any . is clearly the origin of the Aristotelian Scholastics' doctrine of 'transceodentals' which 'run across' all the categories. but appears many because of the communion of forms with bodies. o r says that the beautiful itselfexists or is something." ((i5d)Forms were being. Was this intercommunion not mentioned in the Republic (476c)?And that is a 'naive' passage: each of the forms is itself one. he f0rmulates a theory of the intercomnlunion of certain hrrns. arise ifwe try to postulate one man. to give all the rest in a single word. or1 sarne and other.each of them being always the same and not susceptible of generation and corruption nevertheless are unshakeably this one. One that docs is "tire". Wittgenstein meeting the accusation that he makes a sensation into 'a nothing'. the Greek for "what is". and. Allen. arld in turn on other itself. we pass over smoothly into lormulations that are explicitly natural for evrrv noun or adiective 9 . with actions. we have not yet got the theory offorms. Sludicj in Plato's Metaphysics. on one. are not typical exatnplcs of Platonic fvrnls. at Philosophical Investigations. about size for example. in that it is merely what being (p was not the 'simple' form of the Phaedo. 'Being and Meaning in the Sophist'. Aristotle's bugbear about the forms. the beautiful and the good (for these never change). oxen.) The doctrine of intercommunion of forms might not seem to belong especially to Plato's 'sophisticated' thought. one ox. It concerns the second of the three serious problems which. about their being \*. same and other as 'running tht. Hence we get the full-blown doctrine of a whole world of these simple. benefit of this trick of Greek the transition 'not nothing . their general designat ion was "that which itself is" (7. Adjectival predicates of plum1 sut). in R.. ~ q q b165e. beauties. This is supposed to be a generally admissible starting-point. In the Sophist the interest is rather in forms as participating in othews It is as if Plato had caught himself saying that thejust itselfis and is one.jects without the aid of a word " t h i s ~ s ) " .J noun or ad. q-itselfwould be 'simple'. That fire is itself one thing. The contrast is this: in the Republic the intercommunion of forms is being looked at only from the point of view of the form that is participated in by others. to which fires have an obscure say this would be to set it up as a Platonic form. something. lgoc. one good. which get lit and put out. In the Sophist Plato reflects on being as something other than what is said t o bc. The passage ( qa-b) is perforce obscure. which ibr Plato is a constant feature offorms..In Greek a term functioning as a predicate is commonly without an article. The alternative to "something" is "nothing" . 144. Reflecting o n the whole.g. But even without Greek. as I shall argue. . . Of course. E. one beautiful.jects go into the plural. This doctrine o l . There he would have said that the just itself simply was the being expressed by "just" -but that is naivetC in using the word "being".or~gh'one another and everything else. not becoming. health.the being of a thing is named after the expression "what is" (g2d). he regards then^ all as having parts. and goods are subject to generation and corruption. and with one another. embodying the "worthwhile hypothesis'' of the Phaedo (qsd). and changeless in that being-(p is being. . "big".) The English reader must imagine that adjectives like "equal". that one should be atldetl to the list. strength.hatever each ol. It cannot be clenied that tire itsrlfis sotnething. this is one reason why they are not e. ox. self-predicatiotl.jective p. There are many fires. always the same. 237d-e). And finally. is certainly indicated in the Phaedo. When Plato speaks of 'good itself'. The thought is hard to discern. so as to appear many. "jua". So we have a slide from the innocuous-seeming "q itself is something" to a doctrine of the separate existence of this unitary nature.each1 and Moravcsik2in preferring to avoid the abstract nouns sometinles used by Plato and far more otien used by his translators to renrler his expressions for Forms. fire itself is 'The Third Man Again'. they can rlso always form sub. The reason is that with such nouns. Acfa Philosophica Fennicn Fasc r d . slid. Any thesis about being offered at that stage would be a blanket thesis about thr forrns. Allen (London. (1969). Socrates says. From it indeed. Few English words function gramnlatically so as to produce with a fair degree (d riaturalness the kind ofeffkct that Plato can produce with most Greek tcrnls. Indeed they are not any of the things that are predicated of Mr ~ck. lipon ihr wholr 1 lullor (." The words "this one" are feminine and so must be understood to go with "monad". "not one" (Parmenides. is equally Greek for "wt~atexists". ofien seems too absurd.tlleln is.

One. Why then does Plato in the Philebus suggest there is a clilficult problerri about how they are rnoriads? When he wrote the Republic he woul(l presumably have thought simply that the form one appeared marly bv being participated in by bodies.24 The Ancient Greeks them. 4 Editors have emended the text at several points without grounds in the MSS. The first is whether thew monads n~ust bejudged to exist. But as a proposition about one itselfit is insane and the argument for it is intolerable. and here. Thus we could hardly say anything about anything. The later dialogues conviction that this could not be true.this belongs only to the diJimentfiommther. If we could not ascribe anything to any form except itself. Similarly about same: he is reproachf'ul about saying that something is the sanie (as one thing) and not the same ias another). It is used.because they could not follow the thou~ht. but if the name is the same as the one. We might rather maliciously truncate the argument by remarking that we can still say. which is the thing that i s p. arld other forlris. in the Philebus. but with being. and it is easy to see he was thus naive: he is here wide open to the typical criticism of the sophistiated Plato. The dialectic which follows doubtless shows. is differentiated from anything else. To reject that insane principle is to accept the intercommunion or interweaving of forms. in that nothing. Is the name other than the one? If so. No wonder "much zeal for division tnlrns into perplexity" (the literal rendering of I 5a7 as the MSS have i t ) : b r ifwe diuidc the one fiorrr the man. any ot theni. as soon as we lrave several p's. we are discussing the postulatio~~ one . is easy to discern: it is that of'relusing to admit that pitself is ever also describable as riot q ~ Parmenides . Parmenides himself said much more than "being is" such as "all else is mere names". and hence the connection. as a b r m at all. that Plato believed both theories. i. In the arme en ides Plato seizes on one. himself had tliought this in wtiat way about being.and can only ever say "Being is. when the short grow tall. not by itself: henceonc in itselfwill not be different. multiplicity or changt* in being. what must in that case be permissible substitutions in "the name is the m e of the one" yield the nonsense "the one is the one of the name" (sqqd). and discerns three quite distinct problems. that it is hopeless to ' argue that whatever p may be. to be really based on the same mistake.e. but how are irnperisliable illonads just thal? The New Theory O J Form 0 5 What then is the puzzle about many tor~ns' each being o ~ i In e Reptiblic. we know thcre is further the Ibrrn. it had to follow that another one appeared. ought we to say that those rnonads really exist. much that I have no grasp of. ox. . qucr single. there is more than the one. thus the connection with him seems at first sight rather associative and superficial. which is wholly and thoroughly one. in relation to one. thougli superficiallyunlike. hence no puzzle about tlie contrary things that may bc predicated ofthcm at different tirnes. there is even something right about it. and according to Phto the possibility of propositions and reason depends upon this. being belongs only to g itself among forms. actions. In t11c Republic itself. My interpretation sticks closely t o the text. For if there are so much as names. because it does not belong to one qua one to be different from anything. otherwise one could put "one one" for "one isH. be different from anything. each a tno~iad. We must not suppose that Plato had thought this the right way to argue before the Parmenides. But two things emerge. suggest an ever~tual As regards the Sophist it is fair to suppose that there must be a lot ol'cunnection between the Eleatic visitor's criticism of the Friends of' Fornls and his criticism 01' Parmenides. there is something that isn't being itself. For the dialectic of the P a m i d e s is a contribution to the question what can be said about a form. when arguing for there being only onc ol'any torm.n~that 01' both shared in. called the111single. The text shows that there are supposed to be three 'serious questions'. we might straightway conclude that we could not say that p is p. which it leaves intact. Wt.for only being is. for example. in the first hypothesis ( I ggc):one will not." Plato reaches the conclusion that we cannot say that one is or is one in a more roundabout way and does not point out this truly Parmenidean conclusion. the light heavy. the beaut$~/ancithegood. ally one thing. Parmenides had not been concerned with any such notion as that of the one itself.Thus one will not be different by being one. The Pamenides of the dialogue pmducn a comparabk. Perishing r~iultitudes are not. the second of which it does indeed take pllilosophical acuteness to notice. This could be shown by the sort of argument used against Pamenides in the Sophist.In comparimn with the historical Parmenides this could be called sophisticated. qua one. and hence that q itself cannot be p. that orie thing whose riature they have. So ~nultiplicity of monads should be enough to prove that the lorrns are not one itsel/. ox. which was tlie t'ol. the parts of the dialectic where this is assumed as a principle are an illuminating Gedatutcncxperimed. The results of the first hypothesis are indeed explicitly granted to be generally unacceptable. when the derniurge was supposed to have niadc two pattern beds. and how can they be rnonads?' In the Hepuhlic Plato wrote as if the corrimunion of'fbrms in forms were thr same as the communion of' anythir~gelse in forms. Bur those others existirig imperishably would have to be thisone thing that we calleti thcni. and the third is how tiley are related to the infinitude of becoming. But it is pretty obvious and is one more pointer to the genuine connection between Parmenidean thought and consideration of the theory of forms. and one wonders why he is the man who is made to indicate that the forms will be all right if backed up by a proper dialectic. - . This If however we turn our attention to the forms that do the participating (as opposed to the forni that is participated in) we are struck by the fact that the unity of' each is cssetltial to its being characterizat)lc. or is meant to show. i. teaching that one must not include non-existe~~ts is. not on being for his exhibition of 'dialectic'. though less witty argument: is and om must differ. or say that what is. is not. Plato dicl riot tilentiori generation and corruption as proof' that q's were not tp itse& multiplicity was enough. and so on. He had inferred that thcre can be no variety. beauty or good. Imterpreten generally fail to distinguish the second problem from one of the two others. This is a conclusion we could more readily accept than the method of proof.e.

" The argument goes as follows: There is nothing against what has parts being subject to m as affecting all the parts. one and many squares. not assuming that any breach has been made in the plausible seeming principle: "The truly one must be without parts. or class and member. . Characteristically. have been bedevilled hy a confusion about the question whether something was one. a as to be one and a whole. and so is comparable to the singleness of each of the many p's in relation to the property of being p Indeed. Let us postulate an answer of the first kind. all of which are single numbers. in "separated part" in such a way as to turn it into just such a countable predicate p: the question "How many separated parts has this dismembered watch?" is usually answerable without asking what is to count as a single separated part.g. but the whole itself does exist.though our purpose is to elucidate Plato . But what is subject to one in this way cannot itself be one &el( If we reason right. one and being. At 14zd-c and 144r Plato's Parmenides comes to apparently contradictory conclusions about the relations of one. the many i s d o not in general compose a whole of any sort. so each part in turn is after all not one but two. each single. or 6 bits of apple. each. any difference between Plato's b r m s and the classes of modern set theory is immaterial. each of the numbers being one. and of Christiari theology. is being a $ i d by m. but the unity in question yields no whole. for: ( 1I if being somehow one is an affection of being. one and many positive integers. and being is divided up into as many numbers as there are. is a quite different one from that of' form and particular instance. properly speaking it is an instance of that. namely "of one p" where "p" is a countable predicate. which is found in them because they are parts of one p. plat0 is beating "one" rather as if it were such a predicate 'p' and noting what results we get if we consider his 'one that is' as divided into parts. presented by the open selitence "x is p". we can say. and thus being a totality and a whole and in this way one. In the first argument. Frege objected to the question "How many?" or "How many things?" If asked "How many are there?" we need to know "How many whats?" and the answer must be a countable predicate "9". being is clearly not the m e as av md all will be more than one. being si~nply make a whole. whole and parts. we need to know 'part of what?' Here we should notice two possible styles of answer. not merely by showing thr necessity of' the interweaving of forms. parts both are and are not unit parts. so each of these parts b one. In the second: if'a par1 exists it must be one part. But the relation of whole and part. is not a case of one being found 'simultaneously in several places'. part and whole. This unity of' each of'the a case of "one found at once in many places". bounded and infinite. In the first passage the one bring is a whole of'parts."Is God one?" is ambiguous between "Is there only one God?" and "Has God internal unity?" We have been greatly helped. the one in it has being arid the being is one. The only one thing is the given whole of all these parts from which he started. Combining the two sets ol'conclusions. let alone a whole 9. We niay speak indifherently of the form. in sorting this matter out.26 The Ancient Greeks The New Theory $Forms 37 ~fhirsParmenides is preparing the way fbr the Sophist. In the second. I t is only the whole that is one. I'here are at least three antinomies in these passages taken jointly: the one that exists both is and is not a whole. part and whole. we have to say that the truly one is without any parts at dl. thus we may justi5 his saying that such a part is not one. I t looks as if "one" was supposed to apply only one way at a tirile. the infinite plurality of numbers can be generated by considering one as t~eirig. the predicate ' P ' has to be 'countable'. and let us consider one p divided into several parts. any part is only part of a whole. If the noun after "one" o r "many" is "part". parts both do and d o not constitute a single whole. that is to say. and so 011 act ir$nilrrm. and to them severally when each was a distinct riurnber. differentiated by English (though not Greek) idiom. Thus if we speak of one p among many p's. but Plato is going to use the concepts of'part and wtiole to explain the iritercommunion of' fvrms. or of the class. then being will be less than what is. But the singleness of each single part is of course a case of one being found 'simultaneously in many places'. being transitive. one itself'is infinitely divided. he cannot concede wholeriess to the set. When considering parts that are parts o f a whole. when considering an infinite set ol'parts of t)eirig. by the work of Frege. for this move to be useful. because we have modified the term "pan". it starts from smtch. but makes no attempt to reconcile these opinions by showing how "one" could apply at orice to the parts collectively as parts of a whole. but also by his considerations about one. It is clear that the unity of the parts taken together. or are we to deny totality of bring? This is difficult. t. We come now to the discussion in the Sophist that touches on this theme. Plato's Parmenides cannot concede unity to them severally because they are irifinitely subdivided. Note that. it must be determinate what is one and what are two quite distinct p's. 6 pieces of an apple. he has the result that the one that is is one and many. The question is. Not ~rierely is i t ol'intcrest iri itself. is one separated part.I suppose each to be quite separated off from the others. he has a principle of generating no end of parts. sincc we should not speak of the class of V'S as a whole composed of p's as parts. any more than of p's being found simultaneously in many places.tien. This is extremely important. Since. will never parts. Certain sketches of' philosophy.let us attach "one" and "many" to concept-words and speak of one p and many p's: one and marly men. We must observe that the question "How many parts?" is just as objectionable as "How many things?" unless it is clear in context how parts are meanyo be counted. fbr present purposes. For exan~ple. We may be told "of a p" o r "of p": e. Whereas ( 8 )if being at any rate is not affcrrcd by m so as to be a whole. his principle of division does not in any way characterize the parts so as to give us a countable predicate. Following him rather than Plato tor the moment .

whole. then thevery same holds ofbeing. for one and many (if these can be called contraries). black is of white. but that the other of Jying holds of him. Rather. motion and difference.not meaning that he is doing something else. being will not be. in the r n s e ofnN. One's first thought in reading this passage is t h a ~ it trades sophistically on the ambiguity between the quntiot.. But ( 3 )if the whole has no existence at all. that we have arguments suggesting that g can be predicated only of g. and implicitly by Quine's treatment uf Inass terms as proper names of single icattered b b j e ~ t s . And so. Word and Objcct (Cambridge.and they are meant to be dl the possibilities . of what is said. "Theaetetus is other than flying" . differences between the working of such a word as "being" and such a word as. We should have to determine whether a totality. we may infer. "Not flying" then.ence liwm the non-existence of the whole to the non-cxistence of being at ( 3 ) . Nevertheless it is not wholly abandoned: it is five times insisted upon for rest and motion. 98. . For whatever cornes to he comes to be as a whole (945a-d). itself.e. indeed. g-itself cannot be in any way describable by an opposite of g. so if there is not such a totality. and that non-being has previously been explained as the other of being (ogga8). we may suppose. 1966). Further. But we must not forget that falsehood is to be explained as the mixture with saying of non-being (960b-c). motion's not being the same as such and such is explained as motion's participating in the other in respect of such and such. rest. So the principle is infringed for same and dgerent. whether there is such a thing as the whole or not. is truly predicable. Whereas. such hopeless consequences that the ground is well prepared for abandoning it. There is. otherwise there could be no ink~. ascribed to Theaetetus by the false proposition "Theaetetus is flying" (963b). because being is ascribed to them and neither can be ascribed to the other. there arr not any. and once more all becomes more than one. But inthe dialectical exercises of the parmenides the interdict yields .e. experimental.This admixture of 'the other of being' cannot consist in nonidentity of the form saying with the form being. We are not meant to be left with these results. ifbeing is not a whole. This may escape one's attention because participation in the other of g by a form is often non-identity with g. sole)being and But seems directed first at proving that there must be more than one being. That could be said already to be done (if one is very keen on such a formulation) by his familiar distinction between participating in g and being g itself. the same by participation in the same with regard to itself. Theaetetus' flying is 'being which is other than being' (s63b1I). For Plato has told us that "not big" is not an expression for the small or equal and that "not beautiful" is the name of the other ofthe beauh/ul. it will not ever come to be either. nothing to object to in this. But that beings do not form a totalitv is without implying that ~iothing itselfa very obscure thought: one might say that the beirlgs that there are must be all the beings that there are. It is clear that we can also put: "Theaetetus is not flying" i. wc must note that the issue of all the possibilities considered . It is much as if in modern philosophy one opened discussion with a set of propositions and reasonings about meanillg which were evidently unacceptable. "Are h i n g s wholes?" and "Is there thr such a thing as the whole of being?" But we are at the s~arting-post. ~ The whole passage is concerned with Parmenides' one (i. therefore the question whether 'the whole' exists may be answered "No" exists. p.Thus motion is the same and not the same. "Participation in the other of p" is a formula used not merely for nonidentity with p. is different from a whole: a question raised explicitly in the Thuaetetus.. . The question is not one as to whether a certain abstract property exists. . as. . if "like and unlike. Further observations on the passage: "the whole itself" is a ti~uchpreferable rendering to "wholeness". These considerations show that we have got to construct some sort of account of how one. i. is the name of the other ofjying. and other forms presenting comparable dilticulties are related among themselves and to other ones. We have observed that the common feature of the early theory of forms and the philosophy of the historical Parmenides is the interdict on any sort of contradictory predication: whatever g may be. "apple" have still to be characterized. but whose unacceptability showed that our apparatus for handling the relevant questions needed to be amplified. It is only in such more extreme. In that part of the Parmenides which deals with the forms we have Socrates saying that he would be utterly amazed . In the Sophist we have to judge that the same is d$erent from being. It is true that once we have the apparatus. say. the Eleatic visitor wishes to say that he does not infringe it in the case of being: not-being is not an opposite of being. he says. or large of small. But the means of making the distinction have not been lacking: motion and rest cannot either of them be being. as. and not the same by participation in the other in regard to a host of forms.28 The Ancient Greeks The New Theory of F o m 99 and it will be deprived of' itself'and will not be being. I do not see that Plato distinguishes between 'the copula of predication' and identity by his theory that participation in a form g is compatible with participation in the other of g. part. to participate in that part of the other which stands over against 9.e. one and many. Not to be is to participate in the 0 t h of being: not to be g. rest and motion could combine . No: what Plato is concerned with all the time is the possibility of contradictory predication: that something may be both g and not g (and that in particular sometimes p itself may also be not g). ( I sge).Mass. The only way for being to be a whole is ibr there to be more than this whole. being. also. for all forms. "the other of being" holds.-which is precisely other than nothing but the beaut@l(s57). with being and whole each having assumed its own separate nature. and in acidition to not being.We may say: Isn't this a frightful sophistry? Beings lilay well not form a totality. passages of dialectic as Parmenides intolerably paradoxical.

(3) each existent form is a whole composed of the form and its being. i. Take rest and motion. but on my interpretation i t makes no difference to the sense. how can that exclude anything else? We saw that Plato does not want automatic exclusion of g by not-p. We may add that one will. Not. he certainly commits it in regard to being. the very advantages for which Plato embraces his theory pose a problem. - . That. if'we follow Simplicius. That is to say. If we take "Rest participates in the other of motion". may simply mean that it is not the same as motion. how do we ever get exclusion? We may pause to take very brief stock of the problem of negation. like being. or blue. if it is a fault not to distinguish these. in the same two ways. or the language about being. one and other. and we hear of the part of the other that stands over against the being ofeach. At least in the case of being it must be granted that "x participates in the other of g" does duty both for "x is not the same as p" and "x is not 9". whenJohn is absent as well? On the other hand wemust not think that "A is not red" is without more ado equivalent to "A is yellow. "x does not exist" was not "x does not participate in being" but "x participates in non-being". and use this to explain why "notp" says something. then the idea of some forms as having parts is ofextreme importance. But wen if we have "Rest participates in non-motion" that is still compatible with "Rest participates in motion". That is to say. I prefer the MSS reading. this was just what Plato wanted : but now. Again. Rest participates in the other with regard to motion. The question then arises: why should we then not have bothp and notp? If "not p" says nothing about the subject matter of "p". will be "parcelled out" among all things. if we have this.. the being in one of the forms of the early theory. then we wonder how we are saying something true by saying that something is not the case. The fbrmer predicates ofTheaetetus the other of flying. But then either this. But if'the proposition is false. i. In connection with negation. To rdkr to rny diptych analogy. the latter predicates the other of' being of' what is said about Theaetetus in "Theaetetus is Ilying". no participation of Theaetetus in any form corresponds directly to its being false. This makes what is said a bit of non-being. Thus there will be a part of the other (the bottom right hand layer in my diptych as it lies open) which is a part of being that stands over against being. when a form A participates in a form 8. we decide that there is something in the idea that "It is not beautiful" says "It is other than beautiful" or: the negative proposition uses the negated proposition to fix a possible being-the-case whose sole characteristic is to be other than what is said by the negated proposition. Of course. refusal to combine.The New Theory o f Forms 30 The Ancient Greeks S1 'Thus "Theaetetus is not flying" is not to be identified with '"Theaetetus is flying' is false". If 1 arn right. and does not make out that being is quite exceptional.e. so to say something is not p does not quite generally exclude its also being 9. Suppose. as meaning either "Rest is different from motion" or "Rest participates in non-motion" we probably do not depart from Plato. in any primafncie intolerable sense: we may rather compare Frege's doctrine that the false proposition is the name ofa different object (the False) ti-om what it would have been a name of if it had been true (whereas he rejects the intolcrablc supposition that sense depends on truth-value). how does it say anything? If it says something. how is it distinguished from John's absence. i.e. etc. ( 2 ) the one being is a whole of'parts. however. This will be a part of being to which Theaetetus is related by the (unexplained) relation of participation il' thc proposition is true. of the other ofbeing. If something's not being the case is nothing. (z5Rd-e) it is especially stressed that the other is divicied up into many bits and parcelled out among all things in relation to one another.". one and same hold for all participation by for& in forms. We seem left with having to say "Rest does not participate in motion" in order to explain "Rest does not move". like being. and same. For the language of being divided up and parcelled out occurs also in the Parmrnider in relation to one and to being ( i l g ) . the more generic form. or.e. of each part ofthe other that stands over against being. this is by A's being as it were a pattern in a part of 8. This part of the other will itself be divided into parts each of which stands over against part of being. But he does want it sometimes. "run through" everything. he has as it were hit that part ofthe outermost layer 011 the right hand side of'the diptych which would be marked out by a perpendicular projection from the 'flying' pattern. Thus the falsehood of the false proposition has a close correspondence with the truth of the negative one. to understand "notp" we cannot run to what is or may be the case instead ofp. in one sense of "what is said" this rrlakes what is said different according to whether it is true or false. or the part of these forms? What is being wholly other than Plato's expression for the relation of rest and motion (snge)? Further: what about Theaetetus not flying? That will present us with a problem of how to exclude "Theaetetus is flying"? Why not both at once? Once more: if not flying is a sort of form. This gives us three points : ( 1) the being and unity of each lorm are parts of being and of the one respectively. Are we simply left with non-participation. the speaker has in fact (whether he knows it or not) named a hit of non-being. Non-being is part of being. when we consider "Theaetetus is flying" we will look to the pattern offlyingstamped on the left side. however. when "~Tllraetetusis flying" is said and is false. and it seems immensely unlikely that this part of the argument there was not also pan o l Plato's final vim. among which are the existent unitary forms of the early theory. The lines of projection f'rom the pattern will pass on the right side through the boundary of the hole in that leaf which represents the other of the pattern. then. let us say. In the Sophi. B. it would leave it possible for rest to move. if James' absence is a nothing when James is absent. is a sort of grain stamped on to . then how does it exclude any other ? It is possible that the peculiarities of participation by forms in being. is anomalous. and this Plato does not want. same and other.

For forms and particulars the suggestion seerncd hopeless. (iv)perfectly specific forms considered in themselves. Pannmides. the many forms that are wholly separate and distinct apparently are so much so as not to provide any sort of h e .individual. Thus Plato's aim is to restrict and clarify objectionable contradiction while ur retaining an unobjectionable kind. etc. Inconipatibility. so that the same model ought ro kork lbr more ordinary genera in relation to their species (certainly Plato was much concerned with the proper division of kinds in all his later philosophy). I suggest: (i) and (ii)make the contrast between a form p such that if a lot of things are peach is one single distinct p. And. negative forms will bc incompatible with their positive correlates in the same sense. letters in a (iv)Finally. was not being I). 'whole'. or that he has said they are: that is the serious refltation (nggcd). The serious business is to tollow a ~nari's arguments and show that he has said that the same is different in the relevant way. It appears to me that nothing discussed in the Sophist corresponds to this. The man who can do such refutation will be the dialectician. 'being'. which do not function so. and are conceived to contain other forms in an external fashion. or many elements in a complex.39 The Ancient Greeks The New Theory o f F o m 33 a part of one-being-same. because it was cornpatible with being I) and being q. and A a lurther pattern stamped into part ofthis grained area. (Sophist. It nevertheless remains obscure why this last sort of incompatibility should carry over to particulars. to show that they are different in that way. and so on. the relati011 in question is that of being covered by the fbrni. Theactetus. together with the arguments which we have considered as so important in the Pamenides and the Sophist. (iii)transcendentals . form. and vice versa. to argue that being p excluded being not g.being wholly other than. It is easy to find superficial contradictions. He is marked precisely by his 'adequate discernment' of the different sorts of one-many relation that we get when we consider different types of relations among forms : Dialectical science makes divisionsaccording to classes. and that one of the ideas he had to contend with was that being q was compztible with not being p. (ii) genus species. That is: simply to show that the same is different and the different the same. rather than be stuck with Par~nenides the early theory of forms. as: many species in a genus.tics. he says. And if a man is arguing. (ii) many forms. (See 950b where rest and motion are said to be comprised by being. and not being (p. 145c. IV.m a n y contrast like the other three. but something which I have already mentioned from the Parmenides does: the series of natural numbers is there treated as a series of such forms. Thus inconipatibility of' forms with one another does not autolr~atically carry over. I believe it is usually taken that (i) gives the relation species . and for other relevant verbal and doctrinal echoes. wit11 the exception of the trariscenclentals themselves.other forms. on the other hand. i.will then be not being apart o/. yet I think it does not weigh every word sufficiently. He can be at rest in one way and in motion in another. . (iii) a single form constituted by many wholes in one. so cannot Theaetetus bc Hying too. each found single and separate. one by one part ol'the Form and another by another. why Theaetetus cannot be Hying and not flying. and it is no serious refutation of'a man to detect them in what he says.'s participate in a the Parmenide~. and such forms as 'one'. allnost incomprehensible to us. (iv) many forms that are separate and wholly distinct.I Thus genera in relation to individuals belong to (i)no less than species do. comprised externally by a single form. tor torrns and forms it might be profitable. the big small and the small big. This interpretation is indeed rather closer to the text than others. To adopt this suggestion would be to take up a \lint from the first part of Partnenides asks whether when the many q. If you can do this. when he is not flying? 1 leave B e question as it stands because my purpose is only exegesis and so far as I know Plato off'ersus no solution. i49e. word or particular characteristics in a man (cf. but this is childish. ~ogc). In spite of these ditticulties raised by his theory. We know thatAristotle(Metapl\~.e. But if he has taken two things as the same in some particular way. iv) had an imrnerrse struggle. mutually different. Plato does not want to canonize contradictory and opposite predications. you don't refute him by showing he has treated the same as different. I t certainly fits Plato's account of participation by forms in the transcendentals: and there scems to be no reason to think that he did not regard these as simply supremely generic. (iii). nggdl This passage is obscure and of disputed interpretation. you will adequately discern (i) a single form wholly penetrating many things. is a foolish game anyone can play. appears to me to concern the genus-species division though more obscurely: possibly we get (iii)wherever one could find many wholes gathered together in some sort of unity. and does not take the same kind to be different or a different kind to be the same.

would not insist on the argument. or no one. Soc. I know that no one ever taught him. Soc. Soc. because I have often heard that if you say what is false ydu say nothing. That the slave. Well. Yes. I am sure I agree with what you have just said. Soc. and so it seems that if you are deceived by me you have understood nothing. I continue the dialogue: Men. especially as he was born and brought up in your household. isn't that the time when he was not a human being? Men. obviously. Yes. I shouldn't insist absolutely on the argument. Well. Soc. or always had ? Meno. absolutely do battle. either someone taught him them. Soc. Obviously. isn't itclear that in some othc~ time he had them and had already learnt thern? Men. Soc. for this: that if we think that we must search out what -. 85d9-86~2. Is there anyone who has taught it all to him? You ought to know. true opinions are to be in him during the time both that'hewas a human bring and that Ile wasn't. but he somehow acquired them by himself. . he is not in a positiorl of having got it in the present life. Sonlehow you seem to me to speak well. did you not tell me that no one taught him geometry. Yes. If then the truth about real things is always in our soul. Continued ' anyone is ignorant of we shall be better and braver and less lazy than if we think that it is impossible to find out what we don't know and that we need not search for it. and these opinions when elicited from him by questioning become knowledge. won't his soul have been learned for the whole of time? For it's clear that for the whole of time he either is or is not a huniarl being. Then you taught me that? Men. or do I also believe it to be true? Wait a moment. But if he got it at some time. I don't know how this is something if it is not true. indeed. And to n~yself. yet it seems to me I said something and you understood what I said even if you were deceived. both in argument and in action. If you understand that no one taught him geometry. Soc. So tell me your doubt.Understanding Proofs 35 4 Understanding Proofs Meno. Yes. And I understand it? Men. But if he acquired these opinions at some time. O r hasn't he? Men. Socrates. This knowlrdgc. Or has someone taught him to do geometry? For he will do just the same for the whole of' geometry and all the rest of mathematics. Obviously. Meno. Men. Socrates. Soc. Meno. I am in doubt about this. Men. Or is it? Soc. Is it'not possible for me to have disbelieved you? Men. But he has these opinions. But why should you say his soul is therefore immortal? May he not have existed for a little time before he became a human being.--Prom Philosophy. So oughtn't you boldly to try to search out and be reminded of what you find you don't know now. and say that nevertheless I understood something? Men. and in that little time have acquired these true opinions? Soc. Soc. Soc. Though you Purely by questioning Socrates has elicited frorn an uni~istructedslave the conclusion that the square on the diagonal of a square is twice the original square in area. Yes. then. already had and had learnt these opinions seems to me to be proved beyond doubt. But I should. Then if someone taught him. Could I understand it and not believe it or believe it and not understand it. then. . he either got some time. you also believe it. But what if you were to be deceiving me. what kind of teaching do you imagine it to have been? To explain what I mean. But when I think again of what I have already assented to. or must I do both if I do either? Men.54 (1979). far from small. Soc. Yes. If. yet it is not a small matterwhetherwhat you said and I agreed to was right. Socrates. But if he did not get them in the present life. il'he always had it. Do I just understand it. Socrates. Obviously you understand 'it. let it be as you say. Shall we then refuse to be tempted by this question. Socrates. the soul is immortal. No. . Do you agree? M a . Yes. which is what you don't remember? Men. how what is false is something. For the rest. Men. Soc. Soc. and you will see what I mean. and did I not learn this from you? Men. he was always a knower. Then comes a part of the dialogue which I translate: Socrates. before he was a human being. He rriust have thern.-. if I could. Yet in some way you have understood something. Well. that he has now. for you have understood what I said and I said no one taught him geometry. I feel a doubt. and someone did teach him? Should I be understanding nothing or something? Men.

That he was not taught but that he just at some time acquired these opinions by himself. I'hat is. since you will ask about this proof in turn whether it can be first understood and then seen to be correct. Then if you suppose that he was taught. Sot. by telling him? Men. not if he only repeats it.stand you and not believe you? But is it possible that without urlderstanding I should believe you? Men. Soc. opinions were already in himself and he was only reminded. Yes. Soc. Soc.. In a proof one says "therefore" then . then we shall fall into unsurmountable difficulties. Sot. and such things. No indeed. Yes. Yes. you say. Soc. but when he understands he can deliberate with himself or someone else whether what he understands is true or false. that is what I think. Socrates. Then let us return to the georr~etricnl truth which. I mean when we said: a man first has to understand what he afterwards believes. Why. Tlier~ whcn you tell me no onc taught this slave geolnetry. . or that it is said that it follows? . did you make me understand in such a way that I also could have disbelieved' i Men. Men. whether he was taught or questioned before he was a man or afterwards? But before we ask this I think we have got to go back on our tracks. Suppose that you were considering the proof of that same true opinion that I recalled to the mind of the slave. that the proof is a true proof? Men. or someone else? Men.36 The Ancient Greeks Understanding Proo/s 37 Soc. Well. Men. No. Yes. Rut first he must have understood. I suppose so. Soc. Because I believe there are many more. that must certainly be the case. Then you find out. That is. certainly.". Meno. But if he himself perceived that these opinions were true then these believed you. And should I not be disbelieving what I understood as mud1 as. Yes. I may under. Then it was not because he was told. Then did you also make rne understand what you made me believe? I mean. No. Yes. that is clear. t>utit'I make a diffrrence. Soc. Yes. let us try the second possibility that I mentioned. Yes. Did this person teach him just as you taught me that he had never beer1 taught geometry? Men. Soc. do you think that you could first understand the proof. I should be believing what 1 understood? Men. or the reason by which you find it out. he cantlot know such things unless he himself perceives that they are true. Surely not. Men. not knowing whether it was true or false. But why d o you call this only a pan of the difficulties? Soc. Socrates. that he knew? But if so. from something or by reason of something. Men. Not even if he is told proofs and told that they are ~ r o o f s ? Men. I d o not know if there was a mistake. Yes. when you said no one taught the slave geometry.rld he have t o believe what he was told? Men. That the truth of what is finally said follows from what is first said. must he not? For you said that one first understatlds and tht*nbelie\ees. 1 suppose he was not merely told but himself perceived that it was true. Soc. would you then say he knew? Men. Men. why did he need a teacher? Or couldn't the teacher have done what I did and so done as well as by telling him or even better? Men.. Soc. if I am not wrong. because if I am notwrongwe have madeamistakewhich will mislead us. But will not what you find it out from. Soc. Then he certainly knew these things or had true opinions of them before any teaching? Men. for if' I say I made yc~u untlerstand I make no difference between understandir1g and belief.ue. Men. In what way do you think he acquired them? O r that they came to be in him so that afterwards whenever he was taught or questioned he was reminded of them. be the proof that the proof is correct? Now d o you see a part of the difficultieswe have summoned up? Men. . but if there is no mistake as far as concerns geometry. Or if he did not himself perceive that what was said was true. or get anywhere by only repeating what he has been told. Does one understand that it follows. you will find that this teaching could be nothing but reminding. but nevertheless believed it and learned to say it as a result of being taught. Soc. And what does one understand in understanding these words? Men. he acquired a littlr time before Ile becarrle a human being. What difficulties? Soc. I suppose so. but because he himself perceived that it was true. for what you believe you must first understantl. how could there be a mistake about that? Soc. Obviously. and believed it. and why sho~. Soc. and which we are both supposing someone taught him. and then find out somehow that it was true? Men. I see them. he understood in sach a way that he did not yet know whether it was true. Yes. Soc. Soc. then he knew it was true? What divine authority told him. As far as concerned what you told me. 1 did. But when he was told it was t~. Soc. Who made me believe the slave had not learned geometry? You. if I had Soc. does one not? and "if. then it was riot I but sorrleone else who made you understand what I was going to say. Men. The second? What was that? Soc.

Yes. by my present argument. hen one understands that it follows? Men. and would pass. while someone who was doubtful about the guilt of a criminal who was guilty would not be wrong. does one first understand and then believe that it follows? Men. when one understands "therefore". but what is really the case. as soon as one understands what is said and not later. certainly. and all the more believes. Soc. whether or no we were right before. But someone who thought it doubtful whether something followed that did perfectly follow would be wrong. But if it tollows and in understandingwhat is said with "therefore" or "if . if you do not mind. Soc. and would pass from error to truth when he saw that it followed. why should 1 mind? Men. Socrates. Yes. . Men.38 The Ancient Creeks Understanding Proofs 39 Men. Yes. Yes. Soc. Soc. in a way that may persuade you that we neverttieless cannot avoid the question here. Meno. but in such a sense that understanding does not mean believing. and that you do not understand them if you do not see that they are true. Soc. tliat it follows. Soc. "it follows". if it was in fact true. If you could understand them without knowing them. Soc. and also if I say that your slave is handsome. Socrates. that mathematical truths are the same in this regard as proofs. O r ~ understands e this expression. But in the case of the slave. Meno. Let me say for the moment. Then I ought always to believe what is in fact true. though it might have been false without this making any difference to what one understood in understanding it? As for the cases where there is a doubt. Yes. that one understands that this case is like what really does follow. Soc. that it follows. I warit to deny what I said before. then . but from ignorance to truth when he saw that the criminal was guilty. but . Soc. But that we were then not altogether wrong is something I think I can show." one understands that it follows. But if I aftenvards learn that something does tllow. Namely. both. I have heard it said. still you will not understand me if I say that since I am shorter than you your slave is handsome. No. Why. That cannot bc right. did I understand what you said? Men. and I have mentioned only a few. But does one under~tand what is said? Men. couldn't you know them simply from being told them by those who d o know? For everything that results fmm knowing depends on unbe known to be true if only you have information derstanding. Yes. But I spoke of many difficulties in saying that in proof and mathematics we d o not first understand and then believe. Wait. and your arplnent now is just the one that we avoided about the false staterrlent that the slave had not learned geometry. and like precisely in following. and it is in fact the case. Men. or when you understand at that same moment you will know the truth of what I say. then one knows. For in the case of the slave. Will you not say with me that where one perceives following one does not first understand what is said in "This follows from that" and then perceive that it is true. Soc. since false proofs are given of many things and one finds out that they are false. Rut we agreed. was I wrong to think it doubtful? Men. Mmo. Why do you throw them into your remarks as if you had proved something about them? Soc. Soc. Soc. Though you understand "since" and "if then" and "therefore" and "follows from" and though you understand me if I say that I am shorter than you. riot to be tempted by this problem of the false. not the truths of geometry and the rest ofmathernatics. even though I now regard it as doubttirl? Soc. What you have just been arguing concerns proof. Ifwhat we said was right a false pruol'woirld be nothing. Really I do not know what to say about that. and either you will never understand because there is nothing to be understood. So what you thought absurd is in fact true and a false proof is nothing though it pretends to be something. Soc. Socrates. nothing you have said obliges me to say tht same about mathematical truths. Soc. or that it is said that it follows? Mm. not from the doubtfulness of that which is understood. And could I or could I riot learn afterwards that what you said was true. doesn't that show that I was wrong before? Men. as when a jury is convinced by false rhetoric that something is the case. when before I regarded it as doubtful whether it followed. You yourself have agreed that you cannot know them from being told them. even without reason? Men. ' But what does one understand to be said. Soc. Then in what really does follow. sirlce one ~tnderstands both that i t followsand that it is said that it follows. There is one other that I will mention. and say that one understands only that it is said that it follows. I believe we cannot avoid that problem here. Then someone who remained in doubt in such a case would be less deceived than the others ? Men. Is there such a thing as an irrational true opinion. Mm. and not at separate momhts. . . . Of course you did. But don't consider just your present argument. Yes. the doubt comes of not understanding. not from error. when I understood you even without believing you. real proofs turning up. While I may agree that I cannot understand a proofand yet not know that it is correct. but not for the reasons for which they have believed it? Men. If it were true you could.

such as diagrams. generated together and inseparable. which does bring in matheirlatics and does not only refer to proof: When I persuaded your slave to recall what he already knew. what were you going on with when I interrupted you? Soc. these are all he has W o r e him. Yes. its true opinions. In that case. Soc. have seen with the eye of his soul the real squares and areas that are the objects of geometry? Soc. What I am asking you is whether when the soul views intelligible objects it views them as the eye views colour. answering this question will help us with thc other. then. for it would mean the colour that it saw and its understanding what it meant and knowledge that what it meant was present to it would be the same thing. We are still thinking whether understailding and knowing are the same or ditferent.40 The Ancient Greeks UnderstandingPro@ 41 that what is ullderstood is true. o r judges them as sight judges other things? Men. and if I am not mistaken. This is quite familiar to me. Men. certairlly not. and. Soc. namely that without being told a man discovers the truth. not even figure or size. But when the soul has forgotten what it once knew. you are hurrying the argument on too fast. You did. Forgive me. But if one is instructed one is instructed by someone. But colour and sight are twins. In this alone it would need no instruction and it would mean this by whatever it uttered. A further dificulty about first understanding and theri acknowledging. what it would have to say it' left un- . when it views them. Yes. but once that information is given. Could you explain more clearly the distinction that you are making? Soc. I have no idea what to say. he sees colour. and. without any danger of error. but by the help of them he is led to say what is true about other things. the soul is not reminded ofwhat it once saw. No. are awakened and it knows them in understanding them and could never be told them or made to understand them without knowing them. But why should we not say that in what is called learning. Soc. and can see how these objects here fall shon of those other things. Since. that the soul views all intelligible objects. then. because whatever instruction it can at any time be given will only be reminder? Men. I'o measurement. Well. that is what we should have had to do. and further because those things about which he discovm the truth are not the things before his bodily eyes which the teacher points to. Yes. and does not judge about them getting instruction how what it says should be made to correspond to what it sees? Therefore it is not able to judge rightly or wrongly about what it sees. What I have heard is that colour and sight are connatural. so long as it is given somehow? But if understanding and knowledge are ever the same. then. it seems that I say all that. Try and see if you won't find something to say. little before he became a human being. Wait. Socrates. why should it matter how the information is given. instructed would be nothing but the name of the colour that it saw for the moment. Soc. and is also able to judge wrongly and therefore understanding and knowledge are distinct. Soc. Yes. Would we have known it was true of them if I had drawn them with the greatest possible care? Or should we have needed to have recourse to something else to find that out? Men. Meno. Do you not say. Hence if sight could judge and say what it saw. if by observing you nlearl perceiving by means of the sensrs. Men. I will try. Men. Men. about what was he showing that he had true opinions? Men. But anything else it needs instruction to judge and it learns to judge rightly. Soc. If anyone sees at all. but now for the first timeopens its eye to the objects about which it is taught? 1was convinced that what is called learning is not really learning but recollection by the arguments we have gone through. and he said that the square on the diagonal was equal to twice the original square. for I noticed you drew them rather carelcsslv and what was said was not true of them. I agreed that he was reminded . About squares and areas. if I am willing to doubt what I have been most sure of. But why should he ~ i o ta . Soc. But did we not agree that the soul cannot be instructed by anyone in its knowledges. understanding what it says without knowing it to be true and afterwards finding out and acknowledging that it is true. S w . understanding is enough. and there is nothing else that he necessarily sees. when it is reminded of it fragmentarily and with difficulty. as the eye views colour. Then the squares and areas about which the slave recalled his true opinion were not these nor any he could observe either in this life or in any other. which have been quiescent in it all the time. what it sees is related to it as colour is to the natural eye or rather as are body and shape and everything else besides colour that we learn to distinguish using our faculty of sight? Men. isn't one? Men. Do you imagine that when the soul's eye looks at intelligible objects such as the objects of mathematics. Which ones? Was it for exarnple about these figures that I drew on tlie ground ? Men. as we now have seen. Socrates. Did I not say that sight needs instruction to judge anything but colour? Men. then it will not be possible simply to be tolcl. that he could never be told either in this life or at any other time but must perceive the truth of himself if he is to know it. Men. What else? Soc. but 1 do not know if I shall succeed. whether it is proof or mathematics or sonie other subject matter as well that we have not touched onat all-whatever cannot be known bv one who is merely told it and learns to repeat it will be such that unclerstandirig and knowledge are the same and error is impossible: but what poses as saying something and would commonly be called error is saying nothing.

Obviously. Did we not say that sight and the object of sight were twins. or ofwhat size or at what distance it is? . Didn't you agree just now that when the soul saw the real. neither did the soul's understanding ever come into existence. But now tell me this: d o the objects of the soul's understanding come into existence? I mean. then it makes an end of all the argument about the irritnortality of the soul. Men. And is there not a kinship between twins? Men. The reason why these truths cannot be taught is that a teacher can only draw his attention to signposts and hints whereby he opens his intellectual eye and looks in the right direction and now utters and sees the 11-uthofwhat he says about the real mathematical objects that he is looking at. Soc. Now answer this. No. Yes.Men. as it is sight and colour. Soc.) Men. Metlo. this could be compared with sight when it sees colour and not with sight when it learns to judge what all object is.ject it? .Men. But d o the square and the diagonal and their relations come into existence? Men. Yes. for there is no place for error or instruction but what it utters means what it sees. . Very well. Soc. is something that comes into existence. But now it has struck me that perhaps he has the real things Understanding Proofs 43 betbre his mind's eye. Soc. generated together? Men. not the eye and colour. if you can: does sight work to see colour? I mean. that are twins. Then are not the soul and what it understands similarly twins? and so kin? Men. But you will not object it' we scrutinize your argument t o see il'we should keep it or re. No. that is true. must it labour and be trained and have practice and will it stumble and niake ~~listakes. Meno. it is the soul's understanding and its objects that are twins. as: a horse. and not that it always existed. Soc.42 The Ancient Greeks of'them. Soc. and what he does manage to remember he remembers piecemeal and using many aids and a great deal of hard work. Then if the soul's understanding and these objects are twins. Yes. Soc. But can you show rne how all this argument about knowing and understanding helps the main purpose of our discussion? It still seems to me that we have proved only that the soul must have existed for some time before this life. Yes. it would be without error. Soc. I think I see what you mean. Then. but he stumbles and gropes like a person trying to remember. producing utterances which rnean nothing though it supposes thrm to mean the colour that it sees? Or did we agree that if' left only to announce the colour that it saw. If' i t slioulci be possible to say this. and its uriderstariding of'what it means and its knowledge that what it means is present to it are identical? iMen. I only needed to be reminded in order to return t o my former conviction that what is called teaching is really reminding. or a man. Soc. Socrates. But doesn't a pupil have to d o a great deal of hard work in order to master geometry. Soc. he cannot have the fbrms and mathematical objects present to his soul's view when he has to work and makes mistakes. and still sometimes also when he is advanced (Though you arid I have agreed that these so-called mistakes are utterances in wliich he somehow fails to say anything though he seems to be saying something. For that it comes about that the slave is taught o r not taught in this life. and doesn't he need much training and practice and make what are commonly called mistakes while tie is learning. Well argued. arid the diagrams serve not as reminders but as signposts. Certainlv I shan't ~nind. and still more to learn dialectics. and that I must say that sight does riot have to do any hard work in order to see colour. in the cases that we are talking about.

The ambiguity is first sustained. For it would either be happening or not happening accordingly. ambiguous between being exclusive and being non-exclusive. it is also necessaryfor evetything to be the case or not be the case: the Greek "or" is. but everything of necessity and not 'whichever happens'. Here it is exclusive. and then resolved at the end of the chapter. then just one of the two speakers must be speaking the truth. this experiment will result in the mixture's turning green. on the conditions that every affirmation is true or false. by which my explanations can be verified. that these propositions (or their negations) must be true or false. with the negation sign outside the whole proposition. for singulars too. other than the present. but see the 'Elucidation'. From his remarks 1 infer that these would be correctly rendered "men are . in the other denied. then one side of each antiphasis must be true. 'whichever happens': the Greek phrase suggests both "as it may be" and "as it turns out". But that if'you take "All men are white" and "No men are white" and construct the antiphasis of which each is a side.) for universab not univmal[r qwntiJied: his example rendered literally is "man is white. you will be sent down before the end of term.What Aristotle says ill this sentence is ambiguous. Aristotle is not overtly discussing the expression "whichever happens".e.rnust be part of any proposition. clearly it is necessary for one of them to be speaking truly. it is not the case. An antiphasis is a pair of' propositions in which the same predicate is in one affirmed. But for what is singular and future it isn't like this.Aristotle and the Sea Battle 45 beautiful. then it is false. the other false. . that "or" was non-exclusive. singular andfuture: sc. like the English. . Throughout this paragraph the ambiguity is carefully preserved and concealed. should be true or false. The 'scare-quotes' are mine. and for universals universally quantitied it is always necessary that one sl~ould be true. it is true to say or deny it. so that it is necessary as regards either the affirmation or the negation that it is truror false. the affirmative proposition must be true or false. it isn't like this: is not white". It is true to say or deny it: r)v is the common philosophical imperfect. So if one man says somethingwill be. Notc that Aristotle has not the idea of the negation of a proposition. These have been discussed. in the immediately preceding chapters. For both will not hold at once on such conditions. For both will not hold on such conditions: namely. men are beautiful. of the same subject. as in ( I ) "every affirmation or negation". there will be a relevant discussion tonight. and for singulars too.Socrates is not white" one side is necessarily true. its being white or not white is necessary. This condition is not a universal one. For whichwer happens is not more thus or not thus than it is going to be. as will appear. and if it is false. . for universals universully quant$ed: he does not mean. and if it is white. For. the other necessarily false. I believe that we (nowadays) are not interested in these unquantified propositions. Chapter IX These have been discussed: i. Nor will it be or not be. and the other side must be false. for they are ugly too. that of "All men are white" and "No men are white" one tnust be true and the other false. (4) So nothing is or comes about by chance or 'whichwer happens'. then. not. and if they are ugly they are not From Mind. or not white. that was ( I believe) invented by the Stoics. 65 ( 1956). as has been said: sc. and similarly for the negative. For either someone saying something or someone denying it will be right. of "Socrates is white . 5 Aristotle and the Sea Battle De Interpretatione. . then it is also necessary for werything to be the case or not be the case. (This is what a modern reader cannot take in. namely. while for universals not universally quantified it is not necessary. For what is and for what has come about: he has in fact not mentioned these. and they are also not beautiful. it is necessary that hence the "or" in the conditional "if every affirmation and negation is true o r false" is also exclusive. For if it is true to say that something is white or is not white.". as has been said. "All men are white. though if the "or" is non-exclusive it does. he says. "As the case may be" would have been a good translation if it could have stood as a subject of a sentence. and another says not. For if wery affirmation and negation is true or false. But if the conditions hold. except to say that a verb or a tense .Not all men are white" and "No men are white -Some man is white". (8) (1) For what is and for what has come about. ifevery affirma(3) tion and negation is true or false. it does not apply to the unquantified propositions. as this place by itself would suggest. which he regards as the verb par excellence . or negation. it is necessary: given an antiphasis about the present o r past. And if it is not the case. that this is deliberate can be seen by the contrast with the next sentence. and to point this he says "every affirmation and negation".

". but you don't think the present itidetermiriatr. so why say the tiiture is?" O r rather (as he is not talking about the expression): "Whatc:vcr. but ( 1 1) Still." (5) you say is true. to say that neither is true. this coat is capable of getting cut up. which leads to the rendering: "These a n d similar strange things result.And besides. and so this is not a vety strong objection ifwe are willing t o try whether neither is true. This does not appear to m e to makegood sense. i t is so far as I know a matter of uridcrstandirig the argutncnt \vhetller you tra~islate as here. is obvious. either. It only follows if the previous arguments a r e sound. this will happen. ratlier thus than riot thus". For if by chance. if not. it was trur earlier to say it was going to be white. Takc a sra-battle. even if there isn't one man saying something and another denying it. for example: it would have to be thecase that a sea-battle ncither came about nor didn't come about tornorrow. i These are the queer things about it. it is not open to US.g.: often rendered "since": "since if we do this. But if it is impossible for something riot to conlr about.happcr~swill be just as deterrninately thus o r not thus as it is. And ifthey are going to hold tomorrow: from here to the end of the paragraph the argument is: if it is the case that somethingwill be. The last two sentences o f the paragraph are incontestable. 'whichever happens' won't be: i. For firstly. both being and not being. So it is clear that not evefything is o r comes about of necessity. so that it came about by necessity. But if it wasalways true to say: "it is. this will happen. that there is no 'whichever happens' about what comes about. H e is going to reject the conclusion. if something is white now. . then not by But i/it is itnporsiblejnr something not to come about. the affirmationbeing false the negation will not be true. and it won't get cut up but will wear out first. and equally that there is in general a possibility of being and not being in things that are not always actual. Aristotle is arguing: "We say 'whichever happens' o r 'as the case tnay be' about the present as well as about rhc luture. then it is unable not to come about: tlie reader who works through to the end and understands the solution will exarnirie the dialrctic to see where it should be challenged. so that it was always true to say of any ol the things that have conic about: "it is. And if' sornethi~ig is neither going t o be nor not going to be tomolrow. and for everything that has been it always held. then: impossiblr for that not to be or be going to be. i f . 'whichever happens' won't be. and this being false the affirmation won't be true.either. but that everything is and comes about of necessity. Therefore it is necessary that everything that is going to br should come about. It will turn out that the point is here. . (8) Fur tlter. then it was necessary that this came about. o r deny.: "if we do this. or will be". both are open. In them. ( 9 ) For there is nothing to prevent its being said by one man and denied by another ten thousand years ahead that this will happen. nor its having been ten thousand years ahead or at any time you like. e. if not. But if something (6) is unable not to torrle about it is necessary for it to come about. This seems illogical. So nothing will be 'whichever happens' or by chance. they must hold tomorrow. And if they are going to hold tomorrow. Tlie dialectic is very p o w e r h l . not". then it can't not happen. not". it is not open lo us. So ifin the whole oftime it held: o n e must beware o f supposing that Aristotle thinks the conclusion stated in the apodosis of this sentence follows from the condition. so that whichever of the two was then true to say will of necessity happen. . e. For if anyone has truly said that something will be. in spite uf having familiarized tnvself'with the arttillriess of the chapter.: "isn't o r (SC. And plenty of things are obviously like this. And equally it is capable of not getting cut up. So if in the whole of time it held that the one was the truth. to say that neither is true: and yet Aristotle is often supposed to have adopted this as the solution.46 The Ancient Greeks Arijtotle and the Sea Battle 47 is not more thus or not thus than it is going to be: as the Greek for "or" and for "than" arc thc! same. So that there would be no need to deliberate or take trouble. for clearly this is how things are. however. for example. o r (as is rriore usual) e. And it was always true to say of what comes about: it will be. And: I have diverged from the usual punctuation. for its getting worn out first ( I g) would not have occurred if it had not been capable of not getting cut up.g. and also both becoming ( i n ) and not becoming. In more detail: you say. and) isn't going to bc. 'whichever happens' won't happen. So this applies too to all other processes that are spoken of in terms of this kind of possibility. two things about the future. in spite of the equivalence of the two Greek expressions. And indeed it makes no difference either if people have said the opposite things or not. then when the time comes you must b e able to say those two things in the present o r past tenses. if it is true to say that something is big and white. T h e Oxford translator sits o n the fence here. whether for universals universally quantified or for singulars. But this does not make good sense. if it is necessary that for every affirmation and negation. Now if this is impossible! For we see that things that are going to be take their start from deliberating and from acting. it does not hold". nor is it its having been asserted or denied that makes it going to (10) be or not. then it is unable not to come about. Forfirstly: this goes agaitist what h e has shown at the end of (3):"if it is Mse. These are the queer things about it. as: that it ncither ( 7 ) will br nor will not be. then it will b e the case that it is. Ifwhat . or will be". S o much. one of the opposites should be true and one false. but there is n o reason to think that he rejects the condition: o n the contrary.g. What follows is conclusive. And there is more of the sort. I cannot read this passage without being ~rionletltarily convinced. Still.e. both must hold.

For it is necessary that everythingshould be or not. it is clear that where things are such as to allow of 'whichever happens' and of opposites. moon. nor that it should not happen.: it is a matter of whichever happens. It will by now have become very clear to a reader that the implications of "necessary" in this passage are not what he is used to. see Nichomachean Ethics. for it does not hold for what does not exist but is capable of being or not being. states no facts. An animal too. On the other hand. For it isn't the same: for everythingthat is to be of necessity when it is. Btd it is not the case. butwhicheverhappens. like o d ~ e r l . I 139: there he repeats the wot-d " d p X q " : " r ) roirrcwq dpxq d v O p w n o ~ " :the cause of this sort is man. is necessary. VI. I cannot d o anything else with the typewriter. logical. dbq in a non-temporal m u is. planets and stars. past. one may ask. Hence his thought is not that there are new starting-points constantly coming illto existence. but it is as we have said. this must hold for the antiphasis too. werything that isn't not to be. namely. thougl~1 should like to enter a warning against the idea (which may present itself): ''the nature of deliberation presupposes freedom ot'the will as a condition. and should be going to be or not. in part. which Aristotle could not object to. So that since propositions are true as the facts go. . a prime mover. and that oneshould be true rather than the other. note that this is governed by 'it is necessary'. (14) The existence ofwhat is when it is .' r)bq works rather like the German "schon" (only here of course it would be "noch nicht"). or false.frequent in Creek literature. The word also means "principle". that it is necessary that there will be a sea-battle tomorrow or not. he meant that ifp is a statement about the present or the past. but one that works by deliberating. Thc sentence confirms my view of the point where he would say the dialectic went wrong. . what has this necessity got to d o with logic? Aristotle.) His results could perhaps be summarized as follows. Not that this is Aristotle's grouild for speaking of the general possibility. but it is not thecase that for allp. things that are not always actual: things that are always actual are the the propositional sign to indicate present and past time references on the one hand. e. But it is not (15) the case. but into causes of events and observing that among them is this one. (I d o in what follows. but was. but still it is open for either to happen. I cannot pursue this here. wtiicti is defined in terms ol'deliberation. Hanlcy of Somcrville College for pointing this out to me. take their start: literally: "there is a starting point of things that are going to be". separately speaking. So. is: and cannot be shown to be otherwise. however. it cannot be otherwise. prrrisely qua caltulatillg. . is a prime mover. whichevtl happm: sc. that nr auhuld be true rather than the other: d 'rather and for the most part' above. But Aristotle's approach is riot that of someone enquiring into human nature. present." That is not an Aristotelian idea. English translators of philosophical texts usually either neglect to translate it o r mist I am indebted to M h M. ant1 with other things one is true rather and lbr thc moFt part. But this means that in order to ascribe necessity to certain propositions (the ones. o r future. that are not 'simply' necessary) we have to be informed about particular facts. or a plant. and the other not.48 The Ancient Greeks Aristotle and the Sea Battle 49 with some things 'whichever happens'. acting: he means human action. and: for it simply to be of necessity. A modern gloss. (16) For such things it is necessary that a side of the antiphasis should be true or The existence of what is when it is. that either ofthe sides is necessasy: the am- biguity of the opening "it is necessary that an affirmation (or negation) should be true o r false" is here resolved. As if a calculating rnachine ~iot ~nerely worked. - This is how it is for what is not always existent or not always non-existent. And the same reasoning applies to the antiphasis. must be wrong. is wccssary: i. I hope this will not prove misleading: the purpose is only didactic. p vel not-p is necessary (this covers the unquantified propositions too) and ppis necessary vel not-p. is necessary. But see the 'Elucidation'. and the non-existenceofwhat isn't when it isn't. that would not matter. he seems to have ascribed something like logical necessity to them. I infer that Aristotle thought that correct statements of probability were true propositions. and future time reference on the other. And we learn that when Aristotle said that. and the atfrrnlation is not true rathrr than the negation.p. nor is the idea as undiscussable as it seems at first sight. that either of the sides is necessary. And the same for what isn't. but that it is not necessary that there should be a sea-battle tomonow. for everythingthat is to be is not necessary. If what the typewriter is going to d o is necessary. but that does not mean: dbq. But still.e. we use indices) and. and without which it is not possible for a modern person to understand his argument. with other things one is true rather andjbr the most part: as we should sav: nlore probable. But for it to come about or not is necessary. 1 mean. Wrongly. in terms ofthis Rind ofpossibility: I take it that we have here the starting-point for the development of Aristotle's notion of potentiality. but not this one or that one. though: Aristotle thought that the heavenly bodies and their movements were necessary in this necessaryvel not -pJ is necessary. no1 for false.g. A human being is a prime rnover (in the engineer's sense). not temporal. So it is clear that it is not necessary for every affirmation and negation that this one of the opposites should be true and that one false.. but that does not mean that it is true. I t is first of all thr nature of deliberatiori that rnakes hirn think that the fact of human action proves the dialectic. then either p is necessary or not-p is necessary. separately speaking. simply to be $necessity: there is a temptation to recognize what we are used to under the title "logical necessity" in this phrase. Aristotie thought that what these do is necessary. Then for allp. The general possibility that he speaks of is of course a coridition rccluired if deliberation and 'action' are to be possible.

) I am glad to be told it will be true. Yes . 201e4. Metaphysics. o r it won't". and I know that it was either when I told you the Vice Chancellor would be I-un over. Can't fail to do so? Well. or on the other hand when I said he wouldn't. Having set out the full truth-table and tliscovered that fbr all possibilities you get T in the final column. that is true. Bonitz gives some more examples. (Partly because of this argument: it is a point of honour with me to go. if it is going to rain tomorrow. not certainty o r knowledge.any enquiry would be comic. (Havingstood up.but is it certain? Do you get it from the meteorologists? I have heard that they are sometimes wrong.satisfied. will be. What? Even if it is going to rain tomorrow? Oh. just because the two remarks together cover all the possibilities.i.future. You will orily be making a fool of yourself. But still I have been told nothing about the.I mean if I tell you then 'True'. If it is going to rain tomorrow it is true that it is going to rain tomorrow. AN EL3UCIDATION O F T H E FOREGOING FROM A MODERN P O I N T O F VIEW I'he Vice Chancellor will either be run over next week or not. by its truth value's being T for all possible combinations of the truth possibilities of its components. is not to tell me of any truth-possibility that it is . But is it going to? I told you. if you like. And that I do by discovering which of the truth- . 187a36. of course. of course why. It's riot true. For exaniples.well. listen to this: The Vice Chancellor is going 10 be run over next week. But to tell me "Itwill rain. perhaps not till midnight. I can't say. De Interprelatione. i. Physics. possibilities is fulfilled. But Aristotle says it isn't true that it is going to rain tomorrow! I did not read a single prediction in what Aristotle said. but that is no more than to say that either he will or hewon't be run over. And therefbrc either he will be run over next week or he will not. I should be enormously surprised if Aristotle were to deny this. but I am afraid you were repeating yourself and. for that is to tell me nothing. then it will have been. But whatever I tell you then will have been so now. But d o you mind repeating your information? Either you are sitting in that chair or it will not rain tomorrow. Yes. see Theaetetus. But surely we are talking about truth. If on the other hand you tell mepv-q (9 being different from p) you do give mc some information.or. what is more. that is why its 'truth' gives no information. that I have succeeded in saying something true about the future. with one of these two remarks. Then why did you say it? I was merely trying to make a point: namely.which I agree is information this time .) - - . Now will you actually tell me something about the future? Very well.1 mean I am able to. I agree. There is therefore the possibility of enquiring whether your information is correct. I mean is now correct to say it is true. it doesn't have to rain tomorrow. 16a8. that "It is true that eitherp or 9" is not a sufficientcondition for "Either it is true thatp or it is true that 9". What have you said about the future that is true? I don't know: but this I do know. because I am sitting in this chair. not for certain. you cannot fail to do so. i oo6a 16. Can an Eitherlor be true except by the truth ofone of its components? I seem to remember Quine speaking of Aristotle's 'fantasy'. that means not just that it will be true then but that it was true now. because since I know I am sitting in this chair I know what I have been told is true whether it rains tomorrow or not . But why does that matter? Can't you say anything for certain about tomorrow? I am going to Blackwell's tomorrow. that I have said something true. 1 can't tell you till some time tomorrow. I understand you. then it necessarily will rain tomorrow: (p 3 p) is necessary.e.e. He only implied that it didn't have to be true that it will rain tomorrow. . If what you tell me is an Eitherlor and it embraces all possibilities. Please understand that I was not repeating myself! I think I understand what you were trying to do. what I mean is that if I tell you as I shall be able to -'True' tomorrow . Now I will put it like this: Aristotle seems to think that the truth of a truth-functional expression is independent of the truth values of the component propositions. for all truth possibilities of "It will rain tomorrow". But I thought it was the great point against Aristotle that 'is true' was timeless. you need make no enquiry to affirm the truth ofpv-P . for certain truth-combinations are excluded. I am sorry. But that is a howler! The 'truth' of Either p or not p is determined. as you know very well.50 The Ancient Greeks Aristotle and the Sea Battle 51 translate it. Have you given me any information about the future? Don't tell me you have. and I am asking whether your information. then I discover that your information is correct. now. you tell me nothing. Either you are sitting in that chair o r it will not rain tomorrow. and if one o f the combinations of truthpossibilities which is a truth-condition for pv-9 is fulfilled. Then I arn goirig to the police as soon as I can. I am absolutely determined to go. And that is certain? Yes.

the prophecy together with its fulfilment .or necessarily false that there was . In this sense I say it is necessarily true that there was not. But '. therefore either it is necessary that he should be run over or it is necessary that he should not be run over". I can't see any reason to trust the test if. As Newrnan remarks. and so on? But that can't happen. and when he is ten years old he will be killed by a tyrant. necessarily true that there is a University in Oxford. - - Aristotle's point (as we should put it) is that "Eitherp or notp" is always necessary. only that it does happen). therefore either he will be run over or he will not" you said that I was repeating myself and could not fail to be repeating myself. Then you aren't necessarily going to Blackwell's? Of course riot. You could have avoided repeating yourself if you had said "The Vice Chancellor will either be run over or not. But. it is quite certain . A. A. I believe you as fully as 1 can. and here "necessarily true" has a sense which is unfamiliar to us.. (Unless indeed you say that waiting and seeing is finding out. For such a prediction is a prophecy. for this is a novelty to us . the appearance of the night being continued for the rest of our lives. are in All Souls today. . I go on to ask: Could anything that can happen make it untrue that the sun rises tomorrow? No. but: since it is certainly true i t is necessary. But equally clearly there is no way of finding out.And isn't what is true about the present and the past quite necessary? What does 'necessary' rnean here. A. she would either be awestruck o r think me dotty. am I saying it wouldn't falsify it for me not to be here? But I am here. then eitherp is necessarily true.But what does Aristotle mean when he says that one part of the antiphasis in necessarily true (or blw) when it is the present or the past that was in question? Itight at the beginning. Now suppose that what I say comes true. if my spoon gets up one day and dances a jig on my plate. B. The whole set of circumstances . But do J A.necessarily true" is not simply the same as "true". That is so certain that nothing could falsify it? Yes. .a big civil war raging in England from 1850 to 1870.and this is from our point of view the right way to put it.g. Well. but I can't imagine anything that would make it doubtful. 1 don't mean I can't imagine doubting it.know you will go? Can nothing stop you? Good.that when p describes a present or past situation. situated as I a miracle.anything ffom a change of mind to death or some other catastrophe.52 The Ancient Greeks Aristdk and the Sea Battle 59 B. Not even: the sun's not rising tomorrow? But this is absurd! When I say it is certain 1 am here. NN. for while it may be true that there will be rain tomorrow. B. 8. Not : if' it is true. 7 May 1954. I tully believe you. it is not necessarily true. And then you referred to the truth-tabletautology account of that proposition. though it qualifies perfectly well for philosophical discussion of physically impossible but imaginable occurrences. Well: the sun will rise tomorrow. A. then (perhaps)I shall agree that it is necessarily true that that thing will happen. all the same the sun will have risen. B. A Of course lots of things can stop me. A. 1 haveany doubt ttiat I am here. and she would be quite right. it qualifies ill as a miracle. this necessity we are familiar with. which could turn out one way or the other. We also say this about things which we don't know about the past and the present. Then what is true about the present and the past is necessarily true? Haven't you passed fiom certainty to truth? Do you mean to tell me that something can be certain without being true? ." Clearly this is something that may be true and may not. Now it' you can show me that anythi~ig about the future is so certain that nothing could falsify it. a miracle ought not to be a silly trivial kind of thing e. The question presents itself to us then in this form: does "may" express mere ignorance on our part in both cases? Suppose I say to someone: "In ten years' time you will have a son. At least.or you . when I said "The Vice Chancellor will either be run over o r not. divides into several pieces and then joins up again. If we continued in darkness. and one's theoretical attitude (if one has one at all) to the supposition of such an occurrence ought to be exactly the same as one's theoretical attitude to the supposition that one knew of someone's rising from the dead and so on. Now if I really said this to someone. Similarly if one were discussing impossible predictions one would take such an example as the following: every day I receive - . Could it turn out that this proposition that you.My reason fbr saying so is that if you cared to suggest any test. As everyone would say: there may be or may not. But as you would have been disinclined to say that seeing no possible meaning for an ascription of necessity except what we are used to call 'logical' necessity you could not avoid repeating yourself. B. B. but it is not finding out that it will happen. it is necessary. is untrue? O r is this certain? No. B. But does not precisely the same point apply to what Aristotle says about "Eitherp or notp" whenp is a proposition about the present or the past? B. let me try again: Could anything that can happen make it untrue that you are here? If not. since it obviously doesn't mean ttiat these are what we call necessary propositions? 1rrleari that nothing whatever could make what is certain untrue. Is that as certain as that you are here now? I won't say. and so on. and the sun will rise tomorrow. A. or-p is necessarily true. Are you necessarily here now? 1 don't understand you.

but stops. in saying after his resurrection: "It is necessary that all the things that are written about me in the Law of Moses and the Prophets and the Psalms should be fulfilled". Thus the misunderstanding dates back at least to the fifteenth century. Now. These were condemned as 'scandalous and devious from the path of Catholic truth' and the said Peter withdrew them. (01 Christ. it would still not be certain. (4) It is not enough for the truth of a proposition about the future that the thing will happen. or that the things they mean could not be prevented by the power of God. "The freedom of the will consists in the fact that firture actions carinot be known. because the facts may falsify the prediction. (5) One of two things has to be said: either that there is not present any actual truth in the articles of faith about the future. absurd actions fbr example. In 1474 the following propositions on the truth of future contingents. The 'logical necessity' of which he speaks in the remark on knowledge is thus not just truth-table riecessariness. And whatever I do (1 do random. so it comes out in Wittgenstein. arid in Aristotle: past and present facts are necessary. when Elizabeth speaks to the Blessed Virgin Mary saying: "Blessed are you who have believed. because the things that have been said to you by the Lord will be effected in you". propositions. ifp is a tautology. a university lecturer at Louvain. she seems to suggest that those . put forward by Peter de Rivo. (In more detail.. So why does he say we do not know that the sun will rise? Not. did not yet have truth. and q expresses a fact that A is 'acquainted' with. were condemned in a bull of Sixtus IV. because the facts could go against it. for someq(q 3 p) is a tautology." We are theretore presented with the logical neressariness ofwhat is known's being true. ( 1 In Luke. 'A knows thatp' is senseless. by the Traclatus account: ifA knowsp. HISTORICAL TAILPIECE The De Interpretdione was much read in the Middle Ages. (3)When the Apostle says in Hebrews 10: "The law having the shadow of ' good things to come. in the Traclatus. to see if' he will still have written a true account) the letter records it. "A knows thatp" makes sense fbr anyp that describes a fact about the past or present. I t is interesting to note that Wittgenstein apees with Aristotle about this problem. together with the logical non-necessity of the kind of things that are known. namely: "You will bear a son and call him Jesus. since we are dealing in what can be imagined and therefore can be supposed to happen. But could the facts go against the sun's predicted rising? Is there not a radical disagreement between Wittgenstein and Aristotle here: Aristotle thinks that it is necessity that the sun will rise. The connection of knowing and the known is that of logical necessity. I t is the unfamiliar necessariness of which Aristotle also speaks. but it is required that it should be inevitably going to happen. he seems to suggest that such propositions of the Old Law as were about the future did not yet have determinate truth. seems to have suggested that such propositions were devoid of truth. we must settle whether this would be knowledge of the future and whether its certainty would be a proof that what I did I did necessarily. he would enquire whether the sun's not rising tornorrow is a describable event. Howrferoften and invariably it was verified.) Then this letter about my actions would not have been knowledge even if what it said was always right. I . and that the events of the future cannot be inferred logically from those of the present.54 The Ancient Greeks Aristotle and the Sea Battle 55 a letter from someone giving an accurate account of my actions and experiences fi-om the time of posting to the time I received the letter. Wittgenstein says that we do not kriow that the sun will rise. but because there may not be any more facts: as in death the world does not change. I think. But he also says that we could not say of a world not going according to law how it would look. he will be great" etc. not the very image of the things themselves". So though he thinks that anything describable can happen.

and Professor tukasiewia's example is misleading because it naturally suggests that in calling bronze matter Aristotle is saying: "At this stage think of 'matter' as if it meant 'kind of stuff'. lamely : If it always will be irnpossible that something has happened. but possible if (iii) is light. And it is not an impclssibility tout court but only a future impossibility that can be derived from the proposition which is contrary to the case. Take something of a kind to happen only once if' it happens at all. is Callias or Socrates. This parallels the Neo-Scholastics! The idea that what changes must be something that doesn't change precisely because it is what changes. rhat he has lost it. Let us take something false. bronze) form =shape. A book on logic by a philosopher Joseph. For Aristotle two bronze statues would also. how the three propositions: (i) What is true and past is necessary (ii) The impossible doesll't follow from the possible (iii) What neither is true. E. are familiar ones. l ) v (i). There is a concealed assumption. then it always is impossible that it will happen. However. impossible. plausible. and in just the same sense. . ( 1) I wish to express grateful admiration for the extreme clarity with which He follows Aristotle in first taking: Matter=material or stuff (e. It may easily be assumed that something impossible at every point of time is simply impossible. th~11 assume its truth and see what follows. but his example of the same shape in different matter is a statue of bronze and a statue of stone. Ofcourse. which. is like this bronze ball. But from the statrmcnt of thr cnre this proposition.' and soon after comes the passage that Professor tukasiewiu quotes: "The whole thing. expounds an argument that there must be an ultimate subject of predication which itself has no predicates. and the concept Aristotle is here concerned with. Callias or Socrates. From prOCee-I ofthe Aristote~ianS&g. Amcornbe's in the same volume. ( 2 ) The absurdity ofthe idea of matter. nor will be. could not serve to explain anything. will always be false and r o . and they are different owing to their matter (for this is different). Professor tukasiewia has written. both occur in Aristotle. The key to this must be found by considering how to use the unexccptionable (ii). who used to be well known.6 The Principle of Individuation . individual. So we frarrle the proposition that he will. but the same in form (for the form is indivisible). Jan tulsiewia'r contribution to the s)rmposium precedes G. M. supplementary volume 9 7 ( 1953). as Professor tukasiewia says. Someone losing his virginity will do. both the concept of 'material' suggested by Professor Lukasiewiu's example. is possible are supposed to be incompatible. The hypothesis that things contain something which isn't anything and has no properties is certainly a senseless one. Now let the cute be that a given pcrson neither is nor ever will be a non-virgin." Aristotle says "This ." These passages show that two bronze balls would be a suitable example of the same shape in different material.g. . it is apparently possible lor him to lose his virginity. Appendix A Note on Diodorus Cronus The question is. such-and-such a form in this flesh and these bones. while 'man' and 'animal' are like 'bronze ball' in general". This principle is highly specious. . (3) I am always uncertain what it means to call a concept "metaphysical". be the same shape in different matter. Doe5 something impossible follow if this proposition is true? It does follow that it will be true that he has lost it. But it is a false assumption. is very like the idea that what has predicates must be something without predicates just because it is what has the predicates: both being based on inadequate reading of Aristotle.

e. for perhaps it is Mel. Certainly I feel only impatient when he considers calling units the matter of numbers. because what specially belongs to substance is being separate and being a 'this something' (rdde rd5: e. you heat it. in the first book of the Physics. for it is necessarily based on the possibility of identifying the same bit of matter in our initial experiments: on our having the idea of 'nothing added and nothing taken away'. "[They think that1 there must be some nature. A. Z. not for the pedantic style.. that if you heat it it will turn into a gas. if you ask me to show you the stuff as it really is in itself. You ' occurs. when it is wine.but that I cannot define the stuff (the bit of water. Nor. To be precise.for that would mean that it could not be. just as 1 understand the information I have about the expansion and contraction of water at different temperatures. ( 6 )Thus I do not think it reasonable to take exception to such statements as that matter is in itself indefinite and unknowable: it "has to be understood in what change^". It will perhaps be more than a pint if1 cool it or less if I heat it.r wine. i o s g a r ~ g o . however. He himself approaches it. out of'which the rest come to be whilc: it re~nainsconstar~t. we say that milk. and afier a while it has turned into vinegar. Let rile will turn into so~nethirlg supposc that you show rrle a bottle of wine. "a pint hereabouts". 'this cabbage'. out ot'one another in the sarne way as bricks out of a house arid a (One gets the point of this only by assuming with him house out ot'brick~.g. can I make anything of such an idea as 'place-matter'. For example. e." He says in thel'hysics. Aristotelian. wine. in the context of discussions which are not alive for us and of most of which it would not be possible to give more than an external account. Sinlilarly there are riot any properties. as I gather that "Ding an sich" may have been by Kant. 11'you show rne a lump of stuff'and tell me that it can be moulded into various shapes.g. it is wine. if I tell you that the stuff in this apparatus has changed from being water to being hydrogen and oxygen. in these extended applications. but is not as such of any specific kind or necessarily possessed of this or that property or dimensions. is so useful an instrument as Aristotle clearly thought it was. I should understand you. you leave it. Physics. This is what I understand Aristotle to be referring to when he Not that matter says that matter is not as such (xaB'adrrjv) so much (~oodv). It is important to understand that this is a conceptual statement. g8gb11:bdec ydp cfval rtva pdarv iJ plaviJ nAclovq j1165.reek. that 1 am talking about) as. 'This matter' is. either qualities or dimensions. in thr course ot' arguing against such philosophers: "Water and air aren't. Professor have to take the whole sentence in which "bAq ~ a eadrrjv" tukasiewia's comments strike me a little as if I were to say "A chair as such isn't upholstered or not upholstered". is white and liquid. but not rl: that is. as I have explained.or lacks-qua this bit ofrnatter. And it is of course not separable: that is. That is. k ( &v yiyvcrat rcf. you will show yourself quite at sea about the sense I am using the word "stuff' in."~ Aristotle however wants to say: 'There isn't anything which i t is all the time. e. iv. and it expands. or I can make it a pint without addition or subtraction of matter. and if you electrity it it else. i88a15: odx d adrbq r p d n o ~ d q nAlv@orPS olulas rat olrla t r nAfv@o~v. Now someone asks "But what is it all the timr?" Sorne Greek philosophers would lrave wanted to say it was water or air or Lire or sor~iething in between. . I . it is designatable. r/)qo~qq. I do not know o r understand enough to have a general opinion whether the concept.."' That is.. we go on saying that it is that element whatever happens to it. That last word was being used in a completely familiar sense."~ that water and air Irnist?l do in fact turn into one another.: not characterizedby a particular answer to the questiorl "How tnuch ?" a pint. 2. .is beside the point in our discussion. with such extraordinary properties. and there isn't some third thing that it is all the time. whereby it defeated the law of excluded middle. ifyou told me that the process of change from wine to vinegar involved expansion or cotltraction.' is: not even extended! . Now 'this matter' is rddc. and of course. Nor can we say that it is as such not a certain kind of'stuff. while it is predicated of the matter. Met.^ The change in question is substantial change: "For the rest (of the predicates) are predicated of the substance. to extend its application away from where it is so to speak indigenous. a . which contrived to be of no kind (to be not 71). lorgar*: r a ~ t vdp v rl)codolac K O T $ Y O P € ~ a( d~ ~I6) q . oUrw bc uat Ubwp ral d @ pC< dAAdAwv rat rlol ual y(vovrar. 'this man'. e. and don't colne. For "matter in itself" does not seem to be used as a name or description by Aristotle. and is vinegar.58 7he Ancient (.) (4) We car1 see now why this matter (e. whether one or more than one.we identify something as a pure sample of an element. you could not entertain producing a specimen of it. But any such tlieory -whatever its validity . so that if.g.) I have approached Aristotle's idea of matter by way of 'this matter'. Met. I t wa.\Aa o w i o p 4 1 .. but for inventing such a strange object as 'a chair as such'. for instance.) ( I t may be that we have a theory ofchernical elements. gg4br6: riJvOArlv#vnrvovflCv~ vorlvdvdy~q.g. and this 6 ' ' Met. And the point about negation is clear here too: I cannot say that this stuffis as such not a pint. and were to be laughed at. q ~ t' make the case simple .\ The Principle oflndividuation 59 But the roncept ol'~riatter which Aristotle works on is at least an everyday one. the stuff1 have got in this bottle) is not as such a given kind of stuff' (ri): for the sarne stufl'was wine arid is vinegar. S o not even the volume determines the bit ofwater that we are talking about. which you can say it has.Aristotle says that matter cannot very well be substance (odoia). ( 5 ) I feel doubtful about Professor tukasiewicz's comments on "matter in itself". identifiable. I understand vet? well what you meati. apart from being the various things it can be. and it is what Aristotle means by "dl$"' (Only he tries to use it by analogy in all sorts of contexts.g.

important to notice that Aristotle's 'matter' is not a hypothesis. thinks that it is a concept reached by mentally stripping away all forms until you get to a characterless substrate. are to a man.) . "This matter" contrasts with "undesignate matter". matter is obvi~us. as every bit of matter has at any time its ow11 prcrper place and is different lrom all the other material things in the world. and if1 ask where and what the stuff that I threw into thc sea is.S*: ~ ~ V C 66 PA nmc *=I ( 611 urpl b t r(( r p ( r [r ~ rkpop&l o r m b v . and we say that it has lost its identity. with change ofplace it is what is now here now there. d o h v d v c u 8 ~ ~ r d r l & r b a r rupn --cfr lo . (Perhaps the total disappearance o f a solid object. I might add. say. or. H .60 The Ancient C r e e k The Principle of Individuation 61 stufl'is ~nilk. Bronze is to a bronze as flesh and bones etc.) Matter only exists as somehow designate. another mass of matter. "two-footed animal") carry a reference to matter in their sense. 1qtbl5: ~ ~ ~ . this is because the particles continue to be identified by some property. etc.sally".g. for whom matter is most difficult. with change of quality what is now healthy. without a ripple in the surroundings except the inrush of air to take its place. And if in such cases there is an answer. And 'this matter' is matter thw designate." This leaves it open whether a given bit of matter.. which at a given time has its own proper place and is diAerent from all the other materi. Cabbage is notjust a kind of stuff. "this spot of light. What could be the point erbo~drlCy.hardly the vaguest notion what we should call a case of it. steels. woods (in bowls e. These queer constructions have escaped their notice. its own proper place and distinctness.). ral vbvptv b n o r ~ i p r v o v 9 ~ r drr 6r nbA~v b'i'norffpcvol' BCrarn ort'pqolv. must be supposed always to have had. ral rat' d l l o l w a ~ v d vbv ptv byrt'~.even side by side with a strong feeling of meaning for the word . but . as obflosrd to 'this milk'. loqsagg: ev n d o a c ~ ydp r a l ~ dvr~rnpt'var~ prrapolaiq Par1 T I m 3 bnonrfpcvov r a i ~ pcrapolaic ofovnard rdnov rd vbvpfv t'vrabea. and matter and/omt. but that we have . or absorbed by. Fur.ndlrv 6' C v ( p e o d . whereas the concept 'gold' does not determine an individual thing in this way. ~. now bigger or smaller. with change ctrfsize what is now so much. must we a prion suppose it to consist of particles retaining their identity? and criticism of any such beliet: Aristotle's view of matter is a rejectio~~ Matter only has identity in so far as it is designate. now in process of destruction. We know 110 application for the idea of annihilation: by which I mean. taken universally". taken univer. There is a contrast between a concept like 'cabbage' and a concept like 'gold'.~ a horse. stone balls. This is what I take Aristotle to mean when he Md. Met. ' Catcprie~. not that we do not know of any case of it.6 vbvp~~~rqA~~dvdf~ ndllv O'QAarroviJ pd{ov. [form] is frightfully diffi~ult. Had Aristotle written in English he would certainly have seized on certain peculiarities of English to make his point: we d o speak of bronzes. "rd A elvar" IA being a dative!l. that is. "It seems to be evident that all these things (bronze statues. ' ( 9 ) "In a way. (I wish Greek grammarians could determine something about the expressions "rd r t & elvaz". now the subject as a 'this something' ant1 now the subject in the way of'privation. If. "man". ral rar'ad<qo11. I d o not understand Aristotle's 'form'. Aristotle's prime examples of 'substance' are: a man ( b rtq d~Oponoq). nor aparl tiom such changes should we have any such concept as 'this stuff'. may easily be puzzled by the expression "this matter. "rd r t 6 rlvai A". marbles.1.Suppose I throw a cupful of milk into the sea. ndltv b'dAAo61. but that is not enougll to secure the permanent identifiability of a once designate bit of it. 1 will consider a passage in Professor Lukasinvicr's paper.d things in the world. a given bit of matter is mixed and fused with. and I do not yet know whether he got clear about it himself. they are marked o u t And yet no one wishes to say that the stuff itself has been dearoyed. I t is no longer this milk. but a cabbage is a particular thing. and especially here. dpolwg 64 wal rnr' oc)o(avd vbvpfv C v yrvt'o~r. 2. Aristotle regarded it rather the other way round: "by form I mean substance without matter":" that is. In all changes between opposed characteristics the subject of change is something: e. if they are not marked out by anything. he says "this matter. earmarked. (8) On the analogy between bronre mrd ils shape. a cabbage. "rd rlvar A". It is always. not with a general notion under which it falls as an instance. 13. In order tc) explain what I think Aristotle means by it. re ' I Met. Similarly with substantial rhange it is what is now in process o f ~ ~ ~ e r a tion. a h q ydp dnopardrq. glasses. Hence when Aristotle wishes to generalize it. with which the Metaphyics is strewn. taken universally"? The "this" seems to war with the "universally".r& r( tjv e t v a ~ Crdotoa ral n-v obtav. now the subject in the way of privation".a. calls matter "now the subject as a 'this something'. in itself'it is indelinite (ctdptoroq). And he fell into frightful difficulties here.' The last phrase is obscure. and always t o be going to have. there is 110need for there to be an answer beyond that it became part of the sea.One which occurs for example in one of the passages quoted by Professor (7) tukasiewicz. we cannot mark them out: ifwe do. you get at it ifyou succeed in thinking matter away from substance.g.ndllv 6 f rdpvov. . b e u u r he thought that form was the 'what' of substances: but of course the names of sensible substances and their definitions (e.) arc individuals owing to their matter. etc."~~ Ross. now sickly. '' Met. But this stuH'rnay be changed trom milk into junket.g. . & lo. irons. (The usual criteria for speaking of the same stuR:)But when this matter loses its identity we do not speakof its being destroyed. Z.. Luckily I need not present my half-fomed ideas about it here. losgaj~: rovrp&Mumcral (bAq."~O (10) "Matter is in a way obvious.


The Ancient Greeks

The Princi le o Individuation

P f


The difficulties that Aristotle gets into come out most clearly if'we consider the following:


A thing and its rd d r f v char are the same1$ (Anti-Platonic). rd rf rfv elva~ dvepdnq and rd dvepdny elvar are clearly equivalent

expressi~ns.~~ (3) dvepwnoc; and d v e p d ~ qclvar are not ttie same unless you make dv8pwnoc;=@uxlf; which is right in one way, wrong in another." It is clearly something special about 'soul' and 'rirclc' that they are the.rame as quxn elvar and K ~ K elvar.I5 A ~ All this is supposed to be resolved16by the consideration that the fbrm and the matter are the same, but one duvdper (in potentiality) and the other Pvepyefq (in actualization). But this is still Greek to me. To translate "rb TI rfv clva~": "the essence" produces gibberish - e.g. "Callias is of himself Callias and the essence of Callias."I7 I t is clear that the KaAAfq in this passage is "man" : "Callias is of correct gloss on rd r i rfv elva~ himself Callias and a man". i.e. Callias is of himself that, to be which is being tbr Callias. Proper names do not, as some philosophers have said, 'have denotation but no connotation'; the criterion of identity for Callias is the criterion for there being the same man as the man we called "Callias". 1 have mentioned so much about fbrm, only because I want to consider the "grave inconsistency" which Professor tukasiewicz ascribes to Aristotle. The inconsistency was this: Aristotle says that individuals are indefinable, but he also says that they consist of matter and form and that whatever has fbrm has a definition. I do not think that Aristotle is in fact at all inconsistent at this point. The individual - say Callias- is indefinable, in the sense that there is no definition of him as opposed to another individual of the same species; his definition is ttie definition of the form. "Of the concrete substance in one sense there is an explanation ad yo^), in another not. For together with the matter there is none(for it is indefinite), but there is one according to first substance: e.g. the explanation (Adyo$ of man is that of the soul.""'But this passage and its context are thick with the difficulty that I have described, of which I do not understand the resolution. Hence the defence against Professor Lukasiewiu's particular charge is not worth much.
( 1 1 ) I have the impression that Professor tukasiewiaequates

taken universally", and "intelligible matter". This seems to be a mistake. "Intelligible matter" has to do merely with mathematical objects: on the analogy of sensible matter Aristotle invents 'intelligible matter' to account for the plurality of geometers' circles (e.g.). For when a geometer speaks of two intersecting circles he is not talking about, say, wooden rings. 'Intelligible matter' is an absurd and useless device, of no importance for Aristotle's account of material substance; and it is not 'intelligible matter' but 'this ' matter taken universally' that is said, together with the definition, to form universals like 'man' and 'horse'.
(12) Luckily it is possible to understand what is meant by calling matter the principle of individuation, without understanding about form. It is not off-hand clear that there has to be a principle of individuation. If X and Y are different, the difference may be made clear by appropriate elucidation of the meaning of "X" and "Y". Consider:
( 1 ) "X

( 2 ) "X

and Y are numbers." - "Which numbers?" and Y are men." - "Which men?"

"this matter,

I S Met. 2, logla17: Erpordv rr yctp odr dAAo 6 o ~ r rfvar i r3c taurob odolac, rat rb T I 411 dvar ACyrrar elvat 4 c ~ d a r o u odola. tnt p8v 64 rBv Acyoptvwv nard ovp/3cpqxdc bd(rrcv dv trcpov rlvat, ofovAcurb~ dvepwnoc Errpov rat16 APVK+ dvep6up, rrval. Met. H , loqgbn: @vp)pPvy&p rat @vXfirfvat r a k d v , d v e p 6 n 61 ~ wal d v e p d n o ~ o d radrdv, rlpi) nal4 @#XI) dvepwnoc Acxetjorrar: oorw 6P rivlplv rivl 6'00. 'I Met. Z,1og6a1:rd yitp ndrArprlvar ral rdrloqnal @vxarfvat ral (buxi)radrd. I s Met. H ad@. I' Mtl. A, lossan6: d KaAAlac ra8'adrbv KaAAlac rat rb rfvJvrfvar KaAAb. '"et. 2, 1 o g 7 a ~ 6 rakqc6C : y'[sc. r$covvdAou odolac)tortnwc Adyoc rat odr t o r t v pcrLpPv y&p ri)~OAqgodr torrv(doptorov ydp), rrarit rev npdrqvd'odolavtorrv, ofovdv@pcJnovd ri)~@uxi)cAdyoc.


Both might be answered by giving a 'definite description'. For example "the even prime"; "the smallest integer, greater than one, that is both a square and a cube"; "the philosopher who drank hemlock"; "the philosopher who wrote the Republic". Before we accept the definite description we have to be satisfied that it applies, and in only one case. But, for ( i 1, what satisfies us shows that a man will be contradicting himself, or talking nonsense, if he says "But still there might be another . . .". For ( 2 ) this is not so. But isn't there pointing? - if, at least, the man is there to be pointed to? Pointing doesn't discriminate: you must know what you are pointing at. If I point, and say "That is X", and point again saying "That is Y", nothing in this situation shows that X and Y are not the same. It is of no use to say "But suppose I point to something different?" - for that is just what is in question: what is something different? It is also of no use to appeal to definition by means of place and time; for this you require points of origin, and for points of origin you have to mention actual objects and events: individuals. No individual is preeminent. If I define an individual X by describing its spatial and temporal relation to another individual Y, and Y has no definition, then my definition of X is infected by the lack of definition of Y. An individual can be defined by pointing and saying what (e.g a man) you are pointing at. But this means that there is no difference between the definition of two individuals of the same species. You cannot say it lies in the difference between two acts of pointing, for nothing prevents one from pointing twice at the same thing; and you cannot say: but the difference is


The Ancient Greeks

The Prinn;Ple of Individuation


that you were pointing at dtferernt things; the different is not tirst merely a different thing, and then, in virtue o f this, a different X. Thus there is no detinition 01' individuals except the definition of their kind. What, then, is the diflerence betweell two individuals ofthe same kind ? It is differenceof'matter; and if I am asked to explai~i that, all 1can do is, e.g., to cut something up and show you the bits. That is what is called rnatcrial difference. This is what is meant by calling matter the principle ofindividuation. To me this truth seems clear and evident.
( 1 3 )The statement that matter is the principle of individuation tloes not mean that the identity of an individual consists in the identity of its mattrl. Thus it is not an objection against it that the matter ofa man's body changes in the course of his life.Ig I don't think that "principle of individuation" is an expression any couriterpart of which is in Aristotle. So far as I know, the statement that acis based only on cording to Aristotle mattel. is the principle of individuatio~l his saying that Callias and Socrates are "different in matter, tor it is different" (sc, in each of them). Clearly what is in cluestion here is contemporaries. There is no q\~estion 01' saying that Professor Popper and Socrates differ materially. But Professor Popper and I, fbr example, differ in rnatter. If I say this, I arn riot saying that Professor Popper is who he is because of the matter of which he is conlposed; so it is not a dificulty fbr me that he is materially in a state of flux. But of course if by "What is the principle of individuation?" you mean, or include, the question "What makes a Inart the same man at different times?" - then the answer "matter" is an absurd one. But as we are talking about Aristotle we have no right to take the question in that second sense at all. And I should say there were two cluite dilt'erent questions here which we ought not to mix up. Aristotle writes very interestingly about nourishment and growth in the De Generatione et corruptione I, 5: "someotle may wonder what it is that g r o \ ~ s Is ? it that to which something is added? For example, if someone grows in the leg, this gets bigger, but not that by means of which he grows, i.e. the fuod. Well, why don't both He goes on to say, isn't it because the substance of the one remains, and of the other not?21i.e. the food turns into the man. Further: ". . . flesh and bone and the rest are twofold, as is everything that has form in matter. For both the matter and the form are called flesh and bone. Thus it can be takcn that every part grows - and grows by the accession of something - in respect
l9 This paper originally ended here. Professor Popper asked me to elaborate this section; but what follows reached him after he had co~npleted his paper. lo f)c Cen. ct Corr. 1, 5, gs~a.go: dnoprfurrr b'dv n< rat ri #art id ad(avdprlSov.ndrcpos 4

of its form, but not in respect of its matter."lP That is, we say that the hand grows, o r thejkh o r the bone. (Think of the ambiguity of the question "Is this the clay you were using last week?" - Aristotle would say that when we speak of 'this (bit 00 clay' the word "clay" refers both to the form and to the matter.) Now matter can be added or taken away, but cannot be said to grow, fbr growth is by addition of matter. Thus it is that we use the term designating the hind of thing, to stand for the subject of growth. And then he adds: "It should be thought of like measuring water by the same measure. For something else keeps on becoming (the thing)."25That is, Aristotle compares the form to, say, the mile that we speak of when we say "this mile of river",24 into which and out ofwhich different water is constantly flowing. I find this a very illuminating comparison. It suggests the following picture to me: let us suppose that we could tag (as medical researchers speak of tagging) wely particle of matter that went into Professor Popper - say by making everything that might go into him radio-active. After a f w yean had gone by wouldn't he be a reach of a stream of radio-active particles? I think of it literally quite pictorially: a stream of silvery particles with Professor Popper's outline drawn somewhere in the middle of it. Of course we mark 'this mile of river' by landmarks, as water does not change on entering and leaving it. But food and so on change substantially when they get into Professor Popper, so his form (the flesh and bone of a living man, to put it roughly) does the marking off; and corresponds to the mile of river. I think this demonstrates quite clearly that if you mean anything Aristotelian by calling matter the principle of individuation, you d o not mean that the identity of a person is the identity of the matter of which he is composed.
Ihid. b ~ g odpt : ral dorobv ral Craorov r 8 v ro~odrcuvpoplcuv tor1 brrrdv, donep rat r 8 v d l l r v r&v t v OAn efboc txdvrcuv. ral ydp OAI) Atprar ral r6 eldo(. o Q < ral dorobv. r6 o h brrobv ptpoc ad<dvco8ar ral u p o o r d v r o ~ n v 6 xardpf ~ v 16 elddc torlv tvbe~dpevov, xard br rt)v Olqv our Oorrv. '3 Ibid. 04: bel ydp vo#oar donep eFIrrc perpolq T @ a h @ p6rpq~bbrp* drl y Q ddlo ral dldo 76 yrvdprvov.
I am indebted for this interpretation to Mr P. Geach, who threw it out almost as a joke in casual conversation: but I think it is obviously correct. 1 am grateful to him also for 0th- help in preparing this paper.



npoarbrral T I oiov , el rev rvtjpqv ad{&vc~, adrq p r l i i v , 4 br abldvrt, t j rpoptj, oc. bld ri 63 08s odn ~PVW quCqra1; 2 1 Ibid. 94: 4 d r ~ r o t p & v ~ b vfieo ~~ o i a roc . b'ob, ufovifiqrpop4~;

Thought and Action in Aristotle



Thought and Action in Aristotle
What is 'Practical Truth'?

Is Aristotle iricor~sisteritin the different things he says about npoarprorq, mostly translated "choice", in the diff'erentparts of the Ethicn ?The following seems to be a striking inconsistertcy. In Book 111 ( 11 13a4)he says that what is "tlecidcd by deliberation" is chosen (TO *K rqc pouAqq ~preevnpoarprrov torrrd, but he also olten insists that the uncontrolled man, the ct~parqq, does not choose t o d o what he does; that is to say, what he does in doing the kind of thing thar he disapproves of', is not what Aristotle will call exercising choict; the uncontrolled rriari does not act from choice, *K npoarprurwq, or clloosing, npoatpovprvoc. However, in Book V 1 I I 14zbi8) he mentions the possibility of a calculati~lg uncontrolleci man who will get what he arrived at by calculation, (K ror Aoyropou reu{rrar, and so will have deliberated correctly: dpflwc tornr firpovAt.vp~voc. Thus we have the three theses: ( 1 ) choice is what is tletermiried by deliberation; ( 2 ) what the uncontrolled man does qua uncontrolled, he does not choose to do; (3)the uncontrolled man. even when acting against his convictions, does on occasion determine what i o do hy deliberation. Without a doubt the set ol'passages is inconsistent if we are to understand that any case ot'sotnething being determined by deliberation at all is a c a r of choice, as seems to be suggested by the forn~ulation"what is decided by deliberation is chosen". IS, then, Aristotle is consistent, perhaps his 'choice' is not simply determination by cnlculating or deliberating. There is some reason to think this; though he says that what is detennined by deliberation ( ~ ~ 1 0 (K t.v rqqfiovlqq) is chosen, we may say that the context shows that he himself has in mind a deliberation what to do with a view to one's ends, and that ends are things like being honoured, health, the life of virtue, or material prosperity, or enjoyment of' knowledge, or sensual pleasure. The uncontrolled man, the drpnrqq, is not one whose general object is, say, enjoying a life of sensual pleasure; he simply has the particular purpose of seducing his neighbour's wife. On this view, we remove the inconsistency by saying that 'choice' is of something determined not just by any deliberation, but by deliberation how to obtain an object of one's will ((bouAqos) rather than merely of one's desire (tnreupra):there will be a contrast here even for the d ~ o A a o r the o ~ licentious man. For his will is to satisJy his desires, his sensual appetites; and his decision to


seduce his neighbour's wife, say, is a 'choice', as well as being an expression of his lusts, just because his end in life is to satisfy his lusts; this has to be shown before one can say that a man who is going after objects of 'desire' evilly, has a bad 'choice'. Now - though I think this does represent Aristotle's view - an objection that strikes one is that people's 'ends' aren't in general nearly as definitely one thing or another as Aristotle makes out. If'will' (boulqorc)is simply the ' type of wanting (dpe{r~) that one has in relation to one's final objective in what one is deliberately doing at any time, then there seems no objection to saying that the weak man at 1151az (the uncontrolled man who calculates how toget what tempts him, for he is surelya man of the weak rather than the impulsive type) has a will to seduce his neighbour's wife, or a will for the pleasure of it, at the time when he is cleverly reckoning how to do it. The fact that he has a bad conscience about it doesn't seem to be either here or there for determining whethtr he is making that his aim for the time being; but this fact, that he has a bad conscience about it, is just what makes him uncontrolled er d~olaoro~. rather than licentious, d ~ p a r ~ r a t hthan There is, however, another defence against the charge of inconsistency, which perhaps is not open to the objection that it requires an unrealistic idea of the clearcutness of people's ends. Not all deliberation is with a view to making a 'choice', forming a npoarpeorq, where none has yet been made; some deliberation is with a view to executing a 'choice'. This is made clear at I 144aso; "Virtue makes one's choice right, but as for what has to be done for the sake of it, that doesn't belong to virtue but to another power cleverness." (rqv pev odv npoarpcorv dpeqv xorer tj dperq. ro b'doa C ? K € I V ~ C 4vc~n a e g u ~ nparreoear e O ~ t' K orr rqc d p e r q ~ dAA'f rcpaq bvvape~q.) But also in Book I11 Aristotle speaks of trying to do the thing that a deliberation has terminated in: "if it seems possible, they try to do it. Possible things are the things that might come about through us" ( I I I 9b96). So we might say that something that seems to be a way of achieving your end and to be possible may be decided upon; that you will do this (or at least will try) is a 'choice'; and now there may be further deliberation just how to manage that possible-sekming thing. Now in Book 111 there is no suggestion that wanting (dpeQd of the more immediate means (adopted to execute the remoter means that have already been decided on) is not itself also a 'choice', npoarpeorc. But if we are to reconcile the denial (which also occurs in Book 111. I I I 1b14) that the uncontrolled man in acting, is choosing so to act (npoatpou~evod with the account in Book V I of a calculating uncontrolled man, then we must say that when deliberation how to execute a decision terminates in an action - the man contrives a skilful approach to the woman this will not be a case of 'choice' if the decision itself was not reached by deliberation. ~ h uthe s passages in which Aristotle describes deliberation as going on till we have reached something we can do here and now, and describes 'choice' as being of what deliberation has reached, must m t lead US to think that

From J. R. Bamhrough (ed.),N e w Essay on Plato and Aristotle (London, 1965).

A little tsrther on. But it. an dxoAaoros the decision to seduce her would have been a 'choice'. But although the uncontrolled man perhaps reckons how to proceed . ~ h u the s second defence resolves into the first. thcn he calculates how brst to do this arid shows plenty of cleverness in his calculations. we may say. surely. has an dpc41q for them. .e.once he has given way to the temptation to go after this woman . does not set up a word for 'volition' as I have been using it.deliberations horn an end to the immediate thing. so this delence leaves us in the dark as to what a 'choice' is. But Book VI teaches us. better than deliberation aliich is 711 cir ( I I I ' about the nlearls tiere and now to 'living well in general' . of course. are not 'choices'. Aristotle ought.tlid secrn a relatively clear notion. To return to the weak. that there is no such thing as a 'choice' whicfi is only tcch~iical( I use. . starts with the word "bto" . The second to =-ling a defence was that since some deliberation is d o n with a 'choice'. absolutely speaking. but something one d w . however much calculation may have gone into detemining it. he says he will get what he 'proposes' (nporterrat). odt'dvcu vou rrar btavotaq odt'dveu r)et~qq&ortv &&wq r) npoatproy. 'will'. "rrpoatprut~" ("choice"). arirl ill partir:ular to lit technicdl deliberation. is not anything one makes. (. does not exist without judgement and character. However. But this is a piece of'teclinical deliberation. and that is the object of the wanting (4 bldpcOs rourou). and this verb expresses a volition. The calculating uncontrolled man choosing means of seduction he wants them.xalways. MY yet without moral character". then just wliertt in the chain of. "For doing well. or perhaps rather an intention. or1 Aristode's view. but he did not do so. if it is of what is only a means to the objects of a man's hteuptat. I think. however. or poudv) often seems to lit deliberation about how to execute a decision. namely. For the decision to seduce this woman was sirnply the particular applicatio~l of his general policy ot' pursuing sensual enjoyment. When he describes this man as calculating cleverly. I t scerris at its clearest whet1 Ire is describing the doctor deliberating how to restore health by reducing the imbalance of humours by .68 ]'he Ancient Creeks Thought and Action in Arijtotle 69 rnatter lor a 'choice' has only been rcached when there is no more room l0r deliberation of any kind. even when it's not strictly a technique that's in question). It is puzzling. - . to the technical or executive decision.. I am rlot saying that Aristotle so u s o " n p o n l p m r ~("rlroicc") that the tmninatioa ol'a piece ol'technical deliberation isn't a 'cl~oice'. to be. there is room for calculating how to execute a 'choice'. which you have reached as a result ol'delibcrating how to achieve an erid' -tire lirst Cduse (npw~ov ni~tov). in fact. On the other hand. the succeeding sentence starts "For" so perhaps we should look for the explanation there first. another 'cl~uice'bel~ir~d autive 011J r 13gbr-3).On the contrary. Let us return to the point that a technical 'choice' is never the only 'choice' that is made by the man who makes it. However. etc. correspond to "npoatpctov" ("chosen"). and this is a result of deliberation. uncontrolled man. We rnay well have thought we knew this. If he had been a licentious man. this. and if this has nd - 6I - - - - ". the innocent unnoticeable verb he uses receives no attention fiom him. his volitions in performing the steps that l ~ calculates e will enable him to succeed ." (cdnpatta yap xat ro (vavrtov (v rpa&t dvcv btavotaq xat (eouq odx (urtv. if that particular thing for the sake of which this drcisiosr is being made is not itseydecided upon by deliberation. just as the first defence lelt us wonderitig what Aristotlc supposerl a ~ O U ~a O casc I ~ of. Of course he regards the uncontrolled man as acting voluntarily. that would. . . situatiorls about. his 'desires'. be quite incoesiste~it with the treatlnerit in Rook 111. The definition of 'choice' as dprfic pouArurtxq deliberative wanting would not at first sight seem to justify exactly the same way as the licentious man. who disapproves of' adultery hut is ternpred about his neighhour's wife: be gives way to the temptation anti sets out to seduce her. then unless his 'will' in life L to satisfy these desires (as holds of the licentious man) it is not a 'choice'. i. calcula~ing.rIeliberation. He says choice does not exist without intellect and judgement.) That sentence.a puzzling remark in that passage in Book VI (1lggai7-big) where Aristotle devotes most discussion to this definition of 'choice'. something may be reached as a result of deliberation even when the significant decision what to d o has already b n n made. however. because while the previous sentences give ample grounds for saying that choice involves intelligence. (Aristotle. to have seen that he was here employing a key concept in the theory of action. that I can (lo without having to co~isider how to do it -just where in this chair1 tloes the Iirst 'choice' come? I t must be ad~rlitted that Aristotle's account oftleliberation (poul~ootq. "choice". . "tec:hnical" to cover practical cleverrlcss in brirlging particular." his brings us back to our first defence. the last thing in analysis and first in cxer:utiu~i .. ever1 thougli this is the fruit ol."That is why". That is why choice is appetitive (dperrrtxq)intelligence or intelligent wanting.rrpo~ dAoc. is something for which Aristotle has no regular name fbr he has no general use of a psychological verb or abstract noun corresponding to 'I CKOUU~OV" (usually translated "voluntary") as nPoatpctoeat" ("choose").) That does not seem to help us much. that something is only a 'choice' if it is of means to the objects of a man's 'will' (poulqu~). and its opposite.) So we have to say that the uncontrolled man carries out a deliberation how to execute what would have been a 'choice' if he had been an ctxoAaoro~. as I t l i i ~ ~ we k might not have realizetl ti-om Book 111. Tliat is why lie denies the narne of "nponrpcorc". hence. as must be adn~itted on the basis ofttir text. since appalc~~tly the plrasure u)ugllt by the u~rontrolled nian who calculates is not all object ol' his will. for 'what you can do here and now. For doing well is the end. each of the steps that he reckoned would enable and the volition to perfi~r~n him to succeed would in turn each have been a 'choice' too. he tells us "The end. . they don't seem to give any ground for saying that it involves moral character. There is a technical O I purely c. there is what may give us pause in making this criticism .

but not done them under any such conception? under what conception. say. for him.g. or attract4 by the particular - - - - . Now 1 have described the cases so that the men's ends are clear. The other case is diffmnt. then still more one would want plenty of actions performed under the influence of his new thoughts before one could recognize one as done with a view to this sort of 'doing well' rather than as. how is this? Let us imagine some cases. This one act with a view to this sort of 'doing well' -what is supposed to have preceded it? Has he done things of the same sort.70 The Ancient Creeks Thought and Action in Arktotle gains of each action? Very well. an experiment in wrongdoing. i. 'choice' cannot do all the work Aristotle m u to make it do. then. Well. These are two rather different types of case. 'a good way of carrying on' is the end of any 'choice'. If the agent had never done any scientificresearch or study at all. does not qualify to be a 'choice'. The notion of 'choice' as conceived by Aristotle. although the deliberation is in the interests of a desire which conflicts with what he regards as doing well to d n o i b e his action we need a concept (our 'intention') having to d o with will o r been made by deliberation (~preev f K rqc poulqC). suppose he had them onjust one occasion. but what is to make us call this the first act done with a view to that s o n of 'doing well'? It is not enough for the agent to have those thoughts. The uncontmlled man who has funhm intentions in doing what he does. taking them in. then. no doubt with some piece of imagination attached as if.e. his % O K . Only if they are the thoughts which come to habitually inspire those actions shall we be able to say: that is his end.e. this act would not mean that he was a fraudulent man i. i." Whereas.say in obedience to a mentor. 'choice' is only of those things which are done as means to 'doing well'. But there is no reason to say that the action which is the subject of 'choice' must imlf be the act of a vinue or a vice. whose actions arc dclibnrte. If. a character exists only when there is an habitual performance of the typical acts of that character. then? . e. Another case: someone thinks that he will do well if he spends his life in scientitic research. in both of them it would be natural enough to say that the man is described as having a sort of moral character. and the results of deliberations how to execute it won't be 'choices' either. that is his idea of a good way of going on. clearly is that there is no such thing as your acting with rdnpaoa. or it would be a description of someone under a fantastic illusion. the ctirolaoroq: a life spent doing such things is his idea of a well-spent life-and a fig for moral virtue. beyond saying that 'doing well'. His thesis.that would o as to 'do well' in that kind ofway. and always gets the worst of things. whereas the honest rnan is weak and soft and a fool. to get the money fbr his living expenses he does a disgraceful but not time-consuming thing: one great fraud. drr ro lrapov dbu drw~erv. defrauding them. in view unless you have some sort of moral character. without good or bad habitual action. we may concede that Aristotle is right in saying that it d m not occur without moral character. but I have put in only one act for each. say).word as familiar to us as "practical" is? At any Ate. why can't one have 'choice' without moral character of some sort? 1 think Aristotle does not explain this. the question whether the significant decision is reached by deliberation seems to reduce to the question whether it is made with a view to the objects of the man's 'will' ( ~ o u l ~ o l s Now l . In the second case I described. Someone thinks that it is a good sort of life always to get the better of people by tricking thern. to do that is to be strong and not soft and not a sucker oneself. our question about this was: what does Aristotle suppose 'will' (poulqor~) to be? Why. and to get the best ofwhatever's going.fraudulent. even in a feeble and elementary fashion. what he thinks is rattier that this is a good way to carry on. I used to think it spurious. "One should pursue the doesn't mean: it's virtuous. It is not that the licentious man thinks licentiousness is moral virtue. morally obligatory. then the description of the case would be suspect. this is the kind of life I want. or present pleasure". I'k i e some other Aristotelian concepts. any sort of'decision which does not have in view what one thinks of as a good way ofproceeding in one's life. that is the attitude of the licentious man. of course. Perhaps it is possible to conceive something as the activity you aim to spend your life at even though you never d o it at all. is a very peculiar one. If it had been a winner. On Aristotle's view. to do this he must have leisure. that he had the vice of being . here the single act which is to be the object of a choice is not the kind of act which the agent supposes to be the way to spend his life well. But then either it would have to be something you could understand without doing it (like riding horses. and had no other conception of a good way of spending his life: that was it. Now. 'doing well'. The first case is not credibly described on the supposition that there is only one such act. shouldn't we say that the uncontrolled Illan has a 'will' for the pleasurc he hopes to obtain fiom seducing his neighbour's wife? The answer we get suggested by the passage in Book V1 is: the uncontrolled rnan is not prepared to say: "This is my idea of good work (&ulrpahu).then it was not a 'choice'. or you could only want the name. This would rather be a lunatic obsession than a conception of a certain sort of doing well as the end. he had not done any actions of the a a t before. That will only be so where the objects of 'choice' are (in Greenwood's phrase) constitutive means towards the (~utative) good way of going on. If. only that he not show that he was acting s had indulged in a certain picture of his actions. we asked. someone who had never learnt any mathematics wanted to become a mathematician because of the expression on the face of a mathematician he knew. the fraudulent act was a productive means. however. to do that but: that's the thing to do! Now. would not "prohametic" be a. A particular decision to cheat X will be a 'choice' of something here and now which he makes for the sake of doing well as he conceives it. virtuous or vicious.. and if the man did not perform o t h a fraudulent acts. on the other hand. Either it would be nonsense.e.

Aristotle talks as if' 'desire' were a force ( 1 147a34). Aristotle explicitly wants to exempt knowledge of general principles from 'being dragged about like a slave'. since one does not need to advert to the obvious) its formal character beconles quite clear You have a set of'premises starting with a universal one to the effect that a kind of thing A is. without knowledge o f what he is saying. One hils to know or remember either the last prrlnise or. when mding his point. say.r napalrArlarw ourdarvciv). "They work just the same". There are two features suggesting the necessity of the conclusion. the conclusion. then it is clear that the conclusioll of this reawning is for you to do E. first such and such is needed. dvavrrq rode npwrov. it would be queer of me not to walk. that a cloak must be made.the gerundive form. whom I have picked up at a party. But. by the man who is about to d o it. However. but i t y just ~ babble like a drunk man reciting ~ l ~ ~ e d o c lto es the " particular premise: or possibly to that and the conclusion. is someone's wife" -his failure to find out being explained by his passion. say: "No one should commit adultery" or "lt is disgraafu\ to get veq drunk". and I am a man". but this is only a ~netapllor. There are such cases.75' The Ancient Greeks appetition: not just (ntbuwa. I mean that it is a proof of the conclusion. and say "I have got to walk". on anotller. is your concern is) a B. possibly. ~t least he thought you must d o E unless rnmething prevented you . and fur lies m e may tell inwardly or outwardly when one wants to do wrong.g. Further. "Every man must walk. It is. is an action: ro oupnepaopa ro iparrov aoi~reov npa(l~ t'ort. So here is a 'choice'. and the type of universality in the premise. so he would not have agreed to say what I have just said. but then fails to draw the conclusiorl. Nicomochean Ethics. that he was still under Plato's influence. ~ u let t s consider what this means. particularly when he restricts the explanation "he repeats the thing. if only the premises are true.He will have it that if one a c u agai11st one's convictions. surely. ~fit is all made explicit (as of course it hardly would be in real life. But it is possible that if he had b m challenged about this. and if such and such.i~ I need a cloak. so and so" (el b a a o v (oral. and I am a man". e. an explanation fir better suited to enunciation of the universal premise. I must make a cloak. for that may be only a feeling. o r forced to walk in the other. The man does the thing in question (walks or halts) at once. 'desire'. if nothing prwented me. Now let us look at another example from the Movement ofAnimoLc: I need a covering. For he is marked by an anxiev to make practical reasoning out to be as like u possible to spmulatiw ~ o n i n g . The usual exphnations of this are that Aristotle was a Greek. a covering. speaking quite seriously. and so may commit a d u l t e ~ through culpable ignorance of a particular premise: '7his woman. ~rofitable for a kind ofbeing 8 . I I I 1b16). action has a starting-point ."It Every man has got to walk I am a man I have got to walk is a formally valid deductive argument. "Now no man must walk. Now Aristotle had special ideas about proof. On one ~nterprctation the trouble always concerns one of the particular premises. so he would have said this is not apodeictic (see. he would have said one could amend the syllogism by putting in that a cloak was the best covering or the easint to make o r something of r)ut sort (cf. &n0up1a. A c1oak.and so he sketches the reasoning with a view to execution of the 'choice': "If there's to be a cloak. and this is sweet" (1147asg).the might be the drive of 'desire'. and seems to be referring to a nmessi~tion of the . Disregarding this let us merely note the formal validity of the reasoning as a deduction. I must make what I need. he often gave e m m p l e ~ of II . and proceding through intermediate premises like "Cs are D" and "a procedure D can be and "a c can be obtained by a with another premise to the effect that you carried out by doing E". at nlost he draws it verbally. The fact that it is not. But Aristotlr writes as if these were the only cases of'doing what you believe is wrong. he goes on. and if the action E is are (or someone something that you a n do. one's judgement has always failed in some way under the inRucnce of 'desire' or some other passion. Aristotle says. For example. with their ilnplications braction in view ofone's ends. I 14oagg-5). etc. I need a cloak. Now it is hard to tell whether Aristotle reflected that "I need a cloak" is not a formally valid deductive conclusion from "I need a covering and a cloak is a covering". let us grant that if I agree to the premises and therefore to the conclusion. And similarly f i r failure to get or keep clear before one's mind already known b a s . rode). I suspect that he was also influenced by his own conception of' practical reasoning. ef de rode.. T h o q k and Action in Aristotle 73 practical syllogisms in which there is a certain necessity about the conclusion is necessary to taste everything sweet. Chapter VII. He clearapparently cannot admit the caw where a PmOn f'orrns a headed intention of acting contrary to his convictions. To set out the form of practical reasoning is to set out the form of'deliberation (pov~eoorC). and this last the man does at once. The conclusion.against doing 8.I will call such an argument a proofsyllogism. no criticism of the $logism as a piece of practical reasoning. No dot~bt there is something in that. Aristotle allows a case where the sinner is clear about ail these. "Every man has got to walk" is not a changeless truth. The last two examples come from the Movrmcnl of A n i d . if not prwented from walking in the one case. is. a Inan who disapproves of'adulterymay fail to find o u t something which he easily could have found our. . he says in the Mavtnntl o f A n i d (to1. Does it mean that if you have embarked on the msonirlg you m u d o E? ~risrotle seems to have thought so. 1 should contend.

He coins a word for what we should do. The need for necessity in the connections can fairly be discounted. Here we touch on the difference between Aristotle and Hume.then it wiU be p r a c t i d (St Thomas would call it "theoretical depructicis" (Summer T h e * . For blessedness. at all (6. But you do not get this where various ways of obtaining the end are possible. and second there is the idea of the thing wanted as the starting-point for such thought. which he did not entirely shake off). and given all the premises it is a formal matter what the conclusion is. This is practical reasoning.Whether. especially intellectual virtue. Aristotle's 'will' will then be a 'calm passion' in Hume's terminology. but also a compulsiveness about ' the universal premise.). on the contrary. it seems he wanted a universal premise acceptance of which implies intellectual acknowledgement of it as the guide to action. then it is itself acknowledged as the ultimate guide to action. the whole business is what you want (the dpxv is the dprxrov) can come into play. Aristotle's grand universal premise is that blessedness is activity in accordance with virtue. He wants a "must" in the coriclusiorl in the verbalized form in which he gives it in the Movement of Animals. And we may remark that there are two possible conclusions of the reasoning about heavy waters. If the truth of this premise is acknowledged. the fact that though it is perfectly correct to call practical reasoning 'reasoning'. We have seen two strands in Aristotle's thought. depends on whether 1 actively aim at being rich and am working out this one of the many possibilities with a view to action I might be doing it idly. merely the object of a passion. If one admits that what one wants is no good. It looks as if. is happiness or blessedness.taking the side of and imitating the immortal. art. it is the object of will as the best possible thing for a human being. la.e.which he is keen on because he thinks it helps to make it clear how the syllogism KIVCI. For as seeing can be seen to be what the eye is for. # n ~ @ g ~ For a. ofcourse. - . the kind ofwanting De Anima. qgnbg-7).as Aristotle did (De Anima. and though some practical syllogisms are also (in my sense) proof syllogisms. I draw theconclusion. If by a practical syllogism you mean .e. how it sets the human animal in motion . First there is the explanation of how the human being is set in motion by thought. Hume's doctrine that reason is inert. is required may be compared to Aristotle's "It is not reason as such that sets in motion. i. though what constitutes blessedness is necessarily utterly pleasant. o r blessedness. "It's expedient". in Aristotle's conception. In general. a that terminates in action. Aristotle recognizes this two-way possibility at Metaphysio IX. Consider: - Owning a Launderette would make me wealthy. that for considerations to lead to any action a sentiment. Then we can happily combine the two strands by ~ostulating at the back of all these premises a first premise to the effect that only such-andsuch is doing well. but if you mean a t ~ e of reasoning . and it is easy to spell out the practical reasoning of the suicide.then we have such a universal premise as "Heavy waters are unwholesome". are entailments. But here we must note a certain split in Aristotle's thought.'. is the end that anyone must have so far as he has a rational end.74 The Ancient Greeks Thought and Action in Aristotle 75 conclusion. I suggest that the idea of rational wanting should be explained in terms of what is wanted being wanted qua conducive to or part of 'doing well'. though each time he gives the conclusion he adds . by his own picture of proof and by the Platonic conception of sin as error. There is scope for opening a Launderette in such-and-such a place and so on down to where I might get going. but reason which is with a View to something and is practical" ( 1 iggag6). 1oq6bg-8. Here the De Anima formulation (of a doctrine also expressed in the Nicomachean Ethic$ at 1 140bi6. which we should grasp at to the poor o r v n t that we can . sounds absurd.can be seen to be what the mind is for. .i. it is. no doubt. in the form 'so 1'11. that belongs in the rational part. "such and such a kind of being ought to do such and such a kind of thing". namely "to immortalize" (&8avaa<~~v 1177b~g). 14. a 'must' about it: that is. qggalg) . The argument for this as the true premise is the Nicomochean Ethics itself. but let the universal be "Strong alkalis are deadly poison". or as an academic example. not a means.e. his discovery. That. being the actualization of his rational part and an actualization that is an end. though it is no good (like smoking in some people's view) then one is being led simply by 'desire'. Rut when he is not talking about this automatic-machine aspect of the practical syllogism . But they disagree about the applicability of the descriptions "in accordance with reason" and "not in accordance with reason" to actions and wants. reasoning reaching from a general sort of objective to something one can choose to d o here and now . that is to say so far as he has 'will'."and that's an action". so understanding the enjoyment of the truth . according as you want to be healthy or not. if it is I who have put out thesyllogism. though not so clearly) that the startingpoint of. - 16c. when the thing that one so wants is a pleasure. but still onewants it. For the first he seems to have wanted not only a necessity in the connections which is not always present in practical reasonings. 'the good for man'. it isn't something one wants because it is a pleasure wen though it should be no good. or doing well. and the purpose ofwhich is to act. . Aristotle did not notice some significant features of. i. people would not trouble to work such things out except with a view to action. then this won't be practical. For the highest blessednesshe thought of as something divine. in his enthusiasm for making practical reasoning like theoretical and explaining its power to set one in motion (aided. in general practical syllogisms have a different form from proof syllogisms. A further sign is that when he is lookingat practical syllogism in this lightas necessarily yielding the conclusion his examples of the first universal premises always go "It's needed".

that being his judgement ofwhat it is just to do. even if it is no good) Apart liom being rr~led 'doing well' is what arlyonc wants in some obscure and indeterminate way. My eventual goal has been to expourld the concept of 'practical truth' and the discussion of Nicomachean Ethics. Chapter I1 on 'choice'. it is clearly that 'truth in agreetruth is not the truth of t h o s e p d g ~ sFor ment with right desire' (dAqf3aa dpoAoyoc fxoooa TU dpeCri T n dpf3n) (1 iggago). without the wanting being wrong if they are in accord? Suppose the man has judged truly. sound or unsound. "So. but of judgements about what sort of procedures are just. "since rrloral virtue is a disposition of one's choice. A false judgernent on this necessarily means tltat if there is a 'choice' at all the wanting in it is wrong. 4 t v t n npoaipeoel dyvora. If I am right there is philosophy to the contrary in Aristotle. Can the judgement be false at a lower level than one's idea of doing well. the idea of descriptions undcr which what is done is voluntary is integral to his notion of action (proxir).i. I am not speaking of particular procedures. as I should maintain. S. Suppose." Aristotle goes on. how should these be gods?" . He will secure. And if. then. that ~11t:ydefinitely so count being rich or being farnous or the lik of knowleclge. There is the "yes" in judgement and the "yes" in the will.76 l'he Ancient Creeh Thought and Action in Aristotle 77 sounding like an echo of "to hledize" which means to be 011 the side o f and imitate the Persian. or questions as to facts which the agent can't be expected to have found out. are in desire. actior~ in accordance with moral virtue. or the work (kpyov). as Aristotle would say and as I want to say. and not merely by an extension and in a way that ought to be explained away. pooAqo~c. of practical judgement. . and the one must say and the other pursue the same thing. meaning that one wants to do that sort ofthing. It appears to me that only when we get to questions where it is difficult to know the truth. if' the choice is to be sound. ifonejudges ~. provided that a man forms and executes a good 'choice'. cutncss is the assu~tlptionthat people generally know what they count as 'doing well' . To make this clear. That is brought about . doing well. This then will be why Aristotle said in Book 111 ( I i iobgi) that ignorance in choice. must involve wrong wanting. tile heart ofwhich is. say. is the cause not of involuntariness but of scoundrelism. then these predicates apply to actions (praxcis) strictly and properly. then. but not vice versa. one can say "yes" or "no" both pursuit ant1 avoida~tce to a statement and to a proposal. 011ecould call it that part oi'blessedness for which one's own action is essential. The man who forms and executes an evil 'choice' will also make true Jome description ofwhat he does. but falsely that justice would be done by dividing all the goods available in the country into equal shares according to the number of the population and assigning each share to one person by picking name and number of share out of a hat. that the statement should say that doing such and such is 'doing well'. It is practical truth when the judgements involved in the formation of the 'choice' leading to the action are all true. For to characterize it as 'doing well' is eo ipso to propose it as an object of"wil1' to put it up as a candidate for 'will'. Book VI. Aristotle's unrealistic conception of the clearcutness of' people's ends seems on investigation not to be so bad as it looked. that to act justly is necessary for doing well. We now approach the great question: what does Aristotle mean by "practical truth"? He calls it the good working. is there any chance for the wanting of what isjudged a means to doing well to be right when thejudgement itself is wrong.J far we have only mentioned the judgement on what elinpaoa. . while choice is deliberated wanting.c. For the many objectives The assu~nption of'clearthat are no goorl are allowed fbr in his tl~ought. that such and such a man has his eyes put out or his hands cut off. since in 'choice' the wanting goes after what the judgement declares to be doing well.. will have produced practical falsehood." We may remark that the one must say and the otlter pursue the same thing if there is to be any 'choice' at all. If the worldly man has any wants that are right. terminating in action. these things show that the judgcrncnt must be true and the wanting right. made true . I will start li-or11 I 13gaz1. is. He himself laid down the rule about difficulty at I i igbgg-i i iqax. or that it is justice for a poor man to be punished for assaulting a rich one. but the practical . But he ackrlowledges that in the ordinary course of life fi)r rnost people 'doing well' arnounts to something more mundane: a successhl and Itonourable conduct of life. He. if he is competent.e. Any 'cltoice' that he makes. of practical intelligence. imagine a worldling's idea of doing well. But his description "justice performed" ofwhat he has done will be a lie." That action (since the description of what he does is made true by his doing it). "Since everything that is done about them is false. by passion his is what I want. .ig)~tly. which is spoken of as the good working ( d l . "What affirnlation and negation are in judgement. or the work.i. they don't occur in his 'choices'. and practical judgement isjudgement of the kind described.The notion of truth n fdsehood in uction would quite generally be countered by d e objection that "true" and "false" are wnseleu predicates as applied to what is done.

Part Two Medieval and Modern Philosophers .

For that is what our argument was.necessarily not knowledge. Many might want to accept this so far as concerns human knowledge.St Thomas is indeed partly caught in the trap of argument with which we opened. and from this nothing follows placing any restriction on the objects of knowledge. i A revised version ofthe* . must it not always be capable of turning out false. and Aristotle's doctrine that the sphere of knowledge was only: the kind of thing that cannot be otherwise. or God does not know all that is to come ( la. as we did. was observed rocking on the floor in a discussion and =claiming "God doesn't - - - - . the former is correct only in the sense that if something is not true then my certainty that it is the case is . this thing may not happen.8 Necessity and Truth What is known must be true. hence it readily appears that only the necesszrily trrie can be known. He escapes from it without fully realizing this . . One is likely to learn this criticism in connection with Plato's doctrine that only supra-sensible forms were objects of knowledge properly so-called.ppeuedin the ~ k w f r b q ~upplnm~ (a4 k b m q1 ~ 6 ~ ) . being contingent. The compatibility between the contingency (non-necessity)of these known res and the necessity& dido "What is known is true" is the right way out of the trap . assumes the position as it were of a pure sample in a thought-experiment. which thus. This is probably one root of the Greek ' conception that knowledge is of the changelessly true. 14. Tncth rmd an appeal to the distinction between necessity de re and necessity de dicto. the things that are known. . though the point may be obscured by the fact that the res in question. But one goes straight back into it like this: must not the fact of the knowledge of such a contingency itself be equally contingent? That is to say. such as that so-and-so will win the next election (known only to God) or that I shall ndJCnd thatpot ofcofee a t q elbow too h d to drink (known well to me). that suchand-such a way for the future to turn out is known to be the way it will turn out? For. in spite of Language. (As when Professor Ayer in his youth.ofa set of sixty of a new translation of the Summa TheologMc (sometimes called the Summa Theologica) of St Thomas Aquinas prompts one to ponder St Thomas's attitude to this question. 13). Nowadays an undergraduate early learns to criticize the passage from "What is known is necessarily true" to "Only the necessarily true is known". The appearance of the first volumes. as so often. and so hold that there is after all no escape from the trap. appear themselves as dicta. St Thomas was stopped from accepting it for divine knowledge. art. We find it presented in him in a heightened form: must not God's knowledge be only of what is necessarily m e ? so that either there is no contingency about the future.

the alleged knowledge is not knowledge!" This is again the requirement that the object of knowledge be necessary. g. None of this shows me not to know now. the whole passage is in any case immensely interesting.c):in the same way. "But the reasons lor present knowledge must be present reasons. so that the elfects are contingent. it'on examination the coffee turns out hot. And this can be applied to human knowledge no less than to divine. But. indemonstrable principles which are known of themselves. in the field of speculative reason. Our own knowledge ol' future events is to our k~~owlcdgc only in their present causes. These dismissals are intriguingly obscure. prove me riot to know. in the case where 1 turn out to have been right? "But that makes any proof that you know capallle of refbration. "Emotion always has an object". lgad 3) Thomas's own consideration would put "contingent" where I put "future".rari!vp is true even when the truth ofp follows with logical necessity from il was known that p. And there was a long tradition. false de re pronouncement what ic known is necersary.he felt . I t may be that 1 can conceive rircitmstances that would prove me was labouring wrong. as futu1. lad). F. materiality is attributed to a stone as it is in itself." sonieone says.g.could not get it out satisfactorily. The contingency of the hture events in themselves is said to derive from the contingency of their proximate causes. But up to a short while ago it was extremely common in effect to distinguish such objects: did not Professor Ayer. He is dismissive of a number of logical suggestions."both a human being and lacking any potentiality for laughter". M~mre to this conclusion in his last years." So it is: I cannot have a better proofthat I know the coffee will not be hot. he says everything derives from some first. It belonged to the temper of St Thomas's time and to his own not to profess originality but rather to agree with previous authorities as much as possible. wc rnight argue on his own principles.e. "A cause must be prior to its effect" -by considering whether a counter-example to it can be conceived without contradiction. It is sufficient that the claim does not turn out mistaken. such as that "God knew this future contingent thing would happen" turns out contingent on analysis. when in itself it is not.. Sometimes an inconceivability seems irreducibly of the kind where the two "thuses" receive different substitutions as in "both coloured and not extended" or to take examples claimed by Aquinas . are they so after all? Indeed if one speaks in the manner of Plato of "objects of knowledge" and "objects of opinion" it will be replied by everyone that these things do not have to differ in their objects. than I could have that the coffee will not be hot: and no proof ofthat would withstand its actually turning out to be hot. And why should its being possible to imagine my being wrong. Indeed in at least one place (iasae. he thinks we cannot know them. that does not show that I may be wrong. the thing that is known. Thus he says that quo object of God's knowledge the future u necessary. Thus what might at first seem archaic turns out on reflection to conform to - - . but.82 ibfedieual and Modern Philosophers Necessity and Truth 83 know any negative Iicts!") In this case we come up with the result: If' it was knorun that p is urialtrrably true (becausea past fact) still it does not fbllow that neces. when these do not absolutely necessitate their ellccts. "But. but not as an ob. limiting human knowledge to the necessary. These views put forward in this manner certainly appear archaic.ject of thought. but St Thomas goes farther than this in his adherence to the Greek conception of knowledge. But the conception seems too difficult. "both an existent by sharing in existence and uncaused" (la. in common with many others. even though it is known now. which 1 take to mean the non-necessitation by those causes. Knowledge is not restricted to what could riot imaginably turn out mistaken: given that there arc not more specific grounds for rehsing the title "knowledge" to my claim that something is true. and ifthese do not prove the future contingent assertion. but it is a standard method to test any philosophical assertion . 6c and 44. 94. it has to be eked out by that of a point of view outside time. This appears to St Thomas far too paradoxical for human knowledge. what causes in the world cannot with. he therelore inlcrs that humans cannot know ILture contingents arid niakcs a suspiciously Facile use of Boethius' definition of eternity as interminabi1i. but I say it will not. can be characterized in itself in a way in which it cannot be characterized qua belonging to the actual context of knowing (namely.r vitae tota simul rt pedecta possessio (the complete possession all at once of entlless lifc) in order to compare God's knowledge of future contingents of present facts. Contradiction indeed may have to be rather generously conceived: not every inconceivability can be displayed as of the form "both thus and not thus". Of itself this would not exclude history from being known since Aristotle held that the past is necessary. To adapt to our own purposes the last sentence of Aquinas's article: Since the expression "knowll" refers to the actual context of knowing. on further reflection. G. The comparison with the immateriality of material things in thought shows that the divine mind does not t d e the future to be necessary. xc) we find him saying that werything in this sphere is "founded upon" the principle of contradiction. then you may be wrong is saving it won't. if it turns out hot I shall grant 1 did not know it would not. Accordingly we find St Thomas maintaining this opinion. 14. "il' the coffee can turn out to be hot after all. where the two occurrences of the word "thus" are replaced by the same term." Well. St Tho~nas surely did not need to adopt these devices: he had put his hand whal is on the solution in the distinction between the trur de dicto r~ecessity known is true and thc. but. (la. at least by divine power:' Thus we can know conceivably be tan~peretl nothi~~ ot'thc g li~turc. teach that all that was certain was either a truth of logic or an expression of immediate experience like "I am imagining a red circle" all else being at best merely probable? These doctrines indeed are now not so much taught.

The retreat from atomism. are such as to make us need tools of philosophic description not always unlike those used by a medieval philosopher. we tend to think. But these matters are still very dark. with a not totally unfavourable eye. meaning and truth. is difficult to explain or to translate very directly. and it is noteworthy that concepts of experiencing are only some equal citizens. if we do not. The medieval concept of intentional existence. is far more difficult to carry out successfully than might appear. that is. We can often see what he was at where our great-grandfatherssometimes wen our fathers . ad 2). though it is easy not to realize this. a course it is difficult to approve. or remain in views not arrived at by philosophy while we work at concepts. On what grounds. may even be making a comeback. but how this is so is so intensely obscure that one surveys the obscurities of the scholastic esse inteU@ik. Indeed it is well es- 85 . believe in the Evil Eye? .the original technique of the Latiri translatioris ofAristotle and the first English translation of' is very ditierent. The present transiatior~ what is rather an explanatory commentary than a translation. called 'intentional inexistence' by the scholastically trained philosopher Brentano. which we want to describe.84 Medieval and modern Pliilosophers Necessity and Truth very general philosophical practice: philosophic uiiderstanding. giving a rationalistic account of the phenomenon (la. But the fact that it is now not so difficult to explain such phrases marks some change in philosophy.among those that we want to understand. language. with no substantive being of their own. concerns what rnust be so. one translator otfers us St Thomas. like many present-day Neapolitans. We are nowadays in some ways closer to the ancient and the rnedievals than we are to. say. Or if philosophic understanding is achieved by a successful delineation ol'concepts. There are doubtless many points which go together to form this situation: but if the reviewer were challenged to name what most strikes him he would point to the eclipse in recent philosophy of the notion of6'thegiven" and the manner of the current gradual rcdeparture from atomistic conceptions. not on grounds of' any a priori conception of knowledge. Such a phrase. The logical features of concepts. our philosophic activity is one of describing and clarifying this milieu to ourselves. unless one is to employ the gobbledegook of mere transverbalization . largely: because the people around us d o which St Thomas. then? Why do we not.culture. tablished now that we cannot regard the emotions as merely contingently connected with the kind of objects they have and the aims they tend to make us have. But perhaps we are not all entirely lazy: there is a great deal of work to do before we can face such questions. - . whose actuality is the same thing as the actual occurrence of a thought of such-and-such. of how they happeri to be in this our . So we accept common views. What we believe to exist we credit and what we do not believe to exist we discredit. a present-day English speaking philosopher would be very likely to say 'the lot'. it will be clear. 117. then our great interest is to note what could not be supposed changed about a concept without quite changing what we are talking about in using it.what.or any human . And it seems equally clear that thoughts are characterized by what they are of. We want to get clear about the concepts we habitually use before we trust ourselves as philosophers to use them for purposes beyond our immediate ken. Asked what was given. We start mediis in rebus. no more .could see nothing but useless subtleties and distinctions. a reappearance in modern dress. The sorting out by Aquinas of the de re and de dicto necessities confounded in 'what is known is necessarily true' is something a present-day English philosopher can appreciate. Kant and Hegel. belongs to what St Thomas calls the ralioJormalis obiecti.g. did believe. as we change the propositions of geometry by changing from one geometrical system to another . It is to be feared that the reason is often.

The historical example is an inference of the original cause. I'tie needrd revision then reveals the trovertibly ~iceded position as incredible.r of March. . The topic is our belief in matters falling outside our own experience and memory: When we infer efkcts from causes. that he strikes one asjust making this point. . is an impression of the memory or senses. and these ideas were either in the minds of'such as were immediately present at that action. 34. which causes we must ascertain in the same manner either by a present impression. if by 'senate-house' Hume meant to indicate a (1973). The Senate was not meeting in the senate-house. amend: "When we infer eff'ects from causes or causes hom effects". We must. or by an inference from other causes. ma hat is wtly the Section "On the idea. Let us see what this looks like in the case in hand. and consider for what reason we either believe or reject it. . Section 1V of'Part 111 of Book I of the Treatise is a doubly unusual piece o l philosophical writing for Hurne. cause and effect are taken by him to be inferentially symmetrical. it must. the killing of Caesar.r Srom what is seen or remembered. etc. pp. (SelbyBigge's edition. And this actually is the case with all hypothetical arguments. From Amly~u. and that again from another testimony . . either by perception or by inference from other effects. nor belief of a real existence". and that because this fact is established on the unanimous testimony of historians . . which effects we must building. but a cantilever. . either by an immediateperceptior~ ofour rliemory or senses. o r belief" is in the middle of the part which we would think of as the Part on cause.) But also. The chain of inference has to stop or else "there wou'd be no belief nor evidence. Indeed. . o r causes from effects . the signs of certain ideas. we must come to belief which we d o not base on grounds. or by an iilterence from their causes and so on. we must establish the existence of those effects. Revision is incont o secure coherence. . So for him the tracing back is inference too. testimony . and the only thing that can stop thern. still stand? Yes. all seems uncommonly smooth and acceptable. 'till we arrive at . What is in question isn't a chain nailed at both ends. eye witnesses and spectators of the event. the present perception of certain characters o r letters. But note that it must be purely hypothetical inference. . which are seen or remember'd. This is not to infer eff'ects ftom causes. until we arrive at an object which we see or remember. not our perception. But Hume is arguing not merely that we must have a starting-point. . but that we must reach a startingpoint in thejustification of these inferences. What is its starting-point? It is natural to say the starting-point is the present perception. Does our original amendment "When we infer effects from causes. from its remote effect. But it has to be our perception. 'Tis obvious all this chain of'argument or connexion of causes and effects is at first founded on those characters or letters. The impossibility of running up with our inferences in infiniturn was not occasioned by our incapacity or exhaustion. I believe this is false. until we arrive at sorne object which we see or remember. . 89-3) Now this is a credible account of a kind of'prognosi. then. not that inference must terminate. beyond which there is no room tor doubt or enquiry. . . Let . The inference goes through a chain of effects of causes which are effects of causes. cause and effect are inferentially symmetrical. But as we have said. together with the one that the starting-point must be perception. I For Hume. Thus we believe that Caesar was kill'd in the senate-house' on the ide. For historical belief': When we infer causes from effects. not "we cannot carry our inferences on in inanitum" but "we cannot trace them back in infiniturn"." ' Stickling fbr accuracy. g Hume and Julius Caesar . Read very casually.) "'Tis impossible for us to carry on our inference in infinitum" means: the justtfitation @he groundr ofour inferences cannot go on in infinitum. p = Caesar was killed q = There were [at least ostensible1eyewitnessesof Caesar's killing r = There was testimony from the eyewitnesses = There were records made. . . we must establish the existence of these causes . (ibid. .H u m e and Julius Caesar 87 ascertain in the same manner by a present impressio~l or by an inference from their effectsand so on. the relation of cause and effect is the one bridge by which to reach belief in matters beyond our present impressions or memories. and it collapses. Here are certain characters and letters . Where we have chains of belief on grounds believed on grounds. A little attention. or they were deriv'd liom . That once noted. 'Tis impossible for us to carry on our inferences in injiniturn. etc. He would have been clearer ifhe had said. one reason why this passage of Hume's seems fairly ordinary and acceptable at first sight is. But that cannot be a sufficient exegesis! For what on this account has become of the argument that we cannot go on in inaniturn? The end of the chain is now the death of Caesar o r the perception of its eyewitnesses. there being in them neither any present impressions. but rather causes from eff'ects. . what must be our astonishment o n obser- ving that in illustra~ionHume itlvites us To chuse any point of history. The argument here is that there must be a starting point of the inference to the original cause. deriving from the testimony .

however confusedly. For let us ask: why do we believe that there were eyewitnesses of that killing? Certainly for no other reason than that we believe it happened. Then our belief in that original event is a ground for belief in much of the intermediate transmission. not p from q. and usually could not. That is the last consequent. who knew the people. then s. will be a ~ ~ o l h e r belief q. We don't believe in the identity because we believe in the spatio-temporal continuity of a human pattern from now here to then there. cite the chain ofrecord back to the eyewitnesses as an illustration of the chain of inferences via cause and effect. and his existence was made up for particular ends. But a discovery that a name belonged originally to a period later than the life-time of the supposed bearer of the name at any rate reduces the status of the name: it becomes equivalent to some set of definite descriptions. It is a lot more difficult to see what to say. Fourth. a man whom one recognizes and identifies as a man seen last week. Perhaps we really do infer an effect Iron1 it as cause: "There will have been chaos and panic in the Senate when Caesar was killed. I have heard that the Rabbis held that the 600. then r. if the belief' is not a prrception.a proof that this man was in New York in between.has the rare character of being easily demonstrated while yet it touches the nerve of a problem of some depth. But now: why does anyone believe there were 600. we believe by inference through the links in a chain of record. ( P & i l o ~ h i & Emerkungen. One of the rare pieces of stupidity in the writings ofwittgenstein concerns this matter: - That it is thinkable that we may yet find Caesar's body hangs directly together with the sense of a proposition about Caesar.but let that not delay usi with thc mere idea of Caesar's death. Ifthe written records that we now see are grounds of' our belief. we implicitly depend on an 'apostolical succession' of users of these names or linguistic transforms of them . if r. the immediate. which fbllows from. Third. and so o i l back to p. Now we go in the other direction: since t . So we reason . while that man was not would destroy our belief in the identity." But "we must establish the existence of this cause". r . Mulath mulandis the same holds for the chain of transmission of historical information. Hume andJulius Caesar 89 t = There are characters and letters to be seen which say that Caesar was We must suppose that we start (how? . It is the other way about On the other hand a proof of a bred in the continuity. The mistake which I think it is now not a bit of patronizing superiority more passed through.topher. First. than it is to point clearly to error in Hume. We will suppose that the document itself gets acknowledged by experts in such matters as a In On CmidMy. invented the story of the divine Julius so that Caesars should be worshipped.going back to original users. We do not. justification for a belief p. IV. That would not be very interesting or important. Compare one's belief in the spatio-temporal continuity of the existence of 1 owe this information to Dr Stephen Kau. since s. P. with which we cannot run up in infiniturn.56) What document or inscription could be evidence that Julius Caesar never existed? What would we thinA for example of an inscription saying "I. In using proper names that we take to be the names of people we don't know. And let us make no mistake: it is not otherwise for belief in there having been eyewitnesses to Caesar's assassination.Recause he believes that 600. but only this is a chain of inferences through causes and eflects such as Hume envisages. the ultimate belief must be of a quite differenr character from the derived beliefs: it must be perceptual beliet. There is an implicit corollary: when we believe in historical information belonging to the remote past. o r people in the remote past. s . Augustus witnesses? . As we have seen. - - All this is not just catching Hume out in a mistake. a chain of' reasons for a belief must terminate in something that is believed without being founded on anything else.t killed. Not all these hypothetical propositions are equally witnesses to the crossing of the Red Sea must be credited. they are first and foremost grounds for belief in Caesar's killing. we shall rather have to derive it as a cause ti-om an effect. belief that thc assassination is a solid bit of history. or presently remembered. W h wrote On *C sugption. by deriving it as all effect from a cause. then q. At most. it is not a belief in the historical facts by an inference that passes through the links of such a chain. Second.88 Medieval and Modem Phi1o. but he never existed"? To ask a question Wittgenstein asked much later: what would get judged by what here?s Take something a bit less extreme: a document recounting a conversation about siege-engines between Caesar and Archimedes. we believe that there has been a chain of record. then 1. So Hume's thesis falls into b u r parts. But it is not like that. . 1 un a @dal hrdebccd to On C ' Wittgmrtcin would not have made such a q in this artick.2 600. Hume must believe all this: otherwise he could not. trace the chain of use of the name. It is so also with proper names. to refer to as such . that can very seldom be the case. just as much as it implies. this will not be as (at the beginning) Hume suggests. Belief in recorded history is on the whole a belief thd there har bem a chain of tradition of reports and records going back to contemporary knowledge.and here our reasoning must be "purely suppositious and hypothetical" -: if P. We infer q from p. . from which it emerges that no such man ever lived. But so too does the possibility of finding something written. I t terminaurs in something that we perceive. We can assert this witnesses! That's a lot. belief in something perceived. if q..

Ovid. But not in the sense that the hypothesis that one of my grandfathers was a half brother. But in face of such an hypothesis about Caesar one would have to ask: "What am I allowed to count as widence. I do not believe he could. with its being a piece of'historical writing.I can at best remind myself of. I mean..90 Medieval and Modem Philosophers H u m and Julius Caesar 91 genuine old MS. but its Latinity and the absence of 'giveaways' won its acceptance.haractcr of document that could 'emerge'. oral. with a particular logical status of one hind of certainty. or was not assassinated.) ?'he content proves it to be fictitious. existed and that his life terminated in assassination: this he could call in question only by indulging in Cartesian doubt. then?" People 'in history'. No one can except by finding out the status of the information about him. of a time when there was little contact. works of Caesar? You were told it. whether it was made in the tenth or eleventh century and it comes under much critical scrutiny. Caesar. (But we should not forget that it is by traditional. but. "It wouldn't do. I was taught. And yet he got a detail wrong! And yet again. come to our countries. ~hakespeare'splay. but if I doubt the existence of Caesar. It wasn't an accident that Hurne took the killing of Caesar as his example. if I say I may reasonably call it in question. perhaps. and thcse. It' you go to an expert on Julius Caesar. of the other is such that the supposition of its truth involves destroying bases and standards for discovering any historical facts at all. it seents reasonable to place the writirlg of'it in the first century BC.) The Hellenicity or Latinity is authentically ancient.then the effect of the hypothesis is to make a vacuum in which there is nothing by which to judge anything else. 1 once asked an expert on Galen how he knew that his subject existed. Somehow he doesn't learn at once that this is not a purely fictitious story. if possible. and then someone tells him Caesar was a real historical character. then with equal reason I must doubt the status of the things I've just pointed to" . Contrast an expert on King Arthur. say. who hears of Caesar from a traveller. Of course Wittgenstein cioesn't tell us horn what r. The response was surely a correct one. . he could have called that in question. The hypothesis about Galen is merely one that 'won't do'! That is: one can relate him to better known historical matters. thc De Herum Natura was suspected of being a forgery. He refers to Caesar as a fictitious character. and see that it is incapable of status wen as a wild hypothesis. or its limits. He can check this. as we say. or are editions of ancient books. that detail's being right would not be an important aspect of what lie knew. The style is such as might fit. But I cannot. then? . (To be sure. That means that there were standards by which to judge. Virgil. find out that the corpus of solid historical information belonging to our culture does indeed include this. "we know too much about him -" and went on to mention Galen's connection with Marcus Aurelius as an example.And so also about people 'in . I think. he was taking something which existed in his culture. The ancient Latinity of Horace. So I do not mean that it is vastly improbable. you will find he is an expert on whether Caesar conducted such and such riegotiations with Pompey or when he wrote his books. His reaction was to consider the hypothesis that Galen did not exist. He is accustomed to chronicles and traditional infortnation. The most 1 can do is: hame the hypothesis that Caesar never existed. - . Though I have never given the question any thought before this. I know I had more than one. Dispute exists. But it is very diecult to characterize the peculiar solidity involved. He can check on it. The attempt to construct a serious doubt whether we have writings ol'cicero. . What does the hypothesis amount to in face of our information about the time? But if all that is irrelevant as we could have no reason for doubting the existence of Caesar. that when Lucretius was first published during the Renaissance. I mean: suppose there were a schoolchild who first ran into Caesar through ' Thought to be hiqtory in a time of very sketchy impressions of ancient Persia11history. wr.or I should realize straight away that the 'doubt' put me in a vacuum in which I could not produce reasons why such and such 'historical facts' are more or less doubtful." he said. It could not force an adjusttnent in our idea of the relative dates of Archimedes and Caesar. for exatnple.) He learns that Caesar is supposed to be such a character in our history. . He can learn our languages. he can look into history books and find out that that's what Caesar is. and exists in ours. cven have his own writings! And how do you know that those are ancient historians.the status of "Julius Caesar" as a name of an immensely famous man 'in history'. But I cannot. but continuing to believe in Cicero and Pompey . could it find a ground from which to proceed ? We know about Caesar fi-om the testimony of ancient historians. I mean that either I should start to say: "How could one explain all these references and implications. I cannot check that there was such a person as Julius Caesar. And how did your teachers know? 'They were told it. It might well be that the discovery of' S L I C ~a piece would compel some adjustn~cntin our picture of what was 'on' in the literature of the time. not just fiom explicit teaching. We know it from being taught. are not in any case hypotheses which we have arrived at to explain certain phenomena. itself known by tradition and never subject to question. I already know . information that one knows that these are chronicles. Not on whether Caesar existed. if he had been careful. Cicero and Caesar was such a standard. you know.but that that man.(It is no Piltdown skull.) Or again: suppose a Chinese man. (XenophonYs Cjropaedeia4 is an exaniple of such writing. Do I know I had four? I would have said so. he could wen perhaps have called the date in question (might it not have been a false accretion?) . No more than is the fact of my birth or the existence of my great-grandmothers . these things can change. but by its being implicit in a lot else that we are taught explicitly. but.

that one may. This is a causal explanation of its coming to be at point B at t' and it cannot "without contradiction or absurdity" be supposed to happen without the thing's coming to be at point B an event that might. lo "Whatever has a Beginning of Existence must have a Cause" Hum's Argument Exposed . have occurred otherwise. and wlurt is thenature ofthe intrcncc. Part 111.e.e. A general epistemological reason for doubting one will be a reason for doubting all. that is. Thus. the analogy is not good. Hume distinguishes two questions. Section 11. I suppose. Section 111 of Part 111. a double one: Whj we conclude that suchpartitular c a m must necessarily have such/warticJar efecls. build on to . suppose a given sort of cause to occur without its characteristiceffect.or is a conflation or the like. I may learn that it arrived there from somewhereelse. and vice versa. it travelled from point A to point B from timet to t' and so was at point B at t'.while it is afloat: if this suggests that we can go round tapping wery plank for rottenness. that that first 'principle of causality' cannot possibly be intuitively or drmonstrably certain. Book I of the Treatisc is devoted to proving that "Whatever begins to exist must have a muse ofexistence" is not intuitivelyor demonstrably certain. without contradiction or absurdity. In A Treatise o/ H U M Nature. But not everything can be put up for checking. i. If I ask how something comes to be in a certain place. is not what would nowadays be called a proposition whose truth is logically necessary. Neurath's image is of a ship which we repair . however. But it is prima fkie true in a vast number of cases. and he devotn the rest of the section to disposingof such ~rguments - . and so we might end up with a wholly different ship. It isn't clear whether Hume saw this. and then none of them would have anything to lest it by. There is a counter-example. For there are things that are on a Iwel.and. What Hume calls "intuitively certain" is a proposition whose truth is discoverable purely from examining the ideas contained in it. This semu to be very generally true. and there is the possibility of discovering that some obscure supposed historical tigure is probably mythical. if for example it came into existence at that moment at that place.jom the one to t k dhcrf Not only are these distinct questions but a person may consistently hold that a beginning of existence must always have a cause. we can't accept Hume's principle as impregnable. which we shall soon examine. i. without holding that "such particular causes must have such particular dfects" (or that such particular effects must have such particular causes). and what he mlls "demonstrably certain" is apparently a proposition which follows from something intuitively certain. His most famous thesis is that the ideas of cause and effect are always separable. the one: Why a beginning ofexistme must necessarily always have a cause? and the other. Book I.9 Medieval and Modem Philosopherj history' there are gradations. Things get corrected or amended because of inconsistencies. He first produces a 'proof'.

with the observation that Hume has hardly done itjustice. that it entails q. space and time are relative. Yet Hume doesn't consider this. Hurne replies: Is there any more difficulty in supposing the time and place to be fix'd without a cause. this does not show that every beginning of existence rnust be an effect. I t is not fair to say: you cmnd argue like that. purporting to demonstrate it. This is a very obscure topic. to show that a beginning of existence without a cause is not demonstrably absurd. since they are both upon the sarne fboting. Hume is saying "But why should I find q absurb. which is peculiar t o one liriie arid to orie place. the question arises as to how that time and place could be specified. Thus we specify a time as n years ago. it will equally require one in the other. which suggests that Hume is consciously in Leibniz country. Without existing at a definite place and time. But as soon as you consider it as existing at a given time and place. Well. if a thing that comes into existence must come into existence at a given time and place. Not that the arguments are invalid. 'consider'd in itself it can't be considered as coming into existence at all. This argument indeed seems very uncertain. Or if he should say: "consider'd in itself" excludes "consider'd as existing at a given time and place". The absurdity. Namely: space and time are relative. already given by him. another. and ~ u m does suggested by the argument he cited from Hobbes. And this is indeed the case. then. Or at least sow other existence. that is. Hume could say: the thing is and can be connected with the time and place only by the brute fact of existing then and there: you can think of that without invoking something else that makes it exist at that time and place: But the argument suggests another one. in which we can supposeany object to begin to exist. e not deal with the questions That is the argument. of the one supposition can never be a proof of that of'the other. it must be so in the other: and if that absurdity be not clear without a proof' in the one case. you are conceding it is not intuitively absurd. and rnilst stand or fall by the same reasoning. or to have existed. however. then. he'd see no difficulty in q. Let us now attend to our main business: the argumerit. but the requirement of other existences by which to specify time and place does disprove Hume's great principle "That there is nothing in any object consider'd in itself. when and where it shall begin to exist. The questions are not easy to reformulate and seem better left out. What is Hume's argument? He is saying: You cannot argue to the absurdity ofp. for ifsomeone saw no difficulty in p. but anyway. by producing an argument instead of simply sayingp is absurd. whether the object shall exist or not: the next. which is absurd". which is absurd. this specifiability requires a determinant other than the thing which is supposed to exist. The point is obvious indeed. are in themselves equal. if I don't already findp absurd?" And to this there may well be an answer in this case. a particular finite and non-eternal thing won't exist at all. One is not initially supposing anything of the sort. For as far as concerns time. and which by that means determines arid fixes the existence. for Leibniz argues for the identity ot'indiscernibleson the ground that God must have a reason for putting A liere and now.)This is somewhat cavalier. and so can hardly prove it. So much for the second. This is not to give a cause. and B there and then. Why does he say "The first question is whether an object shall exist"? This seems to consider the matter from the point of view of a creator. antecedently to a thing's existence at a place and time. now f i it is not intuitively absurd. unless something already existent makes that the time and place of the thing's existence. even that is not clear. The first argument. it might be posterior. and we will leave it. Another argument Hume reasonably calls still more ttivolous: every efect must have a cause. Hume says. One says that something that came into existence without produce itsell. but rather "There is at any rate this absurdity in p. In this passage. Hume writes as if the place and time for a thing's existence could be specified independently of its existence. these arguments assutile what they set out to prove. but they cannot prove the conclusion. indeed they do not descrve attention. than to suppose tlie existence to be determin'd in that manner? The first question that occurs on this subject is always. . His making this point will be of some importance tbr our understanding of his positive argument that this 'principle ot' causality' cannot possibly be intuitively or deinonstrably certain. in the place and at the time. Hobbes' argument. from the fact that it entailsq.94 Medieval and Modern Philosophers Beginning o f Existence and Cause 95 as he knows. This is not '2ust reasoning". which is impossible. I t is that "all the points of time and space. it may not be prior. The argument that he is considering assumes that too. 199). and also it is the truth. of which he is ablc to rehie the last three quite easily. since cause and eNtct are relative terms. which give rise to this one: he has apparently not seen that the question arises what the requisite specification of place and time could mean without a prior existence. He considers b u r arguments. there can be no distinction of that place and time from any other unless something else distinguishes them. because you are initially supposing thatp is not absurd. as cited by Hume. p. that it is produced a cause ~iiust by nothing. if they cannot prove it except to sotrieone who already accepts it. which there could not be unless something distinguished A from B. is obscure and is dealt with rather sketchily. If the removal of a cause be intuitively absurd in the one case. any more than the truth that every husband rnust have a wile shows that every man rnust be married. The arguer is not saying "There is no absurdity in p but there is in q". which is incapable of producing anything. and unless there be some cause. seems to go on as follows: Antecedently to a thing's existence at a place and time nothing can connect it with that time and place more than with any other. Since. tlten. (Rather as if the argument were a shaggy dog story. third and fourth arguments which lie considers. it must remain in eternal suspense". then neither is q. which can afford us a reason for drawing a conclusion beyond it" (Selby-Bigge's edition. perhaps q is not 'in' tuitively' absurd but there are further arguments to show that q is absurd. Well. As Hume says. As he says.

the possibility of a beginning of cxistmce without any given cause. as. I imagine a star or a rabbit beginning to exist. o f t h e idea of a cause from that of a beginning of existence is plainly possible for the imagination. So here the argument from imagination is sound. supply a specific cause. Then I can say (6) For any beginning (or modification)of existence E and any particular cause C. To supply such a particular case is both reasonable and conformable to Hume's doctrine of abstract ideas. so good: "separable" presumably means "such that one can think ol' one without eo ipso thinking of the other". and consequently the actual separation of these objects is so far possible. What does this mean? There are tv~o possibilities: that it is possible to imagine a beginning of existence without imagining a cause. The separation. as far as concerns one of them. For example. therefore. He must. and existent the next. another rabbit. (71." I may say. yield . But from Hume's giving an example of' an effect in (g). and any particular cause C. "I know what it would be like to find a kettle boiled without a fire. like husband and wife. And here arises the difficulty. : . so we have (4a) But the proposition does not give me the possibility of imagining an effect without any cause at all. For the argument from imaginability to possibility has a good deal of force on the second interpretation: let me imagine any event . thenmean the second. For quite generally from For any. it does not give me: (8) 1 cail imagine this: there is a beginning (or modification)of existence without any cause. We can either forsake the doctrine of abstract ideas and say that that is to be a sufficient description of our image in the particular case. as we did with "a beginning of existence". (61. It is possible to imagine something's beginning to exist without a . The ideas of cause and effect are distinct. 79-80) . (3) : . is plain: it is "a beginning of existence". and even "I h o w what it would be like to find a rabbit coming into being nof from a parent rabbit".e. From this he draws the conclusion (5) : .. ( 1) (2) All distinct ideas are separable. that it implies no contradiction or absurdity.more so than my quotation shows. But what is the other 'object'? The does not follow: I can imagine that a rose has no colour. Let us even suppose that Hume is right in saying it holds universally. Nor does. I can imagine that a rose is not that colour For m : . without conjoining to it the distinct idea of a cause or productive principle. That is. I t will be easy to think of an object's corning into existence without thinking of a cause. the possible exclusion of any particular cause in the imagination.96 Medieval and Modern Phihophcrs Beginning o f Existence and C u e 97 It is an argument froni imagination. Let it hold for any particular cause I care to introduce. So far. we must take him to mean that the ideas of whatever objects are causes and effects are distinct fiom one another.say the heat of a fire. His argument is rather prolix .. Let us set it out proposition by proposition. and not only can I imagine the one's happening without the other. As all distinct ideas are separable from each other. cause. The first certainly follows from (3)but is too close to it in sense for us seriously to suppose it is what Hume means. and that it is possible to imagin ne a beginning of existence without a cause. The next step is the crucial one: (4) only answer we have is "a cause". . But this sound argument does not yield the desired conclusion. I can imagine E's happening without C. or we can.and especially in view of his calling "frivolous" the argument that the ideas of cause and effect are correlative. This makes one ask "What objects?" The answer. and infer from this (7) For any beginning (or modification)of existence E. Now we can go two ways. We might query (s)on the grounds that cause and effect are correlative..say the boiling ofa kettleand any particular cause of it . The actual separation of these objects is so far possible. it is possible that not there does not follow: It is possible that for nom For example. from: y colour. there is no contradiction or absurdity in the supposition. E can be supposed to happen without C: i. and as the ideas of cause and effect are evideiitly distinct. but the imaginability is of a sort to convince me of the possibility in the sense of 'implying no contradiction or absurdity'. The separation of the idea of a cause from that of a beginning of existence is possible for the imagination. 'twill be easy for us to conceive any object t o be nonexistent this moment. that it implies no contradiction or absurdity. for neither in reality nor according to Hume can there be a bare image of a beginning of existence which is not the beginning of existence of anything in particular. or what (we w granting) follows from it. or the compacting of nebulous material.. (pp.

this supposition is without contradiction or absurdity. it cannot be thought of without thinking of such and such. therefore this is possible. Generalizing. Then the argument is simply: ciple that "It is the same thing that can be thought and can be". : . and our ohsenling that there is no parent rabbit about. The trouble about it is that it is very unconvincing. Indeed I can form an image and give my picture that title. I just imagine a rabbit appearing. With some caution arid restriction. . one wonders why Hume did not give the argument straight in this form. this does not mean that I can think of it as existing without having a definite number of leaves.. the 'precise degree of any quality' cannot be distinguished from the quality. 18). the title ofthe picture. there is no contradiction in supposing) that something comes into existence without a cause. as it were. e.e. but is only the conclusion that the effect can occur without any particular cause which you have imagined it without. For if I say I can imagine a rabbit coming into being without a parent rabbit. and without a cause is nothing but. and accept that the second 'object' is just 'a cause' and no more. we may grant the Pannenidean prin- We can imagine something's coming into existence without a cause. We must. then. It is even wrong in the particular case from which we formed the generalization. The principle we have attributed to Hume as inspiring the argument we have been examining is essentially this one contraposed. His argument for this is that. "This can be imagined. Holding it obviously absurd to suppose a quality to exist. then that thing can be thought of' as lacking that circumstance and hence can exist without it. The attribution is probably correct. Naturally.e.g. in which we forget Hume's doctrine of abstract ideas. try the other tack. But what am I to imagine if I imagine a rabbit coming into being without a cause? Well." but the this is not the desired conclusion. we may put it: If something cannot be without such and such. If this is the right interpretation. though an extra step has been put in. i.98 Medieval and Modern Phdm$lurs Beginning o f Exislerue and Cause 99 (9)A beginning of existence can happen without any cause. Hume's argument can be rendered more intelligible to us if we attribute to him the following principle: If a circumstance need not be thought of' in thinking o f a thing. well and good: I imagine a rabbit coming into being. But Hume's extension of it is certainly wrong. I can imagine or think of a sprig of leaves as existing without there being any definite number of leaves that I think of it as having. nothing whatever follows about what is possible to suppose "without contradiction or absurditywas holding in reality. So if we go this way we have (perhaps)a sound argument from imagination. for we know that Hume thought that "the mind cannotform any notion ofquantity or quality withoutforming a precise notion (degrees ofeach" (p. Hume thought there could be no such thing as an idea of it which was not an idea of any particular degree of it. It is possible (i. That this is the imagination ofa rabbit coming into being. But from my being able to do that. though in no particular degree.

as we say? In illustration. He physically can't push it open. Brentano however is pointing to some conceptual relationship-in him it is an assimilation between will and the emotions.isn't it extremely like that spirit. nisi voluntas in eorum consensione." (Bd. We find Augustine's 'yes' and 'no' in Brentano too: Wenn e h ~ v lnhalt cines Urteils =den kann. cupiditas. distress and joy. . something inner. The contrast isn't so great. though Brentano will grant that people will hardly call it emotion. who is talking about a will that occurs just prior to an act: Augustine calls the principal passions all will and Brentano would l i k to call will a passion. (mutatb mutcmdis) for fear and distress. . If we can insert something psychological. Augustine makes a certain assimilation too: - voluntas est quippe in omnibus: immo omna nihil aliud quam voluntates sunt. that is called joy. .separate imagination andjudgelrlent into two fundamental distinct classes of psychological phenomena. talis voluntas metus est. 8) The list doesn't include fear. even if one assumed an intercalated act of will. that's desire. it clearly belongs to the same class as the 'Mut'." You may think this isn't like Brentano. only more committed to the action? To see that we might do so. that is. given that little extra. it took Frege to distinguish predication from assertion. x. saying no to [dissenting from1 the things we don't want? When we consent. inrofm er als gut gcnchm (im weitcaten Sinne des Wortes) odm ah achkht u q p d m sein kann. XIV. den Versuch zti unternehmen . nisi voluntas in dissensione ab his. "For what are desire and joy but will. but when we consent. "Nay. 7) However.Verlangen. before he actually pushes the door open and steps forward. 5 (1978) I i In just the same situation. Wenri man aber auf die Zwischenglieder achtet und immer nur die nPchststeheriden miteinander vergleicht. Kap. fear and desire. VI) "There is will in all of them. He says of the copula that it ". propositions or their content.he perceives that the swing door is locked. he has arrived at the state of 'Mut'. cum autern dissentimus ab eo quod nolentibus accidit. he hasn't even tried to d o it. . Das cine Extreni ist ein Gefiihl. but how small! And aren't they obviously the same in kind? If there is that last term there at all. Where no 'Mut' is needed one couldn't find a likeness. und sie srhcinen weit vorieinander abzustehen. Nor is he alone in this. when one was describing some act like picking up a glass of milk to drink it. seeking what we want. Itemque cum dissentimus ab eo quod accidere nolumus. and Brentano is surely right in combating the Hurnean thesis that there is no difference between mere images. . s o kann es lnhalt cines Phlnomens der dritten Grundklvre mrden. all the rest. The former enterprise is vitiated by his failurr to distinguish between predication and assertion. nur den Ausdruck von Vorstellungen zurn Ausdrucke cines anerkennenden oder verwerftnden Urteils erggnze. zeigt sich da nicht iiberall der innigste AnschluR und ein fast unmerklicher Obergang? (Bd. These states or events are differentiated from one another by the peculiar colouring associated with each. and. whom he is summoning up the courage to beard. they are nothing but wills. saying yes Iconsentingl to the things we want? And what are fear and sorrow but will. of possible contents of judgement. corisider that he might summon up the 'Mut' and then realize that the action was impossible . this was an almost universal error. 2.I 11 Will and Emotion 101 Will and Emotion Brentano's Psychologie vom empirijchetc Standpunkt is a work which belies its title.something which belongs in the development which is to culminate in action. It would perhaps be correct if "empirischm" had been altered to "empiristbchen"." And likewise. Note that we are persuaded to make the assimilation by a rather special type of example. Now. daB es uns tuteil werde . insofern a alr wahr annehmlich oder d s filach d i d ! ist. imagine a young person standing outside the door of' someone alarming. es uns zu verschaffen . Nam quid est cupiditas et laetitia. spirit to make the attempt . quae nolumus? Sed cum consentimus appetendo ea quae volumus. won't it be almost the same as the 'Mut' itself. and hence in the same class as . Now consider the next thing. and which conies after 'Mut'. isn't the act of will which he puts in as the last member of the series. And so we have will assimilated to emotion. das andere ein Willen. (Dr Civitde Dei Lib. quae volumus? et quid a t metus et tristitia. This is developed into the characterization of emotions (and therefore will) as a set of states o r events whose common theme is acceptability o r unacceptability. in there at all . Brentano's second enterprise is to my mind the more interesting and powerful. But look a little more closely.Mut. the act of will itself.WillensentschluR tur Tat. He has just nerved himself to walk in. Now if that's what happens. he goes on. say. having the thing we want. in which however he doesn't notice the metal tongue of the lock in position. . So there is a difference between this last term and thelMut'. that sentiment of boldness or courage.Hoffnung. I L . he says. Augustine is concerned with just four generic emotions. cum autem consentimus fruendo his quae volumus. Cap. not as true or false but in another way. but might easily have done so: it could go in after longing. he won't indeed push the door open (for he can't) but he will have tried. that nerving of oneself. laetitia vocatur." he says.Sehnsuch~ vermikten Gute .It labours on the one harid to . From Crllur Philo~Ophi~thc Sludim. Kap. He puts the act of will (in a particular case) at one end of a spectrum of emotions: nach derri Betrachten wir als Beispiel die folgende Rrihe: Traurigkeit . and on the other to assonate will and emotion or even to identify an act ofwill as the occurrence of an emotion. talis voluntas aistitia est. one member of that class.

one may say. Geach has written against the idea of there being any sense in positiveness of belief .all these anyone will call positive. It does make sense to say of someone "He's interested in what colour something is. but not what it is. But didn't we say that anything can be given a negative form? It makes no sense to say that someone is interested by the fact that some man is alive and not by the fact that he is not dead. ( 9 ) If you are willing that something should happen. like the young man in the English song "0 no~ o h no n John no John. (4) If you hope that something will happen. and "yes" in response to the corresponding positive. And it might be a characteristic of someone always to relish believing negative things.but how does that fit with our observation that affirmation and denial of the contradictory are equivalent? If that spirit will keep a promise to say "no" to cvetything.102 Medieval and Modern Philosophers Will and Emotion 1°3 The comparison had already been made in respect of desire. No!" Goethe's line is very evocative: is it more?Once again. hope. I think. what qualities things have. admiration. by Aristotle: Por~ b'dnrp P v bravoiq rardtpaorq ~ aandpaorg l rovr'lv opC&r 6io<rqxal pv)vj (Whatascription and denial are in judgement. Whereas hatred. friendliness. must you fear that it will not? Here (pace Spinoza) the answer seems to be: not generally.5 covers sensual desire (epilhumia)and anger (thumos)as well as wish. e. distress ('Unlust') at the idea of its not occurring? To this the answer seems to be clear. 'sighted'.more strongly than most people . (3) If you want something to happen. Belief equals disbelief in the contradictory and any proposition can be given a negative form. As soon as you make this comparison you are faced with the equivalence ot"'no".) (Nitomachean Ethics. All of them speak of. Belief is as its objects are. not in what colours it is not". in a very generic sense of the term. In the passage I quote some might claini that he doesn't mean to refer to the passions. only disbelief in something positive. distress.) This is the only way in which it makes sense to speak of belief as positive: 'positive' belief must mean belief in positive things. because the passage is leading up to the explanation of choice. I i gga. But hope and fear are tied up with expectation in complicated ways. surprise. (It is certainly not a theory of all predicates. 'blind'. is readily acceptable. Similarly when we say that one who believes p disbelieves not p we haven't said yet whether he can also believe not p. at least if the thought of'that is entertained.g.with which he wants to contrast will but he has. fearfulness. scorn. it seems to mean something in the following way: someone may be little interested in what is not the case. decision and choice. Aristotle's theory of the categories is a theory of things which are positive in our present sense. lack of interest. We may accept the idea of certain objects of belief as positive (whether or not their expression contains a negation) though we need have no general theory of all propositions as ultimately positive or negative in sense. 'nerve' . one wishes that he may - . and there is no doubt a host of cases where hope of something does involve fear of the contrary. what they are. Does this carry through to emotion and will? Here we can ask various questions: ( 1 ) IS pleasure ('Lust') at the idea of' something's occurring equivalent to. sadness. and what they aren't. Love and hate take p e ~ mobjects. or make a comparison with yes and no when they consider will and desire. 9 1-91 Aristotle's orexi. that there is no such connection. what qualities things don't have. contempt. Turning back to will and emotion. belief that he was alive positive. But since the passions are certainly causes of pursuit and avoidance this doesn't seem sound. joy. if I say what 'will' is not. this doesn't mean that wants must be consistent . cheer. And that is the reason for the 'negativeness' of hate. Unless it means that he tends . Though.all these will readily be counted 'negative'. l hating a person. seems a pretty positive thing when considered in believe what he is told rather than to disbelieve it. pursuit and avoidanceare in desire. must you want it not to fail to happen:' Here the answer is positive. or does it necessarily involve pain. curiosity. gladness on anyone's behalf. And it might be complained against me. But perhaps this is no more than a 'taking as' which comes naturally to everyone? Hatred. orexis. and another the opposite. but when offered the choice: how will you distribute the terms "positive" and "negative" between hatred and love.nothing is said about whether you can also want it to fail to happen. as tending in the positive direction: he says what exists. not noticed this aspect. we can get what concessions we like out of him. Thus we might make a classification of certain pairs of contraries as ones which exhaust the possibilities for their subjects. pleasure. and only interested in what is the case. rather than what doesn't exist. one character may be fearful more than hopeful. envy (in the sense of disliking another's gain or good). then one will retreat and call hatred the negative emotion. Therefore to say that someone is characterized by believingness rather than disbelievingness is to say nothing. Love. Why? Mephistopheles says in Fawt: "Ich bin der Geist der stets verneint" . the answer is yes. when they exist and are capable of having the predicates hold of them. But does it mean nothing to say someone's belief attitudes are positive rather than negative? Well. if we think of that spirit as the spirit of destruction of ~ositive things. must you be unwilling that it should not happen? Obviously not. But disbelief would not as such be negative. as Ceach says. So we can after all put him with Augustine and Brentano here. Nevertheless. Thus it does make sense to speak of a man's opinions. depression. the idea of positive attitudes. In this way belief that someone was dead could be called negative. in response to a negative. of his belief. hopelessness. by the way. dislike and spite. gloom.

I Will and Emotion 1 ° 3 . and "ripening" in the decision of the will: Aber ligt nicht demungcachtet schon in dm Sehnsucht ein Keim des Strebens? und spriebt diaer nicht auf in der Hoffnung. "I'd just - - - - . involves desires which do. Brentano. the need to do so would never arise. but if anyone did assimilate soap and washing I'd want to oppose it. The voluntariness of the voluntary act doesn't consist in anything of the sort. which doesn't 'take' propositional objects. whatever inner went precedes an act. sleepiness. That an act is voluntary doesn't mean that it is preceded by an act of will. something in the line of feelings (the 'Mut') which he also finds present." It is a point that somewhat embarrasses him. I need not deny one similarity at all I mean the similarity in language. Nor even is thirst a feeling that it would be lovely to drink. is a feeling of being liable to throw up soon. then they cannot be explained to us by pointing to the emotions. to bring it about that he not be ill. The genetic explanation by reference to familiar objects of experience: "You know what fear and hope. bis endlich das Verhgen danach zugleich die Scheu vor jedem Opfer und den Wunsch jeder lllngcren Erwagung Dberwiegt und w zum Willensenwchu0 gereift ist? Once again. I have made up my mind to catch a certain train and I leave in time (or not quite in time). It is in fact as odd as identifying hunger with the voluntariness of eating. you may compare pursuit and avoidance to affirmation and negation. ~ h ism another reason for the error of psychological confusion here. and it is useless to look here for the explanation of negativeness in will arid emotion. I don't deny this similarity. 'This point is by itself enough to show Brentano radically wrong in his explanation of the ideas of good and evil.104 Medieval and Modcrn Philosophers I o the 'Geist der stets verneint' thinks that be destroyed or diminished. So long as you stick to propositions in gencral (or other possible objects of judgement) as giving the objects of emotion and will. Not to such ordinary examples as the following ones: I feel inclined to shut the window and 1 do shut it. that admixture of reasons and thoughts which is so characteristic of human emotion. being on edge. Brentano knows quite well that will is not a feeling: "da6 er einen Willensentschlu6 filhle wird wohl keiner sagen.) If we have to use them to differentiate emotions from psycho-somatic sensations. takes steps. feeling full. of elements in a flow of feelings with the voluntariness of an act. o n c ~ w e curnine the situation in detail. or pursuit and avoidance. Neither of them would have been the will in the voluntary act if the act had been voluntary. one can still ask i f it was voluntary when it occurred. We all know it h difficult to find the evem which shall be the act of will within a voluntary action. Nausea.g.this won't work because we will already have to mention good and evil in explaining what we meant by the words for the emotions. And the logical similarity is in the language connected with e. und mtfaltet sich. for example. And s 'alles was entsteht ist wert daf3 es zu Grunde geht' that comes into being deserves to be abolished. the ideas that can be got from having all of these in your repertory are the ideas of good and mil" . and that he forgets from time to time. thirst. Now I arn in opposition to Brentarlo in respect of his assimilation. But it is equally difficult to find the evmt which is the feeling. itches. love and hate are. In spite of this. But why not? The answer is that these sensations don't involve reference to good and evil. as when he says that inner experience shows that there is nowhere a sharp boundary between feeling and will. rather I energetically draw attention to it. One doesn't want to call these "emotion". In this resides the sum of correctness in Brentano's assimilation of the two kinds of thing. The physical event was almost the same. but you arc going to get pursuit and avoidance as equivalent to one another according as the same matter is itself presented negatively or positively. consider how. I would quite radically distinguish will and emotion and I say that Breritano assimilated them because he didn't realize how unlike they are. nausea. At least. feeling inert. That suggests that he sees or wants to see an act of will as itself a content of consciousnessand think5 that feeling merges into it. don't you? Well. You aren't going to get any real contrast of positive and negative. one wills. And he speaks of a "seed" of striving as already there in the feeling of yearning. For it's a necessary and useful point for helping to distinguish between emotions and complex bodily sensations such as dizziness. I'd want to distinguish soap from washing. but that it is itselfan act ofwill. the emotion. who assumes it does. In proofof this. "unfolding" in wishing and in getting one's courage up. weariness. emotion and will. finds bad that someone is ill. (If I have understood him. bei dem Gedanken an ein etwaiga eigenes Zutun. Crouching down on the edge of the swimming bath I hadjust nerved myself (Brentano's 'Mut') and positively determined to roll head first into the water suddenly you pushed me. But the whole idea is an error. it is not a feeling that it would be good or lovely to throw up. in dem Wunschezu handeln und in d m Mute dam. The roots of a real polar opposition of emotions are surely to be sought here. a confusion of radically different kinds of thing. desire on the one hand and will on the other.wen though one might give expression to it by saying so and thereby become emotional about it. you may indeed say that there is yes and no in them. One regrets. The idea of neighbouring members of a series which can hardly be distinguished from one another seems quite inappropriate to these cases.@& something introspectively almost indistinguishable from something else. or putting them in the same class. of this "seed" as "sprouting" in hope. Yet it is precisely here that there is a logical similarity between will and einotion. Even emotion such as love. 1 think. What matter that I find the nerving ttyseyand the &cision extremely alike? Neither of thein was the will in the voluntary act of rolling into the water for ex hfidhesi there wasn't any such voluntary act. we seem to have our attention directed to a very special sort of example. Indeed that was rather suggested by his spectrum.

106 Medieval and Modern Philosophers Will and Emotion 107 nerved myself. or when one stopped. that he noticed the elusiveness of the feelings themselves.i. I don't think.anger had taken away my inclination to make it. That is why 1 say that this sensation was a sensation belonging to my being angry. but because it is done by me either for its own sake or for the sake of something else. e. What was that? Was there a feeling of anger or fright which occurred at a particular time? Well. These things elude one when one approaches the matter with a certain expectation of what one will be able to find. Also the emotion is mentioned as a reason or 'motive' for some actions. It would be of interest to discuss the causality .or (b) to have a certain intention .but they can't be the anger or and that this had certain consequences. will is far more different in kind from the'emotions than he ever even conceived as a position for his opponents to take up. - . as well as a cause of others. that intelligence grasps what conduces to what and what the situation is in which it operates. because .e. however. a thought? "I was very or "very frightened". daR das Verm6gen das Wahre vorzustellen. werden. The last is the important case for us .the others are side issues.g. drawing in of the breath.or fear?''. das Gefiihl sei. But at least the effect is of a quite diff'erent kind from the cause: the effect is a voluntary action taking place no doubt at a definite time. But the contention fails.i. 1 have an intention of returning to England this month. that it frames the conceptions of those generic (right or wrong) ends which are characteristic of human beings. This new dimension of 'What for?' enters into the description of the act and belongs to the intelligence of the agent. some of which can be called effects. die Erkenntnis. whereas a tickle that suddenly perhaps attacked my nose at the same time had nothing to do with it. to the effect that only now are we at last realizing that. Thus he quotes with approval a remark. It is difficult for Brentano not to turn out to be an emotivist. This point of similarity. how many different types there are here. 1 mean that e. if you mean that at some particular time it was true that I was angry (frightened). though it may be certain that one did fear it at a certain given date. Lack of space prevents this.because I was so angry.for I have deliberately chosen what might easily be either. "How did you know that was a sensation of anger . though sometimes possible. It belongs to intelligence in two ways: one. for one can ask someone "How does (or did) anger (or fear) take you? Where did you feel it? In your chest? In your head? In your legs?" Now suppose Eomeone says: With me then it was a constriction in the chest and a trembling in my knees. The states ofexnotion. Brentano himself implicitly uses it: he keeps on speaking of the different colouring of the different enlotions .e. undoubtedly cause both voluntary and involuntary actions. but it takes intelligence of the human sort to be an aholastos in Aristotle's sense and make pleasure in general one's goal. the cause. anger inspired it. other animals may be dominated by an appetite for pleasure. There need be no answer to the question when one began to fear something. und daB beide ja nicht miteinander rnilssen ve~wechselt So we see the drive behind his contention of the identity in kind between will and the emotions: it doesn't belong to the intelligence to frame the ideas of' good and evil. dasjenige aber. and two. das Cute zu empfinden. which is quite like the elusiveness of the act ofwill.leaving one to understand perhaps that the will has yet another colour." we say: but what war that? A certain tension of the muscles.but no such thing as a distinct mental act is generally necessary when one acts voluntarily . is by no means necessary. would be no ground for an assimilation. is void for uncertainty. as knowledge (Erhenntnis)is the source of the power of imagining what is true. whether or not they are states of actual excitation. I wrote that letter because I was angry. I abandoned a ~ r o p o s e d outing because I was angry. But what was the feeling of anger or fright itself? One may find certain physical sensations . For he thinks the source of the ideas of good and bad is ~ u r e l experiences y of love and hate (taken very broadly). so feeling is the source of the power of experiencing what is good: Man hat es in unseren Tagen allererst einzusehen angefangen. yes. A positive act of mine is voluntary. and have had it all along without thinking of it (otherwisethan by booking the passage: certainly that is nothing of the same kind as a feeling)or (c)to try to do something (which is usually to do something elre) or (dl to act voluntarily. not because it is accompanied o r preceded by an act of will.g. a state which lacks a central core and the assignment of which to a definite time. At this point one wants to say: The feeling of anger (or fear) suffuses the physical sensations and the reactions in thought and action. And this metaphor of suffusion is a very powerful one. To will is either (a) to make some decision . however.of Kant's in Untersuchungcn tlber die Deudicfit dcr Gsundsiftrc der natiirlich Thologic und Moral. Examples: 1 upset the coffee involuntarily.

g. 1931). names not. say. 1959). Mind. in Stenius' phrase. if an elementaryproposition muld be reached in a particub ou. symbolizes an elementaryproposition.In Wittgenstein. Oddly. Frege had distinguished concept and object. the reference of predicates. So far as concerned the meaning of a functional could be regarded as the result of completing a function "f(x)" by "b" as giving a Value of "x". but equally Socrates has many things true of hirn. Names ofpredicates are more like 'arrows' because their meaning is 'directed"' (p. 9 ( I 965). This is Wittgenstein's usage in 4. F. This expression is mine. not Wittgenstein's. rather. An Inhoducfion to Wiltgnutein'J TrMlDtus (London. 144that (all)names resemble 'points'. unlike names. These errors can be avoided. 1960). i. This. in combination. 1 still think Stenius quite wrong about this. and Thus. reversible polarity. concepts. ~ u t nohing to say - - . Referring to theway in which predicate-expressions. more than one name. He said that both parts of the proposition name and predicate-expression stood for incomplete objects and that there was no essential difference between universals and particulars. say "a-b" were a specifiable elementary proposition. however. e. or universals are not just more objects. Properties. yields a new predicate. waiting to be filled by an object. but in respect of having argument-places. I hold. but he says it is a term used for what he calls "predicates" (relations and properties). and Irving Copi and myself. then there will be predicates ofdifferent type-monadic. - From Anab~is. It is this logical network alone that is 'universal' and in its connection arises the possibility of negation. as Wittgenstein himself later put it. which he later used to characterize the Tractatus position. a proposition in which a non-relational property was ascribed to a subject. a concept had as it were a hole in it. whose Erik Stenius. f and b. etc. There is an ambiguity of the word "predicate" that needs clearing up.s6. 175). as Ramsey took it. atlother. concepts go over entirely into logical forms. This view still appears to me to be correct.' Professor Stenius does indeed cast sorne doubt on the suitability ofthe traditional tern1 "universals". concepts. This leads 11in1 to a really monstrous piece of exegesis: he simply overrides Wittgenstein's statement that names resernble points in that they lack Sinn. mysteriously. Predicate-qressions. I mean that negation. This fact has always made me uneasy over exegetes' taking sides on this question: Erik Stenius giving one answer. I myself came to think that there could be no such fact though. were essentially 'unsaturated'. This prohibition worried me. E. Wit~gmrtrin's Tractalus (Oxford. while Riamsey's essential idea is kept. 'Objects. or a diffmnt function "g (x)" by "a". i. logical sum. is not the way in which "predicate" was used in early Russell. too. but when attached to a name it does not yield any name. This still seems certainly right. this was simply objects. but I now think that Ramsey was righter than I ever realized.12 Retractation Here is a striking fact about the Tractatw: the question whether the objecls spoken of there are individuals only or are also universals seems not to have crossed Wittgenstein's mind. 67 (1958). Ranasey. the T r U u forbids one to say s o . "a logical network sprinkled with names".individuals and predicates. Now if predicates are what is symin the Tro~tatu bolized by the remainder of a proposition when one or more names are deleted. he says "Thus Wittgenstein is wrong when he says in g.8 Ranlsey disputed the distinction between individuals and universals on the ground that there was no less reason to speak of Socrates as attaching to wisdom. but not necessarily one in whose sense d y two objects occur. What Ramsey failed to see was the distinction between the sign of a function and a name. logical product. P. dyadic. directed meaning. And he insists that there are these two quite different kinds of 'ot~jects' . one of the notations for an elementary proposition in the Tractatus. as I put it. the one-in-many theme does not come out in only one way as between Socrates and wisdom. as Stenius takes it. 'Univenalr' in The Fmmdatimu ofMathmtotic~ tondon. are not names. "a-b" were the elementary proposition fully analysed into a collcltemtion of n a m e s .)" would be quivaknt m "efl. H m "f" would not be another of a. Protesting that "f (b)" need not for Wittgenstein represent only a twoobject atomic fact. or "g" h e r na. I exposition he found in Russell. by that phrase "a logical network sprinkled with names". "f(b)".of b. Wisdorn is fbund in many objects. rhou& bob "f(b)" md "g(. nor yet another sort of object. but I see no reason to insist that a functional expression always 'covers'. say.e. he seems to have thought of "f(b)". than to speak of wisdom as attaching to Socrates. but there is nothing wainst a functional expression "f" covering oniy one name. attached to a predicate. I said. I maintained. he did not perceive that what sipally distinqishes names fi-omexpressions for predicates is that expressions for predicates can be negated.G . etc. any predicate would have to cover a plurality of names. negatability or. "f(b)". predicates. which Wittgenstein chooses to write as. for example.e. for at that time it certainly appeared to me that one must say there could not be. is 74 : "one cannot ask: are there unanalysable subject-predicatepropositions?" Hence he implies here that one cannot ask whether there are possible atomic facts containingjust two objects. . however. Ramsey objected to this idea. the completely analysed proposition is. Irving Copi. I complained that Ramsey wrote as if. There a 'subject-predicateproposition' was a non-relational one.Anscornbe. Wittgenstein made the gulf between concept and object far greater than Frege did. M. but I think it is justified. as if the possibleatomic fact it would symbolize must consist of two objects. Both could exist only in facts. admit of negation. and Relations in the Tractutw'. and the only objects admitted are the bearers of names.

or the objects constituting them will be universals.5 indeed the contrast. It should be dear now. if I am right. I did not appreciate the force of his argument that they were no more 'universal' than the hearers of the names of the Tractatus theory. so why should w the objects into which red is analysed? This is perhaps a proper reply. according to him. between name and function. (EvenifA and B are different and widely separated things. ' See Edwin Allaire.or simply the 'formal mode' forthe contrast between individual and universal. Does that make them into universals? Well.individual things which are the bearers of singular names and universal things which are the bearers of general names . and it appears to be correct.) Denying this. corresponds to a function. . That is to say. I thought that in spite of all he said those objects which. "Red" and "green" must in any case each cover a pluralitj of objects. in logical configuration if they are plural. nothing in reality except the objects themselves. 'Trmlatlu. for as "material ~ r o ~ e r t i e they s " are "first formed by a configuration" (s.~ I went on to say that this Ramseyesque view was not Wittgenstein's: "it is only the logical network that is 'universal' ". that from the Tractutus point of view the distinction between individuals and universals . - ' An Introdudron to Wittgnrrtn'n'J Tractatus. not to any sort of object. the latter. if red is composite. because Ramsey failed to distinguish name and function. then equally A and B. But I had no evidence for saying the view was not Wittgenstein's: I mistakenly thought this followed from the fact that "only the 'logical network' was universal". Now let us consider the objects covered by 'red' and 'green' in "A is red". and I thought it was the same as. 6. That is Ramsey's argument. "Universal" in this sense is a term applicable only to meaningless. because the objects must be indiuidualr. I think Ramsey would have supposed they would be the same. be 'universals'. This contrast is disregarded by both Stenius and Ramsey.opgi). It appears to me that the point is of far more than exegetical interest. "A is green" and "B is red".or. log. There is no ground for a distinction of types of object. But if the features of negatability and of having argument places are what mark the expressions for universals. yet it is difficult not to feel that an object that can exist all over the world in different facts has rather the character of a univer~al. This solves the problems raised by the incompatibility of the two propositions. were represented by functions would have to. of into a variety of facts. strongly insisted on in the Tractatus. if it d o e so simply because they are found in a nexus with A in the one case and B in the other. 19 (rggg). I have also always thought that exactly the same objects were involved in the sense of "A is red" and "A is green" : the same objects in a different configuration.110 Medieval and Modem Philosophers Retractation 111 as to whether there are or are not such elementary propositions. The concept of a universal is a bastard progeny of two quite distinct concepts those of function and of the existence of an object in many f & . I wrote of it as a purely Ramseyesque view: The question arises whether the objects that would be named in place of our using the colour-word "red" in the two cases.' Howwer I did think that the objects covered by "red" in "A is red" and '3 is red" would be different where A and B were different. The former is linguistic. and ordinary language uses propositions of the subject-predicateform without enquiry into their analysability. But suppose after all the objects covered by "red" in "A is red" and "B is red" are the same. The fact that we must contrast name and function does not in the least oblige us to deny that the objects (whateverthey are) covered by "red" in "A is red" could a150 enter into the fact that B is red. But for the moment let this suffice. not peculiar to these objects which are involved in expressions for properties and relations. then these objects will not be universals. And ' functions are purely linguistic. There i. Thus far exegesis.3751'~ Analysis. and so failed to take Ramsey's point. p. And no doubt he would have poohpoohed the feeling that in that case these objects would have the character of universals rather than 'individuals': w e don't think A is a universal because it can enter e think this of red. but there is no place in logical enquiry for the question whether functions might or could not 'cover' single names. wen if they do exist separately all over the world at once. would be different. This point is of course independent of the other. are merely noting a convention" (RFM. If we assent to "Essence is express4 by grammar" we may very likely say "The words for what I am talking about have to have this grammar. But what I mean by "linguistic idealism" would go further and say "Essence is created by grammar". t h e had not been that linguistic practice with a word. H. so that the two together are said to make up the red. For the property mentioned is a property of language. In their sensibly perceptible aspects those words differ from oIie another. 11." E. . I call "linguistic idealism". I . XII). For the 'essential' is "the mark of a concept. horse. Wittgenstein appears to suggest that what 'corresponds to' our concept of colour is not (aswe might have wished to say)a distinct feature of things. certain similarities of colour. and whole in every part". But now: if we say the essence expressed by the grammar of the word "red" is itself the creation or product of that grammar aren't we saying that nothing mwld have been red if there had not been human language? If. The first is like the suggestion made by Plato in the Cratylw.The Question of Linguistic Idealism 113 The Question of Linguistic Idealism "If anyone believes that certain concepts are absolutely the right ones. ' - t The language for talking about sensation must have first-third person my mind at least . cheval. tnnos . which we will call its logical shape. as that the word for something is inflected.the question: have we in his last philosphical thought what rlligtlt be called a linguistic idealism? Linguistic. von Wright'. And that we do not want to say. ) 980) What we want to be assured of is that 'what we realize' actually exists and is not a mere projection of the forms of our thinking upon reality. is not the suggestion that the essence expressed by the grammar of colour-words is created by that grammar. Now language is certainly at least a complicated set of proceedings. sensation). 73 (any edition). But there is a crude grammar common to all.g. But here we don't mean: This property (say. whatever its merits. 74). but rather the 'very general fact of nature' that colour and shape are independent. equus. if it applies through and through. brass. as may happen in one language and not in another~(as: in English there is or was a call "To horse!") . say iron. it would cease to be languageabout sensations. this grammar. and that having different concepts would mean not realizing something that we realize . or that certain phrases containing the word are compounded in certain ways. 'This logical shape is the gramrrlar of the words. bronze. even. So a word has something. Suppose we accept this last. people might have some colour-shape concepts. its real essence. words for the sanie thing ia different languages . and none of colour or of shape.not a matter of the fine details that would be described by a commentator on Panini. was that it was "whole in the whole. more particuhrly. Going back to my opening quotation: someone pight say "I don't want to say that such-and-such concepts are absolutely the right ones. : Rrmmb a the Foundalionr ojMathnnolicr IRFM]. whose existence shows that the users of the word have the concept 'red'. 1-3. "If you talk about essence . we might accept this dictum in the following way: as Plato suggested in the Cratylus. From 'Essays on Wittgenstcin in honour of G . This isn't a matter of the most superficial grammatical features. And he also wrote: "Essence is expressed by grammar" (PI. and do not here pause to show that the fixing of a word as such a word is done by its grammar.e. For do we ourselves have a concept wherever we notice a similarity? (SeeZettel. appears in another place. the second. that they 'fail to realize what we realize' . So now it looks as if either the grammar corresponded to something of the object. So we should rather say: Language that doesn't have these features. We don't have to suppose." And these are grammatical remarks. Acta ~hilarophicaFennica 88. mostly with articulate noises. I. Yet it does head one off from an attempt to concentrate on and state the essence that the grammar expresses. steel. $971). and the formation of concepts difl'erent from usual ories will become intelligible to him" (PhilosophicnlInvestigations [Pll. This is one of the passages from Wittgenstein arousing . remarks about the way in which words are used. Now. is not about sensations but about something else. in a world where colour and shape were not independent." Or: "Nothing is called a part of red. for example. I . for example. and if you took language about sensations and changed this aspect of it. answering to the essence that it catches hold of (or expresses). because he describes concepts in terms of linguistic practices. their logicai shape is tile same.g.e. by which each is e.of first-third person asymmetry) is rightly ascribed to this kind of thing (say. but nothing dividing up its colour. or the 'object' were itself dependent on language. This suggestion. say: "We call something dividing up a square."Mark of a concept" is of course an expression taken from Frege. I only want to h o w if they are right ones at all. not the property of an object".g. 1. which it has whether there is language about it or not. (1976). also not red.are like the sanie tool made of different materials. This would appear in Wittgenstein as. a count-noun which is the name of a kind of whole living thing. A scholastic statement about colour. while another part.then let him imagine certain very general facts of nature to be different from what we are used to." To take Wittgenstein's own example. which appears in one place and is not red. A tool which is designed to catch hold of something will perhaps have a shape corresponding to the shape oi'the object.

when counted wrong. "Pointing twice to the same" is an expression that docs not yet deter - . having a count-noun. we are talking our language. However.although there were horses around? That might be difficult. But we can cmeive Ofdflerent concepts. . and the reports be differently treated according as they are first-hand of things currently present to the reporter o r as they are second-hand or only remembered. there would have been pain?" Nothing. which he identifies with a 'nominal essence' and calls a creature of the understanding. rather. read. if the word is to be the word for that experience? But the word is not just a response to that experience at that time: what else is the word to apply to? The experience can't dictate what is to be put together with it. for example.e. say the one described?" Well.whether it is to be an exclamation or to have a syntax.pain according to our concept or according to some other possible concept. That is to say. that is not the slightest reason to say there wouldn't have been horses. they might apply it to cows on seeing them. and if the latter. the essence expressed by "horse" would not have existed but fbr human . that there were not a great variety of species of mammals. But those terms don't signify a mental act such as forming an image or having a representation before me. P e q k using this would then not be using language whose grammar expressed the same essences. from which. and hence that I mean that. I suggested saying that language about sensations had to have a certain grammar had to have the first-third person asymmetry. and there were no count-noun corresponding to "horse". e. The essence is not what I mean or am speaking of: it is rather that through which I understand or think of (mean) etc. These essences. i. To sum up : Essence is expressed by grammar. "The answer to that is surely: Whatever is the same as it. it is not experiencing pain that gives f the word "pain". I t would. No image or representation could determine future or past application of the word. ti>rexample. though it overlaps with ours sufficiently to call it that. are not created by grammar. he supposes it too is a horse. for example.~ we follow the suggestion of my opening language and t h o ~ g h t Could quotation and imagine 'general facts of nature' so different that people did not have the concept 'horse' . l'hus a tribe can be imagined with a different concept of pain. be an open question whether they did or not. Locke's error was first to think that the 'idea' was the object of thinking. it is that because of which my use of the word is a case of meaning a kind of animal. that though there doubtless would have been horses. This is where he answers the question. This is determined by the grammar's expressing an essence. if there never had been humans around talking about horses. t h q might not thereby be missing anything that we realize. was what a word stands for and so what we mean when we use it. then. suppose that they 'fail to realize something that we realize'. we might not tind here the concept "horse" and yet we might not "uch an error was committed by Lockc:according to him essences are general ideas. to pain without wounds) by saying "Do we make a concept wherever we see a similarity:"' Similarly he says "You learned theconcept pain when you learned language. But when Wittgenstein deprecates thinking that people with different concepts must be missing something that we realize. So it is pain as we mean "pain" that we are saying would have existed anyway. "But don't they notice the similarity:'" (sc. which are expressed by grammar. and second to think that when we speak of 'a kind ofanimal' we must mean a general idea. are treated as personal stupidity. i. fbr example. then.e.1 14 Medieval and Modern Philosophers The Question offinguistic Idealism "5 Wittgenstein imagined a tribe with a different concept of pain from ours: their concept applies only where there is visible damage. if asked to explain what the word "horse" signifies. or say "horse". he presumably doesn't mean that they could not be supposed to miss something that we realize. so may not language similarly relate to sensations without expressing or reporting them? Sensible properties. How could an experience dictate the you the meaning o grammar of a word? You may say: doesn't it make certain demands on the grammar." But do yo11 know what sameness that is before you have the word? The experience at any rate does not tell you the limits of use of the word or what sort ol instrurnerlt it is to be . if there were animals. may be mentioned and scenes described.still. seeing a donkey." -What sameness have you in mind? You will say: "The sameness that is expressed by the word. should I not then ' give up my idea about sensations? It seemed that language may at any rate relate to horses without having a word for a horse. he might say "But isn't it the same as you pointed to before?" showing that the identity in question is identity of kind. O r again. he will grasp that the word is the name of a kind. but some verb form signifying something like horsepresence. I understand or mean or think of a kind of animal when I hear. Suppose. If."But which 'pain' are we talking about . Now what is there in all this to make a dilIiculty about saying: "Even if there had never been any human language so that there was 'no concept of pain' at all . but only horses and humans. if our hearer understands our explanation. But am I not saying here that language about horses doesn't have to have the grammar that expresses the essence. And similarly. I am master of this grammar: it is by that grammar's expressing an essence that the word I am using means a kind of animal. but there are no words for the sense modalities and the words for the sensible properties have no purely 'subjective' use: 1 mean no use in which they are not subject to correction by other people.g. whether that of a property of objects perceived. such as a wound. It is enormously difficult to steer in the narrow channel here: to avoid the falsehoods of idealism and the stupidities of empiricist realism. The first-hand reports of sensible properties. horse. of language without the same grammar. Whereas. I t must be a misunderstanding of 'essence' to think otherwise: to think. doesn't have to contain a count-noun which is the name of a kind of animal? It seems that I am. we may point to a horse saying "That is a horse" and to another one saying "So is that". and general and ~~r~iversal are crrations of the mind. what I and others have called and will call a "horse"." That is.

it sounds as if the former expression introduced sorne new objects to count over and above horses a l d giraffesand men.llly separate letters Letters of the alphabet of distinct type fount? Type founts? Etc. M. though they didn't invent a. before there were humans with their linguistic practices. And it is cotnmon to otherwise dissimilar words of different languages. by which horses are spoken and so also thought of by means of the word. the question still is "same what in ice. and not in either case a 'creature of the mind'. of weighing. I suppose. If anything is the creation of the mind it is the grarnmar of one's expressions. as if some new reality were being ~ostulated. C d r i d p Mar. As indeed it is. But so powerful is the suggestiorlof the superficial forms of that when we speak of' 'counting kinds of animal' as opposed to 'counting horses' (for note that the former expressio~i is ambiguous). "Linguistic practice" here does not mean merely the production of words properly arranged into sentences on occasions which we vaguely call "suitable".T. ' See Kui* M 19%. the response to and joining in various other specially human activities. it is that about the grammar of the word. Nor is it the object ofthinking. Humans would not. they invented They invented doing that with the ideas of o w and of (there was no similar representation of a half in the Roman numemi s~tem')and thdinvented treating it as a representative of half ofanything. We shall ask the question: Does this existence. that human possession of concepts is so dependent is not quite so evident. you might my they +. If what we mean by "idea" is the essence expressed by that grammar. . And though they didn't invent halves. for example.. a visible. or this truth. For example. When the view is introduced that counting kinds is counting a different sort of objects from counting individuals (for after all. if the ~rocedure LL~e counted ' ~ e that one". activities of measuring. it is difficult to see how anything could be identified as.There is only one letter e in the Roman alphabet. NwLrr N dS * . language with a syntactical organization or even in any conventional system of signs may have dealings with his fellow-humans which clearly involve his making and complying with varied requests. of moving about in a huge variety of ways. With the method is associated a concept C which is used to say 'what' one is counting. Why then are thcg more of a human invention than natural numbers? Did he mean: the natural numbers are (as it were)given: we could do what we liked with them. It is plausible to say that we would have no concept of lGngth apart from some activity of measuring. a he competent use of language is a criterion for the possession of the concepts symbolized in it. But the same sort of thing holds for facts involving fractions. and no concept of precise comparative lengths of distant objects if the activity of measuring had not a quite elaborate use of words internoven into it. and so we are at liberty to say: to have such-andsuch linguistic practices is to have such-and-such concepts.a series of natural numbers. is relevant. of giving and receiving and putting into special places. acted joke. or days. number?" or "same for purposes of counting how?" .The rock-bottom explanation of' counting each of' these is the counting procedures themselves. a giraffe(a $raff'e not having been counted before). ~t is as with letters on a page. depend upon human linguistic practice? That the rnedng ofexpressions is so dependent is evident.116 Medieval and Modern Philosophers The Question of Linguistic Idealism 117 mine what counts as that: the question only has a determinate answer when we know what identity. So the natural numbers were no more a human invention than wolves. the earning and spending of money. A deaf-mute untrained in a . On the other hand we take it as obvious that there is a host of concepts and considerations which are quite inaccessible to him. when Kronecker said "God made the whole numbers. Any e is numerically the same letter 4 t h ~~m alphabet as any other. what method of counting. and mathematicaltruths are human inventions? Humans invented such a thing as P x 9 . (What do I mean by this? I suppose something like the fallowing: if there isn't a practice of counting objects. the granlmatical object of the verb 'to count' is ciifferer~t!) then of course the question arises: where are some of these objects to be found? In the realm of Platonic forms? In the mind? ~t is asto1lishing how deeply people are sunk in this bog. for if you say "Same in number". all else is human construction" what can we take him to mean? At first sight. for example. but to assign a new number to.then it appears that one is counting kinds. into which a use of language is interwoven. One will not understand Wittgenstein well unless one notices that he has not anywhere SO much as got his feet muddied by it. Whereas we have only introduced a different way of counting what is before us. then in the case of the idea of a horse that is no creature of the mind. of consulting tables and calendars and signs and acting in a way which is connected with that consultation. has two e*si and r]. say. Somewhere in one of his notebooks he Wrote: "Nulnerical identity: a bad concept". which is created by pure grammatical misunderstanding. .I. P . They invented treating half as a number. for example. Told to 'count these' one assumes or needs to be told some particular method of counting. A horse has is to say been counted and another horse comes along. for purposes of counting".) But this fact does not dispose us to think that there could have been no such event as that a wolf killed three deer in seven days. have had the concept of the natural numbers if they had not had the practice of counting objects. But what one is counting is in any caie out there before one. The fbregoing considerations lead to the following test. it is not surprising to learn that such a man can make or laugh at a joke. also it is fruitless to speculatehow much of his capacity is dependent on the existence of the language-using life of those he lives among. I t is important that it includes activities other than the production of language. deer. Greek. ifwe want to know whether Wittgenstein is a 'linguistic idealist'. little. Separate marks? Physir. on the other hand.

This was one of Humr's greatest observations: he called it the "natural unintelligibility" of promises. he 'literally' has to. then under the contract you 'have to' do this. We have to draw attention to the use of modals: "must". it is accompanied by d i n g the learner do what he 'has to'. or descriptions which look like descriptions of states of aff airs. With each of these there is associated a certain use of modal notions. i t is evident that they contain descriptions of conduct. the peculiar philosophical problems that arise about rules and promises would not arise. how? Hume's discussion is primarily in terms of the 'act of mind' supposedly expressed by a promise. Or they may apparently be 'permissive'.Probably this is the right direction to look in. the problem of meaning. We have agreed (with small reservations) that the existence of human concepts can be sorr~ewhat generally equated with the existence of a great varict): of human linguistic practices. a great many things whose existet~ce does depend on human linguistic practice. Both a rule and a contract. rituals. rights and promises. and from seeing that here linguistic practice creates the essence which grammar touches the nerve of great philowphical problems. he is complelled. if he reacts to the training with human intelligence. Thus at first. does the supposed obligation come about? And if so. Now it may seem to be a question how there are two distinct problems here. you 'have to' do certain things. of course. But this mandatory or permissive 9tyle does not signify that someone is ordering or giving permission to do anything. or of any other procedure. that is. the second question would still arise. positive and negative. as we might say. the "you have to" itself becomes an instrument of getting him to d o such things.or else does not. ceremonious promdine. The significance of the promise is the imposition of a kind of necessity upon him to make the description come true. For let us suppose that instead of promising. The dependence is in many cases an unproblematic and trivial fact. you 'can't' stop him. And similarly. their mainverb being modilied by 'may'. One niay then become equally puzzled about what they are. even ifper impossibile the promise was after all 'naturally intelligible'. Ile gives signs of understanding and he complies or disobeys. it might be thought that if for the sake of the argument it was granted that there could be such a meaning. 'The case ofthe promise or contract is different. so that the first problem of its meaningfulness was solved. There is a game played with children in which people's hands are successively laid on top of one another. For it can be derived purely from consideration of the meaning problem that he uncovered. then the obligation is generated by the utterance of a sign with the meaning. Another is associated with giving details of 'how to' d o various sorts of work where others are to learn from y w r orplanation. etc. That situation we understand. of how any 'obligation' could be generated.That of breaking a promise! Making the promise consisted in giving a sign of generating a new injustice: one of not making a description come true which one was otherwise at perfect liberty not to make come true! That is the meaning of a promiw. The rightness of this conclusion is independent both of Hume's psychology and of his theory of the foundation of morals in a peculiar sentiment. have to". But that. But what injustice? . consider the utterance "You have to move your king". But in others it is not trivial . These descriptions may occur in the present or future trnsr c)r in a peculiar mood which we may call the mandatory future. one may become puzzled at how they can generate necessities. you 'cannot' do others. But if I d o say that. If moralists have found this offensive. something is to be an offence against its meaning? Its meaning has not properly been explained. The one whose hand is at thebottom 'has to' pull it out and lay it on top. or again 'cannot' do that. and second. are themselves either possibly or even necessarily linguistic instruments. Because ol'the rulrs of a game. But what can this necessity be? We may say: it is of either making the description come true. This is indeed equivalent to "The rules of the - - ----- . Another with more or kss e h b o n t e p d o r mantes of a m t variety of kinds games. - 1 Hume divided the problem up into two: first. If one stares at such instruments in a spirit of philosophic enquiry. we have remarked. For since the first question was how the meaning of the sign of promising could be that by giving it an obligation was generated. this will have been by misunderstanding it. And now this leads us on to the possibility ofwhat we might call a partial idealism. i." That of course is perfectly intelligible. I t is notjust a question of a description which someone then makes come true . But Hume is right. . etiquette. But then. by no means inlplies any dependence on human thought and language. If you have made a contract. When someone has a right to do something. or being guilty ofan injustice. His chapter is celebrated and has led to a large literature on the substantive question: what account to give of promises and the obligation arising from them. the two questions are separate. Hume's own conclusion was "That promises have no force antecedent to human conventions". even if we grant there can be such a meaning. such as the membership o f a committee. but the essentials of his argument do not depend on that and are reproduced in such an exposition as I have given. the second question is already ' answered: if the meaning is that an obligation is generated. Someone is told to do something by someone else. since it lacks the peculiar character of the promise by which the sign itself is supposed to be a profession of generating the constraint (obligation). This is one of the primitive uses of "you have to". But there arc. 1 say "Let there be a constraint upon me to act thus and so.. But how cun the meaning of a sign be that as a result of its being given. said perhaps to a learner in chess. on the part of the things that fall under the concepts. One such use of "you have to" is associated with 'keeping your word'. The cases i have in mind arc three: namely rules.118 Medieval and Modern Philosophers The Question of Linguistic Idealism 119 invented half's being a number. o f coursr. which is far more extensive than has usually been considered in philosophy.e. (in some circumstances) making your word come true. Were that the situation. it will be a question whether there then does come about any such constraint.

You can easily move another piece instead. which are accomplished by the art of linguistic practice . In the dance an incorrect step means that you aren't doing what you are trying to do. and there isn't a distinct meaning for "being a move in the game" (unlessby analogy from other such games) which explains or logically implies the "You have to" which come5 into the learning. like that of 'a relation'. VII. the point excites no anxiety. this existence. Learning the game. on the grounds that anything supposed to be known is supposed to be revisable. whose meaning is that one must act in accordance with it. and that it may be a rule without being formulated. That was why the problem of meaning was felt so acutely in the case of promising. and a conclusion is justified.g.g. the long foomoa. When the rules define a practice which we are free to engage in or not as we like. for example. in some cases by the truth of its prernisses. Another sort of practice with a use of modals interwoven into it generates customary rights such as are to be found in any human society that I wer heard of. in the other: "If you don't treat these rules as binding. so these "You must's" and "You can't's" are the more basic expmsions in logical thinking. it is otherwise." But isn't the latter an odd thing to say especially if someone i s only making correct moves? In what move is he 'treating the rules as binding' or failing to do so? The difference will appear in his attitude to 'incorrect' moves. and acceptance of it is just not minding doing it wrong or not minding a casual variant. Horses and giraffes. If someone invents variant rules.'. It is the only thing one can milk out o f a m i t y of^^ into a proposition' " (PI. . but the restriction on doing what it is another's to do (for example) or on preventing him. The rules forbid it. Nevertheless the problem is parallel. And just as "You can't move your king" is the more basic expression for one learning chess. "Consider 'The only correlate in l a n y a e to a necessity of nature is an ~I'bitTary rule. either in fact o r in Wittgenstein. some people will think of these as perhaps up for revision. you can't do that if you are going by this rule.these s e e m to be found everywhere. But they are not what Hume calls "naturally intelligible" that is to say. made from given rules to particular practice? Always there is the logical mud: you 'can't' have this rmd that. you don't 'have to' move your king. e. In the one case it's "If ~ o don't u make these steps you won't be doing that dance". and in explaining what a rule is beyond a mere regularity. "Is this truth. what are the new . and internalizing such acceptance. and of which we may after all construct new variants with different rules. and what can that be? In another sense. not logical truths. the concept 'a right' may indeed not exist. for the necessity is supposed to be generated by tire existence of the rule. But this account neglects the role of rules in a practice in which something's being a rule is supposed to justify or necessitate this action. 'The requirement of acting so because of a rule is not generated by the rule's being uttered. e. In chess the question would be: "Is he changing the game? If so. . g if you want to dance a certain dance. grasping the idea of a rule: all these hang together. since it lies at the bottom of his learning the concept of the game and its rules. for everyone would agree that the necessity of acting so because o f a promise was supposed to be generated by its being given. The parallel between rules and promises is obscured t)y the fact that a promise is a sign. they are not expmsions of perception or experien~e. whereas we want to say that a rule is not a sign but is rather the meaning of a sign. Of course if we speak of rules. I. is the product of sophisticated reflection on the data. I t is this concept that marks something's being treated as a rule rather than as a description of behaviour that you may want to engage in. in some by the steps taken in reaching it.~ They are understood by those of normal intelligence as they are trained in the practices of reasoning. . except that of acting so ifyou wish to engage in the practice. The concept 'a right'. such as making a supposition or drawing a diagram or constructing a table. Where the regular manner of acting defines a practice. colours and shapes the existence of these is not such a product. he's in check". Here is a description of the behaviour called "dancing this dance" and you look to the description with a view to making it a true description of your behaviour..190 Medieval and Modern Philosophers The Qwstion offinguistic Idealism 19 1 game require that you move your king". is part of learning the concept The rules of the game require . Hume might ask what act of mind the assent to a rule is. and if we speak of logical truths. you aren't playing chess. And it is this that is in Hume's sense 'naturally unintelligible'. But when it comes to rules of logic. you will make the steps and other movernents which belong to that dance. there is the question whether these rules have been followed in some exercise.Ja * ~ u m e n v r r n w ~ b :him i-md gHN judgement and the-ofarermmrrall e 1. Both ideasare a mere distraction. "But that wouldn't be a move in the game!" True: but that only means: no other move at this point is called a "move of the game". not by rules of logic but. one will say. I should perhaps have divided it up: Is it so actually? Is it so according to Wittgensdn's philosophy? Now we have partial answers. is the subject matter of logic.ical necessities belonging to the nature of such things these rmn to be r e p d e d by him as 'gunmatical rules'.111. We move in just such a circle as Humt complained of for promises.' 'Requiring' is putting some sort of necessity upon you. the transition. According to what rules is the deduction. acquiring the concept of the 'you have to' which is expressed. we may avoid the problem by saying that the rules of course don't impose any necessity upon you. beyond a resolution to act in a certain regular manner. unless someone compels you. E . Valid inference. a system with more than two truth values. learning the very idea of such a game. some people will entertain the idea of constructing variants. . But the me~phy. ) 37s). 1 rules?" The incorrect move 'isn't allowed'. the product of human linguistic practice?" This was my test question. that it is given in a formula for acting. This b a form in - - - quaWy 'kk.' Accepting "You have to move your king. And why? Because it is a rule. But a child who is learning may not yet grasp the concept: 'The rules ofthe game require . you must grant this in face of that. .

If something is a logical inference? ' then "no experience can contradict the conclusion without contradicting the premises.19 x rMedieva1 and Modem Philo~ophers The Qucstion ofLinguistic IdGalism 199 which Wittgensteirl advances propositions about which he finds something convincirig but which he is worried about. the product of reasoning. which are far more compls F b it ia o b v i o ~ that ~ mathermticl is a huge structure. Now this process Wittgenstein himself once described: "You can ward off each attack as it comes" (Personal conversation). Wittgenstein proposed the following: if one could learn to characterize the presence and absence of colours by truth-functional formulae like "Either no green. So ostensible proofs of absurdity are assumed to be rebuttable. if right. Could he say that quite generally? Is it really 'arbitrary'? He always seemed to say in particular cases that something that appears as a metaphysical necessity is a proposition of grammar.I . In natural theology there is attempted reasoning from the objects of the world to something outside the world.accurate observations? Of course we have ' m a n s of representation' at our disposal here: we can talk of 'appearances'. and the arbitrariness that then seems to belong to what is counted as logically possible. Wittgenstein certainly worked and thought in a tradition for which this was impossible. But he could deny that this was a case of rightly claiming to derive new information. but doesn't it imply the 'no new information' thesis about logical inference (narrowly construed)? Elsewhere (RFM.cxcludc another. or would-be rationality.] So does it depend wholly on our grammar what will be called (logically)possible and what not i. than to allow that one observation should logically. and when we arc tempted in philosophy to count - - - - - 1 should sl b .Y sentmce-like formation that we know how to d o something with. If so. which is counted as what was meant: that is what fixes the meaning: And so it is about following the rules of'correct reasoning. On the other hand contradiction and absurdity is not embraced. The last statement is not precise. Wittgenstein's attitude to the whole of religion in a way assimilated it to the mysteries: thus he detested natural theology. I. of impossible states of affairs? And so in fact Wittgenstein's characterization of the three 'observations' would be challenged. One draws the conclusion as one 'must'. It refers e. 1s grammar 'arbitrary'? (See his remarks on "Every rod has a length" IPI. or white". VII. He preferred it left jagged. On the other hand. [I take this to mean: what is not the case. this means at the very least that it is neither possible to demonstrate them nor possible to show once for all that they are not contradictory arid absurd. That is what "thinking" means ( f t k ' ~ M . and might put my interpretation in the lbrm of another arrow. Wittgenstein even took "he moves without moving" as a contradictiori in intent. But might it not be that w don't let any situation be characterized as one in which there are these three . in religion. But again. then it must be possible that all of them should occur. In the Catholic faith. and pictures. this tioes not rriearl just the practices of' arranging words together and uttering them in appropriate contexts. 25). and was impatient ot'being told that that at least was not so. The 'no new information' thesis could hardly be put more strongly: that is to say. what part of this was philosophical (and therefore something which.. "This can be disproved. Logic then dictates our refusal to call something an "observation". aren't there even single impressions. but a smooth surface has been constructed. $251. then here Wittgenstein was a linguistic idealist. to action on the rule. snd edn. 6 . When I see an arrow at an airport pointing vertically upwards. On the contrary. VII. it is difficult to say. not every technique has its application in our life. or complex of structum. For he could say what I just said.horizontal and pointing in the direction I an1 looking ill when 1 see the first. what that grammar permits?"But that is surely arbitrary! Is it arbitrary? It is not ~ V ~ I . of 'logic dictating' what an observation must be. He wished to take the contradiction as seriously intended and at the same time to treat it with respect. actually going this way by the logid infaena u a practice and do not deal with his views on mthrmaticr.)However thismay be if there is such a thing as idealism about rules and about the necessity of doing this if you are to be in conformity with this rule. still it can at most d o what a painting or relief or film does: and so at any rate it can't put there what is not the case.1. He would describe'this with a characteristic simile: there is something all jagged and irregular. others ought to see) and what part personal. just because a logical conclusion contradicted a possible distinct observation. might there not be three distinct observations which would correspond to an inconsistent triad of such formulae? But if they are distinct observations. But the attitude of one who does that. ig I). non movetur). This was connected with his dislike of rationality. if what it represents is the case. are canvassed in the following passage: "If a proposition is conceived as a picture of a possible state ofaffairs and said to show its possibility. the Incarnation. But that means that you are moving about within your means of representation" (RFM. 64). and some people have a desire to encase it in a smooth ball: looking within you see the jagged edges and spikes. he is willing rather to allow that an actual observation may contradict another actual observation.. then what will Wittgenstein say about 'illogical' thinking? As I would. or wishes that that should be done. certain beliefs (such as the Trinity. the first "moves" being transitive and the second intransitive (movet. But the arrows with their interpretations await action: what one actually does. To repeat. their source is quite other. The signpost or any directive arrow may be interpreted by some new rule. each one in turn. 1 don't know how to distribute this between philosophical observation on the one hand and personal reaction on the other. The dependence of logical possibility on grammar.e.g. religious mysteries are not a theory. that it isn't thinking At the Moral Science Club he once quoted a passage h*m St Augustine about God which with the characteristic rhetoric of St Augustine sounded contradictory. nnd edn. but I still believe it" is not an attitude of faith at all. the Eucharist) are called "mysteries". is not that of willingness to profess contradiction. He insists that these tlrings are the creation of human linguistic practice. lmy consideratiom h m ottly concernWittgenstein's attitude to logic. I rnerltally 'reinterpret' this. - .

$108). Otherwise such an utterance may be nothing at 111. ltnpossibilityeven has a certain role: one examines a formula to see that the valencies are right. and was not supposed to be necessarily true for all times. They are not ours. That is not to say that he never sees grounds for criticism. can detect error only in its own domain. That sounds quite different in the circumstances which actually hold. In On Certainty Wittgenstein was led to consider Moore's claim to know. a different picture of the world. a human construction. we cannot move in them. To say something is 'superstition . Above all. A scientist cannot condemn superstitious practices on the basis of his science.jective. I. And yet it. J I I I). in thoughts belonging to its own system of proceedings.g.1. 1876-7 of more than a year each.not mistake' is certainly a reproach. e. but that should not incapacitate his readen! What he wrote was a u e when he wrote it. 1. is no more right than when in a dream he s a p "It's raining" while it a c t d l y is raining. $676: He who says "I'm dreaming" in a dream. At the date at which he wrote the certainty of never having been on the moon was altogether different from. Talking about these matters I once asked Wittgensteinwhether. In 1871 a Russian explorer. to the way it would sound if some men had been on the moon. "Of course the compound need not exist.) This means that I have to be actually operating the languagea6 My proceedings with it have to belong in the system of thought that is in question. Wittgenstein's considerations about never having been on the moon have constituted a fool-trap which a surprising number ofpeople have fallen into. That is not agreement in opinions. But in none of these cases is Wittgenstein willing to speak of a 'mistake'. is quite clearly regarded as a linguistic creation. $ 0 4 1 ) . it may be some strange secondary application of words. among other things. therefore.104 Medieval and Modem Philosophers The Question oftinguistic Idealism 105 some quite useless thing as a proposition." I believe that the objection is a religious one. say. and things aren't like that but I suppose that they are. Wittgenstein invites us to compare 'logically possible' with 'chemically possible'. What are the implications of 'agreement it1 language'? I t runs through Wittgenstein's thought that you haven't got a mistake just because you have as a complete utterance a string ofwords contrary to one in which some truth is expressed. The situation he imagined in this remark has come about. H . What he calls a 'mistake' is something for which there are unsatisfied criteria of correctness.not mistake" :superstition "itself produced by grammatiul illusions". such as is winced by the Aztecs in their quinquagcsimal fear that the course of nature would change? Or a strange belief like those of Tibetan Lamaism? O r to magical beliefs and practices? Of coune. ..It is what humans say that is true and false. It was a criticism of a particular kind of philosophic thought. This leads us on to another way of raising the question of a sort of idealism. or of a special fbrm of beliefwhich flies in the face of what would be understood to falsifjlit but for its peculiarity. that he had never been far from the surface of the earth. (There is a corresponding condition for being right. Science can correct only scientific error. if he had a friend who went in for witch-doctoring. being right? As readers of the 'Notes on Frazer' will know. 1 want to say: my not having been on the moon is as sure a thing for me as any grounds I could give for it" (Ccrt. Now our question is: outside this one example. The notation enables us to construct the formula H o p . visited New Guinea and tha. and some perhaps without knowing it. he would want to stop him.. .0 . It is ob. These alien practices and language games are simply there. Wittgenstein rejected the idea that 'our science' show that magical practices and beliefs are e p r s .H .0 . The exclusion belongs to the system. it may be a mere manifestation of ignorance like a child's. with its rigour. " 'I &now I have never been on the moon'. I. But there is no need for him to hold such a philosophy in order to pursue his science. A compound tnight be called chemically possible if a formula with the right valencies existed. "So you are saying that human agreement decides what is true and what is false? . About the merits of other proceedings it has nothing to say except perhaps for making predictions. It may be madness. he thought it stupid to take magic for mistaken science.but the system then rules it out. He may do so on j e basis of a 'scientistic' philosophy. Nikolai Miklouho-Maclay. 1. but I don't know why. can't have less than no con~pound corresponding to it in reality" (PI. even speaking audibly. I. but even a formula HO. does Wittgenstein's view preclude our saying "They are utterly wrong.g. . but does it preclude having grounds. it is not up to me to decide what is allowable here. $52 i ). $ i lo) or 'a queer reaction' or a manikstation of some different 'picture ofthe world'. it may he 'superstition' (PI. $590). In this case one could give grounds for this knowledge .0 . He thought about this for a little and said "Yes. and whom he ." (PI. is not a way of 'bargaining its rigour out of it' (PI. that is often because we have not considered its application sutficiently (PI. that is. In his work up to On Certainty we might think we could discern a straightforward thesis: there can be no such thing as 'rational grounds' for our criticizing practices and beliefs that are so different from our own. and they agree in the language they use. I can be accused of making a mistake when 1 know what it is for a given proposition (say)to be true. The non-arbitrariness of logic. criteria which correspond to the intention of the speaker. He had a Polynesian servant who died.rftcr lived there for two periods ( I 871-8. This is an interesting illustration. On Cnkrinty ICrrt. . and that the earth had existed for a long time before he was born. Now it is clear that Wittgenstein himself did not hold such a philosophy and did not think it belonged to respect for science and interest in scientificmatters to do so. in one way it can't predude objecting. the certainty of never having been in China. (In this sense Aristotle might have called Plato's Theory of Forms 'dsuperstition. they are in darkness?" Does it preclude objecting to e.

do not count for him. $ n 7 1). saying perhaps there is a way of getting there which I don't know. interpretations of a rule. no one has wer been on the moon. beyond 'I have never been to such and such a place. but soon be convinced by what we tell him seriously". Not merely is nothing of the sort evcr seriously reported to us by reasonable people. . the statenlent: "I know that never . 9 8 6 ) Now let us contrast these passages with the following ones. let them be never so sure of their belief. and have compelling grounds for believing that'? And here one would have to say what are compelling grounds" (Cert. He told them that he was a man from a place called Russia. Various developmentsarc possible: t h y wonder if t h y will ever meet the counonaua on the moon. . Moore has a real occasion lvr saying that. They also believed that Maclay himself had come tiom the moon. come to an end and then one is convinced. 1964. and it will accordingly be able to produn apparently telling grounds of the one or the other? But is t h m then no objective truth? Isn't it true. a d the people of the s t r q c a i kl u ~ m r this. $ 2 7 2 ) . 8 It dunp the situation once people have god to the moon by spceroctet. because we wouldn't know what to identify as someone's beliwing that. we cannot confidently understand his signs as signs of doubt" (Cert. . also fiom On Certainty. There is no demonstration against him.This is because (a)there is a gap between reasons and conviction. the moon is a long way off and it is impossible to climb up there or fly there. What we believe depends on what we learn. Wittgenstein goes on to ask: "But what does it say. but those who get there know at once that t h y arc t h m . Isn't this altogether like the way one can instruct a child to believe in a God. or one acts. And. or There is a great difference between the tribe described in $ 106 and the tribe that captured Moore in J 064.. . believed that hemeant that the man had flown away over the sea. Road Belong Cargo. but there might be people who believe that it is possible and that it sometimes happens. and so they supposed that Russia was in the moon. They held a meeting of senior people to ask him questions. then it is certain that no one has evcr been on the moon.)' The grounds were not beyond their grasp which was why Moore couldn't explain to the savage tribe. $058).If now the child insists..146 Medieval and Modern Philosophers The Qystion o f Linguistic Idealism 197 buried at sea. in the other: "What reply could I make?" But here it isn't that he can't 'give them the grounds'. (Cert." Wittgenstein comments: "We should feel ourselves intellectually very distant from someone who said this. We all beliwe that it's not possible to go to the moon. -What do we 'tell him seriously'? That it's all nonsense and that these ground were su#n'ent - - ' See Peter Lawrence. that someone has been on the moon? If we are thinking within our system. We say: these people do not know a lot that we know. and 1say it was only a joke. Wittgenstein asks: "What reply could I make to him?" . is more malleable and "will not stick to such a belief.. What count as reasons for us. Ifwe compare our systern of knowledge with theirs. But suppose that instead of all these answers m met the reply: "We don't know how one gm to the moon. For this demands answers to the questions "How did he overcome the force of gravity?" "How wuld he live without an atmosphere?" and a thousand others which could not be uwmred. etc.Now compelling grounds are objective (Cert. justifications. then theirs is evidently by far the poorer one. They. Or: they dismiss the idea of such an encounter u a misunderstanding.. Let us consider the child who goes on believing the story in $106. he gestured towards the horizon. But a child is not as strong as a tribe. like explanations. When the people asked where the man was. because they have fantastic ideas of hurnan ability to fly and know nothing about physics.. Reasons. $$l06-8) That is. in the sense that it is not up to me to decide what is a telling ground (Cert. and who indeed grant that there are no ordinary means of climbing up to it or flying there? -But a child will not ordinarily stick to such a belief and will soon be convinced by what we tell him seriously. And similarly for the man in $ 108 who says "We don't know how one gets to the moon. The child tells me the story. which are on the same theme but have important differences and a different moral : Suppose some adult had told a child that he had been on the moon. In one case Wittgenstein says "He can't give them the grounds".the same question as in the case of the tribe." Here." But there's no refuting such a man. but he can't give them the grounds fbr his certainty. it is said. those grounds pass them by. that there is no question of its being possibly true that Maclay was ever on the nloon. but those who get there know at once that they are there.. what reply could I make to him? What reply could 1 make to the adults of a tribe who believe that people sometimes go to the moon (perhaps that is how they interpret their dreams). $$~64-5).' It is a similar tribe that Wittgenstein imagines in the fbllowing: I could imagine Moore being captured by a wild tribe and their expressing the suspicion that he has come from somewhere between the earth and the moon.. Etc. don't convince them. And (b) because this is not one of those cases where "if someone gives signs of doubt where we do not doubt. We know that those people were fantastically wrong. o r false. Mancheater. And not a case where the mere idea of accepting this (namely what he thinks) destroys the whole business of judging anything. among which was "whether he could or would ever die". goes this way. but our whole system of physics forbids us to believe it. Not a case where the hypothesis that someone believes something can't be sustained. the man hadn't been on the moon. that none exists. and wen you can't explain everything. This would be an occasion for making that statement (Cert. ." We should feel ounelves intellectuallyvery distant from someone who said this. $964). (Ccrt. (And they mean something different by 'getting' to the moon from what he would mean.they are wrong and we know it. Moore tells them that he knows etc. and even you can't explain everything. $154). We may 'just have to put up with it' (Cut..

possibly do so?" -There was no analogue in the moon story for the 'apparently telling grounds' on the other side. not just that the earth has existed for only a short time. we tell him not to believe the story.i. What then? Are we not to tell him 'seriously' not to believe the story? Are we to encourage hit11 to 'have an open mind'.rgph i have just quoted could hardly co-exist with our physics. He could perhaps be given that and still retain the idea of his superlative position." In another passage Wittgenstein imagines Moore arguing with a king who was brought up to believe that the world began with him. For this king has been brought up to believe. and therefore believed this. $ g ~ )That . then there is a peculiar line drawn at a recent date. and a picture & s t couldn't co-exist with our physics might well facilitate a belief in some people's visiting the moon. - - Could Moore really prove his belief to be the right one? I do not say Moore could not convert the king to his own view. that 'past' was supposed to be merely part of the scene as it came into being then as Philip Gosse thought the fossils were created in the rocks in 4004 BC. we can find an analogue for the objection to the story of the man who had been on the moon. On the other hand the statement b t -ne had done so raised difficulties From his y t m gknowlcdge. Etc. We know what people have been doing for thousands of years.. Whm did he come to be? If. we 'had lo put up with it'. 9 6 s ) - - i . to him is part of the scene that came into being with him.whatinduce one to go over to this point of view." Could he reply: "No more than 1 lie to a child whom 1 teach arithmetic. long time. you must do this here" was itself the creation of language. Just that was learning to follow rules. and we give the reasons: we thereby tell him implicitly not to form and stick to an idea of getting to the moon for which those grounds agairlst its having happened would be irrelevant. is. e should be W e might teach him that the earth has already . if this is the rule. and that is an implicit directive not to believe it and not form such an idea. never having been very far from the earth's surface. and insist that he see only this as the way of following a given rule"? He could not fairly make that reply.. This is not just the same thing as his 'being given our picture of the world' perhaps in connection with learning our physics. What has physics to do with transport to the moon in dreams? Turning now to the example of the earth's having existed for a long time: I can imagine a man who had grown up in quite special circumstances and been taught that the. $107 : The comparison with "the way one can instruct a child to believe in a God or that none exists" derives its interest from the rest of the paragraph : "and it will accordingly be able to produce apparently telling grounds fbr the one or the other". $ 94) that he had not. Or again: we tell him that there is (can be) no truth in the story.. I t is at any rate clear that this is a quite different treatment of "No one has been on the moon" from the case of the savage tribe. (Cert. for a long time W trying to give him our picture of the world. he is brought to 'see the world in a different way'. These people might not: they might have the same physics as we do. If the scene in various ways included suggestions of a past. But there rsc connections: the 'picture of the world' acquired by the man in the came into existence 50 years ago. Here are the buildings that were built hundreds of years ago. it seemed. but it would be a conversion of a special kind : the king would be brought to look at the world in a different way. (Cerl. in the conception of himself that he has been given. For about that he says "What could induce me to believe the opposite? Either a memory. and the child's acquisition of this 'must' is. he has lived only as long as his apparent age. or having been told. And if Moore converts him. Examining the parallel. Nothing in my picture of the world speaks in favour of the opposite" (Cert. 'The child has heard that someone made the world. . What we would say happened before his time. This would happen through a kind of persuasion. And these considerations show why Wittgenstein should make lndu manu validas potis est moderanter habenas? Namely "Who could . he is given a superlative position in the world.I 98 Medieval and Modern Philosophers The Question of Linguistic Idealism 199 for saying so? Why can't we say that to the man? In his case. $ 9 4 ) This is again different from the case of the previous man (and indeed from the man at the beginning of the same paragraph). how could anyone . No. One then simply s a y something like "That's how it must be". quis habere profundi This'idea of one's 'picture of the world' did not come into play in the same way in Wittgenstein's considerations about never having been on the moon.e. been far from the earth. That is. ~verything I have seen or heard gives me the conviction that no man has ever been far from the earth. part and parcel of its acquiring concepts and practices which it belongs to the human intelligence to have. They had a poor system of knowledge. But the same thing can't be said here. Suppose I were to say: "Wittgenstein. The objection is that of'1. it is n o t an ac-l part of his 'picture of the world' that "inherited background a@nst which I distinguish between true and hlse"(Cert. but that the world began with him. . But our teaching of that man will presumably be primarily of history: "The earth has been inhabited for a long..e.ucretius: Quis regere immensi summam. these are. The "must" in "You see. of how they can become comperent to entertain all sorts of'possibilities. to work out every possible conception and entertain every strange possibility? Absurd! That idea is based on a false picture of how people can learn. keeps it going and runs it. you show quite clearly that that would be lying to the child. The 'pkntrcofthe world' is not the same thing as the 'system of knowledge'. as we have seen. i. Remember that one is sometimes convinced of the correctness of a view by its sinrplicirg or spmetty. that no one had. The king's rejection of this base line and'of such a position for himself in the world is rightly described as 'conversion'.

$ 6 1 I 1. one or another living linguistic practice which we have. verifying are all processes that go on within. $ 6 12) So what is in question here is not: cultural relativism. can't tell you what is the next step in applying them? In the end you take the rule this way. Again. are part of a 'world-picture'. "If we call [what they think1 'wrong'. For those were cases where the 'doubt'. NO! . That is to say." (Cerl. Doubts whether this is a tree or whether his name was L. the other wrong. I couldn't say to them 'I have never been to the moon . A world-picture partly lies behind a knowledge system. one calls something error which counts as knowledge in another system. but the movement around it determines its immobility. The other men who beliwed that the earth had only existed for a short time were comparable to the savage tribe. for this has to be the back-reference of "I said I would 'combat' the other man but wouldn't I give him reasom? Certainly. and so on. Similarly he enacains the same sort of idea about things like knowing his own name. there are some propositions which are quite solid for me. No one ever taught me that my hands don't disappear when I am not paying attention to them. ' world-pictures are incompatible. May not the Dalai Lama learn our physics and astronorrly and history.. $$15x-3) I want to say: we should not regard the struggling investigations of On Certainty as all saying the same thing.. and the confident action of a s h o p k v KNing one with apples. w a s I b d ! ' " (dRFM. but I d o not learn them explicitly: discover them like the axis around which a body rotates. This 'axis' is not fixed in tile sense that anything holds it fast.). But this king's knowledge system need not be poor at all. intepretations. 135-6). Wittgenstein rejects the idea of one of them's being right. Now is that a personal profession or a logical point? Certainly it is the latter. confirming. 8 The is to people who consult oracles instead of physicists. belief in one's 'world-picture'. Wittgenstein wrote: "If the shopkeeperwanted to investigate each of his apples without any reason. I'm c r q " (Cut. Finding grounds. Wittgenstein means only that they are a foundation which is not moved by any of these proceedings.' And if they asked 'May you not be mistaken?' I should have to answer: No." (Cert.. The conflict between the principles of Western medicine and acupuncture medicine might serve as a good example here. I may be mistaken. I hardly wer have as I apply a rule. doesn't that mean that we are starting out from our language-game and combating theirs?"1° This is put as a question. why doesn't he (then)have to investige the investigation?And can one now speak of believing here (1 mean in the m s e of religious belief.g.Of course. 5 459) His answer to 'Tan one now speak of believing here?" is rather plainly "Don't". yet: ~ g h It - - " Cd. testing. that are the 'immovable foundation' of' these proceedings. There are assumptions. Rules and the particular rule are defined by practice: a rule doesn't tell you how you 'must' apply it. . That is to say: even if he might come to count the idea of visits to the moon when dreaming as possibly part of their system ofknowledge." (Cert.The Qucstion offinguistic Idealism 130 Medieval and Modern Philosophers 13 1 those remarks about simplicity and symmetry... I really am using a dictionary. reasoning. beliefs. that it makes sense to ask whether such-and-such a volcano was active a thousand years ago. which in fact. an intellectual disaster or intellectual enlightenment? O r is it really another form of: Rules. but his answer is clearly "Yes". proving. (Cert. For the assumption is of "two principles which really meet and can't be reconciled" and "each man declares the other a fool and heretic" (Cert. like reasons. we have a "disagreement in the language they use" -but it really is a disagreement. spoils everything by assimilating rciigious belief. But when. by taking the step. not in the sense ofan interpretation. $667). for example. but by acting.It is not the same. has no real content. about which "Ifthat'swong. not surmise)?All psycholngical words here distract from the main point. One knowledge system may be far richer than another. by making 1 can assimilation^. that if I need to check on others by looking them up in a dictionary." We cannot get him right. the question arises: has one the right to do that? O r has one to be 'moving within the system' to call anything error? "Even if I came to a country where they beliwed that people were taken to the moon in dreams. of course.. W.^ But to return to the question with which I started: it may seem that if wer Thus Norman Malcoltn in his paper '"the Groundlessness of Belief '. in order to be certain ofwhat hcwas doing. speaking with this knowledge system behind one.clr whether the world has existed a long time or whether the kettle will heat on the tire or whether he had never been to the moon are themselves not all subjected to the same treatment. Wittgenstein thought of a motto once: "I'll show you differences. But how can it help us with our problem to say "At the end of reason comes persuasion"? Is it futile to say here: But won't the persuasion be right or wrong. Not all these things. Their knowledge system would be very poor. I.g.which containsa fine exposition of some of Wittgenstein's thought and an interesting thesis about the 'pathology of philosophy'. but how far do they go? At the end of reasons comes persuasion. $57. Not everything that is 'unmoved' is a foundation. even though to believe is not in either case to surmise. for example. but only commit frightful confusions. belief in a system of knowledge. But. with their interpretations. I cannot doubt or question anything unless there are some things I do not doubt or question . and disagreement is just imagined by the philosopher. just as it may be connected with far greater capacities of travel.e.e. Thus the assumption of Descartes' deceiving demon is not indeed excluded: "If I were to awake from the enchantment I should say 'Why. And a world-picture is not the same thing as a religious belief. And among these would be such facts as belong to my 'picture of the world' . even then he could not call his present view of the hypothesi~that he'd been on the moon a mistake. give out in the end. In all this I did see a sort of 'linguistic idealism'. cannot finally dictate how you go.. that I know the meaning of some of'the words that I use. and yet go on believing that he is the same person as all the previous Dalai Lamas? A main theme of On Certainty is the 'groundlessness' of one's worldpicture. say. By this.

" (e.. in the one case in persuasion. .and now it actually looks as if I were wrong (Cert. namely that that sequence of phenomena is rare. They are unreal. In short. not competent to judge. or again when I have said "I can't bewrong" about the kind of thing I have a right to say it about .. in the other in decision. . . But to say I can't be making a mistake and yet to add: I may rightly or wrongly come to believe I wasn't competent to judge .but only the conflict. But the doubt is empty: "What difference does it make for me to 'assume' this I the deceiving demon]? I might say "Yes. does not discredit the sentence "I can't be making a mistake". But the 'language-game' of assertion. . 5575)." . and asks "How much difference would that make?" But still: it is "always by favour of Nature that one knows something" (Cert. Rut such a case.The situation terminates. . (Like the doubt whether another means the same as I by "red".g. the calculation is wrongbut that's how I calculate" (RFM. That one knows somethingis not guaranteed b~ the language-game.135-6). NB: this is "may". or decision I' Suppose he has said "I can't be making a mistake" -ofcourse.. The argument from mere conceivability leads only to empty. even though there are occasions on which. $9481-4). . . and there are cases. that would indeed completely change the character of the language-game. $$643-5) The Qurstion ofLinguistic Idealism 133 . It is only in face of the 'presumptuous' use of "know" by Moore that he wants to exclaim: "You don't know anything!" (Cert. attained 'realism without empiricism' (RFM. "I don't know if there ever was such a person as Queen Victoria"). would be linguistic idealism with a vengeance. $ 578). That one knows something is nd guaranteed the languqe-game.rightly come to believe 1wasn't competent to judge"? True. That I may come to think I was wrong. where. rightly or wrongly..Medieval and Modern Philosophers not wake up. have to learn from 'a higher authority' ? (Cert. ornamental doubt. and not "might". $407) "When one hears Moore say "I know that that's a tree" one suddenly understands those who think that that hasn't been made out at all. The case of conflict remains unfinished business. Otherwise wouldn't it discredit all assertion? I can't be making a mistake . But I can't deny it.) It is quite otherwise either when there is a conflict of irreconcilable principles in real life. Wittgenstein now makes the observation: "If that always or often occurred. I t is as if 'I know' does not tolerate a metaphysical emphasis" (Cert. depends for its character on a 'general fact of nature'. But . He could say the same in the case of knowing his own name. 9 646) the two phenomena are contemplatedjust as - . believe I realize that 1 was not competent to judge. as in face of the idea of the deceiving demon. 1. which for speaking humans is so important a part of the whole business of knowing and being certain. But here Wittgenstein has at last succeeded in his difficult much use is that? "I can't be making a mistake" expresses certainty: doesn't that consort ill with the thought. Theconclusion is not an idealistic one. or persuasion.. if someone with our normal education says "I don't know. in the sort of case where that's an appropriate expression.. To say it was. In the next entry (Ccrt." for "know" in such cases. have my eyes opened. . phenomena: the "I can't be making a mistake" and the coming to believe (rightly or wrongly) that I was then not competent to judge. which is aroused by the picture of the private object. but they would legitimately arise if the false picture were true. is not news. (Cert.isn't that saying "1 can't be wrong. But here.. And now isn't it as ifwittgenstein were saying: there isn't a right or wrong . this particular battle has been fought out. gg). $505). .. Certainly one can imagine cases. sure. .but I may indeed sometime. This is so. or its possibility. "I may." With that observation. " . these were illegitimate doubts. . $641). yet I may be wrong"? The distinction between mistake and something else that can't be called "mistake" just rescues us from the contradiction. or always putting it in front of one's assertions. after 'waking up' one never doubts again which was fancy and which was reality. He concedes them only in face of a false picture of legitimate certainty. rightly or wrongly. certainty is defended in face of 'legitimate' doubt (Cert. he has third edn. VI. one would want to ask: What do you mean? How are you using the word "know"? What would you caU "knowing that"? This explains why Wittgenstein sometimes dwells on the possibility of substituting "I believe .that is the hard part. I don't usually look at that possibility.

in history. loo assimilation. 14 certainty. 1i 7-18. 76-7. 3% theory of categories. Aristotle. s 1-3. i 03 understanding and.J. 5s Aquinas.83 principle of individuation. is4 Anaxagoras. 51. 81. 10s Cimo. 36-9 Boethius. 106-7 defined in tenns of deliberation. 15 . 14-15 Hume's principle of. 8s Brentano. 9s Aristotle. 188. 86-91 calculation. matter and. St. 44. Aristotle's theory of. 101. 95. 44-55 desire. xi 'act of mind'. 84 Augustine. l o e l o r . 59-60 chemical elements. 106 asymmetry. xi. 133 positiveness of. 84. 93-4 early theory of forms. Brentano. 16. 93-4 cause. 8 W .Index acknowledgement. 17-18 concept. Aristotle. classes as. 93 'intuitively certain'. 73-5. *Q knowledge. 16-1 7. Aristotle. 66-7 I. 107. 106 early theory of forms.54 change. 131 addition. 8 1. 101-8. loo. Anicius. 106 Cebes. 57-65 revised theory of forms. Parmenidea. matter and.115 atomism. SO. 96-8. 1 13. and linguisticidealism. 103 causality. 57-8 and object. Franz. 196. 'language-game'. concept of. first-third person. gi class. 133 predication and. see olto mathematics. and premise. Platonic form. x. 10s early theory of forms. x. 180-0 1 becoming. Brentano's conception. 58 'chemically possible'.83 Plato. 69. '73. 1lg action.48. language and. 48 rules and. 11-14 1 7 . 'demonstrably certain'. Aristotle. St Thomas. 11g. 06833 collections. is*. 8 1-5 Aristotelian Scholastics. and sight. 93 language and. 104-6 Caesar. Hume on. betwem will and emotion. 77 Brentano..4g. lo8-g conclusion. x. and beginning of existence. 74 necessity and truth. r 3. i s colour. Hume's discussion. 60 antiphasis. xi.5*. 85 will and emotion. 66-77 arithmetic. 13s cf. I 4 annihilation. see also numbers affirmation. 18.Julius. numben assertion. Aristotle. matter and. r e r m t from. s4n being. cf. viicix 'metaphysical'.A. 15 and dfect. 40-4s combinability of number. m t h . is4 Dt In("pretationc. 89 learning and. relation of monads t o infinitude of. 97-31 belief. I so Ayw. 66 categories. 44-9 agreement. 15. 103 thought and action in. i 84 choice.

Ilving.-93.74. 84 Goethe.. 16. 93-9 concept of intentional. 15 cf. counting. x future contingent%x future contingents-conl. of beginning of joy. 93. 6511. 90 decision. 96-8 imagination. language and. 75 and will. 119 kinds. 61. 93. 49 h u s t (Goethe). 194 principle of individuation. loo 'just itself: Plato. 181. 57. as beings. 189 grammar.110--11 individuation. Aristotle. 5 9 76-7 effect. cause and. I l4n. 101.93-4. 101 De Generatione et Corruptione (Aristotle). 9g Diodorus Cronus. 85 will and. 196n learning. Aquinas. 3 incompatability. 193 essence.. and will.136 Index revised theory of fornls. of cause from effect. 85 and thought.66 evil. 119-16. essence as general. 88n 'keeping your word'. 106 Katz. 84. 199-3 growth. 16. 51. 1 0 9 3 Hume. 9-PO Platonic. 180-81 hypoihesis. notion of the. language and. 93 languageand. C. Truth and Loge (Aycr). see also numbers language and. 75 De Civitate Dei (~ugustine). Philip. Dr Stephen. go individual. and lat to's theory of forms. 15. 75 beginning of existence and cause. realism without. muth. 9 1-33 fiundationr ojArithmetic (Frege). 1 14-15 counting. log generation. Aristotle.. 100 rwised theory of forms. 64-5 Hartley. 108. 84 history. belief in matters outside. 108-9 predication and assertion. 86 interpretation. 94 genus-species division. doctrine of. W. Hume. 599 54-5 Gistence. I 9 7 Copi. 69. 1 16-1 7 Cratylus (Plato). and will. I 98 revised theory of forms. and judgement. name and. 16n Kneale. will and. 81 Aristotle. 16 . 105. 116. 87-8 idea. 1 2 1 explanation. 1oo impossible. David. 119. 61-3. 94 count-noun. 18gn intellect. I 15-16 numerial. 104 ~evclopnunl o f logic. 104. P. principle of. 8 inference. g.48 dialectical science. 46. 39-3 difference. counting. 'strongest available'. Johann Wolfgang von. 8 0 'natural unintelligibility' of promises. 4gn hate. 1a 6 Kncale. and linguistic idealism. 62-4 deliberation. 183. confusion with principles. 48 Aristotle. 69. 1rg-31. 89 Ethics (Aristotle). 107 exclusion. choice and. 116 indefinable. M. xi. 40942 Moore. will and. 96-7 function. expectation and. Law of. 34-5 given. Aquinas. 118 contradiction. Brmtano.64-5 imaginability. 49 cause and beginning of.133 cf. Galen. 1 1 7 Frege. 116-1 7 and principle of individuatid. 13-14 is of changeleasly true. 93 Kant. . and revised theory of forms. 83 Pannenides. gl Hobbcs. xi linguistic idealism. 86-8. I 19n concept and object. RenC. Aristotle.76-7 imagination. 13. expressed by grammar. 90 hypothetical argument. 13s system of. form and. I I 8-3 identity.65 rwised theory of. 93-9 cause and effm. 40--4o luyplage. 8 1 Lawrence. 108. I 18-3 Lquagc. 9-20. X. 81 Mms 34. 6g. action defined in terms of.44-55 Dc Rerum Natura (~ucretius). x division. 119-14. 116 Hume's doctrine of abstract. 67-8. 89. 119 Extensionality. reasons and. lo I divine knowledge of the future. definite. 115-16. 119 eternity.90 De Anima (Aristotle). leln.. 104.64 and universal. 59. Aquinas' discussion of. 79. 1 10-1 1 future. 66-7 9 Descartes. g 1 Geach. 49. 56 distress.107 Gosse. 70 definition. 1o t judgement. 103 Hegel. Thomas. . Imrnanuel. 16n knowledge. 69 intelligemx. 95. 7 1. 46 divine knowledge of the. 109 experience. 101. language and. 9 7 'doing well'. 19.106 eniotions. ~ q u i n a s8 . 8Gg. 3-8 religion and. 64 De Inlcrprclationc (Aristotle). 16n dialectic. log fear. 15 contingency. viii. language and. 114n. 1 0 9 feeling. 118-19. 98 separable. 39 conviction. 3-8 expectation. lo8 corruption. 106 forms. 8e-3 and early theory of forms. The (Kneale and Kneale). 45. xi linguistic. 100-107 empiricism. principle. 86 language and. 57-65 inexpressibility. 8 9 3 future time reference. log good. 48. 1 19. Plato. 95 hope. language and. xi. 44. 139 reached by deliberation. 1s1 consequences. Aristotle. 80 Mmo.96-8.54 and understanding. M. 1 13 Cyropaedeia (Xenophon). laneuage and. 107 intercommunion of forms. 3 1 existence. 1 1-1 9 falsity. r 9 l n Aristotle and. choice and. 33 geomeny* early theory of forms. xi desire. and revised theory of forms. 84. 63 descriptive terms. Peter. Gottlob. 131 description. 7 . Aristotle. 115. 109. Aquinas. can't clearly be conceived to be. 18. 96 idealism. early theory of forms. Georg Wilhelm.1 contract. 66. 16 fractions. 106 Aquinw. 3. ix early theory of. Aristotlc.

93 understanding. 109 . x-xi and Aristotle.Joseph. 53 proposition. s 1. Karl. 89 physics. 9-13. 49 'logically possible'. 195. ~ 5 . 73. . 59-3. 'intelligible'. 96 memory. I la. 190-11 proof-syllogism.57n.196n magic. 19. Wittgenstein. 57 Neurath. 76 morals. 131 139 and belief. 116-1 8 unit as matter of. 58 Plato's revised theory of forms. I n 9-3 name. 94 length. 194 Plato 3 Philosophy of Molhnnalics (Wedberg). xi object-cont. 73 Hume. Aristotle. 194 " necessity and. 103-4 Lucretius. 5 p'ophecy. and revised theory of forms. 108-1 1 'objects of knowledge'. o h . Norman. 749 75. g language and. 131. g and function. 83 'objects of opinion'. so on what cannot be thought.54. ix-x. 16-90 revised theory of forms. 5s-3. 53. 139 Phacdo (Plato). 194 love. mathematics. lo. viii opinion. 64 number. 195-79 199. 74 mystery. lr8-go. 119--90. I 0 9 3 Pompey.. 89. ix revised theory of forms. 10Neo-Scholastics. I 13 revised theory of forms. 3s copula of. nn. early theory of forms. loo contradictory. ~rist"t~e. 119 Movement o f A n i d s (Aristotle).Index learning--cant. 108-10 predication. sg Phaednu tPlato). 4 nothingness of what-is-not. I r 7-9 P b s i a (Aristotle). 15. 98-30. 59~60-69 - Maclay. Jan.og-y Platonic theory of forms. 148 tukasiewicz. x. sg. 81-5 negatibility. ng. 84 de re. and concept of length. 49 Parmenides. loen. xi. 1 s3n Meno. 14 miracle. M. 101. 8gn. I 9 9 moral virtue. choice and. 139 place. 41 Leibniz. Aristotle. 53 Nicomachean Ethics (Aristotle).56 perception. 194. and assertion. 91. 93. 74 proofs. 17-90 language and. 96-7 Marcus Aurelius. 94 PhilosOpAical Inveshgations (Wittgenstein). Aristotle. I 1a PhilosOphirche Bcmcrkungm (Wittgenstein). Karl. . qg. 48. 30-31 participation. 117 ~ o c k eJohn.56 de dicfo. in revised theory of forms. g l 'mark of a concept'. form of the. 'intelligible matter' and. I 19. 49 prognosis. 95 definition by means of. 1 14n~ 1 15 logic. 57-65 'undesignate'.30 non-identity of. Index moral character. Wittgenstein. 89. 49. language and. E. I n Wittgenstein. and early theory of forms. I n. 117-18. 1 9 9 measuring. lo. 30-31 Parmenides. . will as. 57 prediction. 1 17n Meno (Plate).83 language. 6% . 9-90. 39. . 3-8 revised theory of forms. 57. 10 Philebw (Plato). 195 Malcolm. Parmenides (Plato). Aristotle's notion of. measuring and. as cause of everything. sg self. Hume. 59 Number WordJ and Nwnbrr S $ k (Menningff). early theory of forms. ix. s 1. 118-19.99 Popper.48. 108. 95 necessity. g 1. 1 18. I 18-90 'natural unintelligibility' of. 89 on knowing. 13 probability. 73. 53 mistake. 1a I . go Newman. 15-20 knowledge. a9 nourishment. I 1. Plato's revised theory of forms. 49. Hume. vii noun. . 74 mind. 84 language and. P I . 34-43 properties. 1 g On Cerfainty(Wittgenstein). 93P94 ~ o o r eG. 61 meaning. 87 early theory of forms. I 07 positive attitude. 7 n early theory of forms. vii. I o8-g early theory of forms. general theory of positiwlnegatiw. lg private ostensive definition. 4 5 3 and matter. I 93.r8. 96-8.50. 10s past time reference. 198.196. 116. 1 16 language and. ix and rwised theory of forms. 86 promises. go. viii. Cardinal. I 9 pleasure. 30-3 1 of predicate.4 0 matter. and beginning of existence. and belief. and Plato's theory of form. s 1-7. language and. 103 .59 'picture of the world'. 117n object. 19 1-9 and truth. 139-3 monad. 36 persuasion. vii-xi early theory of forms. st4-5. 4 4 . Wittgenstein.57 language and. 34-43 Mefaphysics(Aristotle). Aristotle. . see also matter mathematical objects. lo.58. 185-6.. 16. 16-90 language and. 13 rwised theory of forms. and early theory of forms. o 1 3 . . 4 Aristotle. 59 principle of individuation. 13. 3s-3 pan. I 90. 31. 59 Plato. 64. 1sgn. 196 understanding and. ason. 11 1 negation. 86 Menninger. Plato's revised theory of forms. 29-8. concept and. Aristotle. 15.p-3 principles. I a I predicates. ss.73. 38 'other'. 63 'place-matter'. 99-30 non-existence.6 . g. 193 non-being. 8 1. objects of. 4 property-variable. 64-5 potentiality. g. 48 practical reasoning. 19. 130 one. 1 19n material difference. 69-7 1 Moral Science Club. 61. 117 membership of a class. 79-5 practice. N. language and. 54 present time reference. language and. . 63 'matter in itself'. 108-10 non-substitutability of. 1o1 understanding and. 761 109 'no new information' thesis. 13. transition of rules to. 3s passion. irrational true. g. Hume's theory of foundation ot. self-predication. 83 obligation. 13011 "many". ~ 4 ~ 3 9 . so ultimate subject of.

19 9 realism. 108-1 1 word. I o rights. doctrine of. 131 record. 41-9 truth-condition. 98 Quine. 5. language and. 4-43 similarity. *9-30. Aristotle. 16-90. n3. 131 Russell. 47.50-55. 15 soul. 91. 100-107 f d o m of. 10. doctrine of. 8 ~ 0 and Aristotle. a8 Trattatw L. go understanding and. 23 Sophist (Plato). 63 theory of. 94 universals. and Plato's theory of forms. 13 separation. vii. log same.xi. 187 'practical'. chain of. 66. 23. Anders. 54. W. 1 18 Rivo. 108-1 1 rationality. understanding.g. 19 9-3 Republic(Plato). A. theory of. philosophic. F. and empiricism. cf. language and. 59-60 'subject-predicate proposition'.30 sameness. knowledge and. Brentano's metaphor of. 34-43 unexceptionable. 83 Aristotle. szn 'stuff'. language and. 86. 85 Parmenides. 54 ~oad Belong Cargo (Lawrence). 1I sight. 5 volume. 194. 13 religion. 99 Treatbe o f Human Nature. i i 3 . 56 unit. Plato. 95 resemblance. P. language and. 1 19-13. 33. I i n-gg Tractahu.98 Meno. I 80-99. 198.. n8n Xenophon. loo psycho-somatic sensation. 93 necessity and. 19-80. og. Ludwig. 96. 74 symmetry. 42 Theaetetus. 76-7 understanding. I 18. i 93 thought. 'one thing'.196n rules. 88-9 ~ t a r t i n ~ . 9-1a . Plato's revised theory of forms. notions of. 133 reality. 108-1 1 unthinkability of what-is-not. gg. 94-5 definition by means of. Allen). beginning of existence. 47 Wittgenstein. I 95 self-predication. and matter. 106 Summa Thcologico (Aquinas). 98 quantity. 89 relations. whole. inin truth* Aristotle. 98. 44 Studies in Plato's Metaphysics (cd. 61. gq. 108-lo Stoics. 19.4 1-3 space. 1a 7 Hume.49. 105 Wedberg. 199-30 Simplicius. 51 truth-functional expression.Pope. 98. language and. 13 teaching. go. a5 sl~fiusion. and revised theory of forms. 50 theology. 5 ~ ~ 5 4 truth value. 115. n I .. in religion. 59 unity. is. see predicatiorl sensation. 50 type. o. 75 and persuasion. x. Aristotle. 15 understanding proofs. 39 Theatletus (Plato). 48 Whitehead.~ o 87 int. go Zettel (Wittgenstein). 74. 10-1 1 simplicity. xi. 1 ng-30 Tarski. 3-8 time. of. A ( H ume). 11. 44 Psychologie vom nnpirischen Sfandpurutt (Brentano). 81 superstition. 88. 66-70. Aristotle. N.a9. pre-existence of. 75-6 and emotion. 113 reason. 58 voluntary act. ns. 104 quality. 116--17 Word and Objccl (Quine). language and. 44. Stenius. 75 characterization of. and colour. vii . 31. how to use. xi substitution.ogr~o-Phif0~0phi~uj (Wittgenstein). Erik. Plato's revised theory of forms. 63 totality. Peter de. 99-30. igon Wittgenstein. xi Aquinas. 4 1-7. 13 Sd Theoty and its Logic (Quine). 54 Wittgenstein. nn. matter of number. 94-5 spatio-temporal continuity. 125 syllogisnl. 108 early theory of forms. see also forms Index variables. 1s 'whichever happens'. 73. 54 Socrates. 34-43 'something'. 1 1. 16 set-theory. ix. 51 truth-table. notions. V.Index proposition-conl negated. reference of. 50 truth-possibilities. 34-5.33 Sophists. vii.56 Hume. and matter. linguistic idealism. cf. early theory of forms. 115 separate existence. 3 I. 11. 81-5 objective.93. Bertrand. 50 Ramsey. 1 14 science. xi whole. z4. 108-1 1 'transcendentals'. 59. matter and. 44-7. 98 will. belief and. 1 9 revised theory of forms. Brentano. 84 understanding proofs. 1-16. and beginning of existence. xi. and conviction. go Sixtus IV.. 54 linguistic idealism. belief' and. log substance.