Aristotle and Beyond: Essays on Metaphysics and Ethics

by Sarah Broadie

Overview

Written over a period of thirty-five years, these essays explore the topics of causation, time, fate, determinism, natural teleology, different conceptions of the human soul, the idea of the highest good and the human significance of leisure. While most of the essays take as their starting-point some theme in Ancient Greek philosophy, they are meant not as exegesis but as distinctive and independent contributions to live philosophizing. Written with clarity, precision without technicality, and philosophical imagination, they will engage a wide range of readers, including scholars and students of Ancient Greek philosophy and others working on more contemporary analytical concerns.

Product Details

ISBN-13:	9781107405851
Publisher:	Cambridge University Press
Publication date:	07/26/2012
Edition description:	Reprint
Pages:	214
Product dimensions:	8.90(w) x 5.90(h) x 0.60(d)

About the Author

Sarah Broadie is Wardlaw Professor of Philosophy at the University of St Andrews.

Read an Excerpt

Cambridge University Press
978-0-521-87024-5 - Aristotle and Beyond - Essays on Metaphysics and Ethics - by Sarah Broadie Excerpt

CHAPTER 1

Affecting and being affected

I INTRODUCTION

What is it for one thing to affect or to act on another? Examples come to mind: bending a bar of iron, moving a ball, melting wax. But there are other relationships one can have with an iron bar: one can also look at it, think about it, want it, approach it. Are not these too kinds of doing to it? We might say: even thinking about something is treating it in a certain way. Yet we should hardly admit that thought, sight, desire, and approaching affect or act on their objects. If we want to say that these are ways of treating things, then we have to say that not all cases of treating things are cases of affecting them.¹ To put it in the formal mode: not every transitive verb ‘φ’, i.e. not every verb ‘φ’ with active and passive voices, is such that ‘x φs y’ entails ‘x acts on or affects y’. Let us call the values of ‘φ’ for which this entailment does hold, affective verbs, and those for which it does not, non-affective. Now the question of this paper is: what are the logical characteristics of affective, as opposed to non-affective, verbs? Or to put it in the material mode, using the term ‘operation’neutrally: what are the logical characteristics of affective as opposed to non-affective operations?

The question is not: how do we know that thinking about something, or wanting it, does not affect it and that melting it does do so? – as though perhaps we cannot be sure of this. I am going to assume that we are able to distinguish correctly between affective and non-affective operations. The inquiry is simply into the formal features that constitute them one or the other.

Some more examples of affective verbs: ‘to stretch’, ‘to bruise’, ‘to shake’, ‘to heat’, ‘to cure’, ‘to calm’, ‘to compress’. Let us call this list A. Examples of non-affective verbs: ‘to remember’, ‘to mention’, ‘to learn’, ‘to understand’, ‘to love’, ‘to miss’, ‘to outgrow’, ‘to precede’. Let this be list B.

All the affective verbs so far mentioned are verbs of changing: i.e. for these values of ‘φ’, ‘x φs y’ entails ‘x changes y’. But it is a mistake to identify changing with affecting. To affect a thing is either to change it or to keep it the same, in some respect. For instance, if you hold a hydrogen balloon to the ground, this affects it no less than if you had pulled it to the ground. Hence we should add to list A such verbs as ‘to hold back’ and ‘to keep rigid’, which imply prevention from change.

To say that affecting a thing is either to change it or to prevent it from changing does not help much to answer the question we have raised. For now the question rephrases itself as: what are the distinguishing marks of changing and keeping the same, as opposed to other sorts of operations? The answer cannot be in terms of changing and unchanging truth-values. That is, it cannot be right to say: ‘φ’ is a verb of change if ‘x φs y’ entails that some proposition is now true of y that was not true before, and ‘φ’ is a verb of keeping the same if ‘x φs y’ entails that some proposition is still true of y that was true before. This lets through too much. Notoriously, if Phaedo outgrows Socrates, Socrates need not have changed; yet a new proposition is true of Socrates, namely that he is shorter than Phaedo. Again, if Phaedo keeps his eye on Socrates, it is true of Socrates that Phaedo is looking at him and that Phaedo was looking at him before; yet in doing this, Phaedo does not prevent Socrates from changing in any respect.

In what follows, I have been able to give only necessary, not sufficient conditions for a verb's being affective. These necessary conditions emerge from considering the following questions: (1) What logical characteristics must the object of an affective verb possess? (2) What are the logical features of affections, or the properties that things get through being affected? (3) What is the basis of the relation between an affective verb and the verbs which are its contraries? The meaning of contrariety between transitive verbs is explained later. These questions take up the next three sections of the paper. It will turn out, I think, that none of the verbs in list B fits the resultant conditions. However, there is one distinct class of non-affective verbs which do fit them, thus showing them not to be sufficient conditions. In the last section, I discuss this class of verbs and point, though vaguely, to what it is that marks them off from affective verbs.

II THE OBJECT OF AN AFFECTIVE VERB

(a) Affecting a thing includes changing it in some respect and keeping it the same in some respect. Now one can only keep something the same in a respect in which it could change. For instance, one cannot keep a horse equine. Therefore an object can be affected only insofar as it can change. Now change, and therefore affection, can happen only to things that exist. (So existence itself is not a property with which things can be affected. ‘To create’, ‘to destroy’, ‘to keep in existence’ are therefore not affective verbs, although they are in some ways like them.) We thus have the following argument against some of the verbs in list B above: something can be mentioned, loved, missed, thought of, desired, although it does not exist. Hence to mention, to love, etc. something, is not to affect it, and these verbs are non-affective.

(b) If ‘φ’ is an affective verb, given the truth of ‘x φs a C’, it must be possible to answer the question ‘Which C?’ And if x φs y, where ‘φ’ is affective, it must be possible to answer questions of the form ‘Is y P or not P?’ for all applicable values of ‘P’. In other words, what is affected must be a particular thing and a fully determinate thing. Indeed both these requirements, I believe, follow from the requirement that what is affected must exist. This condition, like the existence condition, rules out some of the verbs in B, but none of those in A, from being affective. One can want a child, love a good comedy, think of a sunset, without there having to be answers to the question ‘Which?’; and without there having to be answers to all possible questions about the nature of what is wanted, loved, thought of. The same is true of mentioning.

(c) Another characteristic of the verbs already ruled out in (a) and (b) is that for these values of ‘φ’, ‘x φs –’ is an opaque context. Though all Cs be D and x φs a C, it does not follow, for these values, that x φs a D. Again, if x φs y and y = z, it does not follow that x φs z. Now it seems clear that the context ‘x affects –’ is not opaque; that whether or not a thing is affected does not depend on the aspect under which it is considered. If all Cs are D and x affects a C, then it affects a D. Similarly, ‘x affects y and y = z’ is conclusive grounds for ‘x affects z’. None of the verbs in list A gives rise to such opaque contexts as we are considering.² And it can be shown that if the schema ‘x affects –’ is not opaque, any schema that implies it is not opaque either. That is to say, when ‘x φs –’ implies ‘x affects –’, i.e. when ‘φ’ is affective, ‘x φs –’ is not opaque. For suppose the opposite, that ‘φ’ is both affective and opaque. If it is affective, then ‘x φs y and y = z’ is conclusive ground for ‘x affects y and y = z’; and this is conclusive for ‘x affects z’. But if it is opaque, ‘x φs y and y = z’ does not imply ‘x φs z’. Now it could be that we have no reason to think that x affects z unless we have reason to think that x φs z. But as ‘x φs y and y = z’ by itself is no reason for the latter, by itself it is no reason for the former. So given ‘x φs y and y = z’, we cannot conclude to ‘x affects z’. So this proposition both is and is not conclusive ground for ‘x affects z’. Thus, any verb ‘φ’ such that ‘x φs –’ is opaque, is not an affective verb.

(d) The conditions above, that the grammatical object of an affective verb must designate an existent, particular and determinate, and that it must be intersubstitutable with extensional equivalents, only rule out about half the verbs mentioned in list B. Another argument is possible, which perhaps rules out more of them but which is weaker in other ways; it is as follows: ‘Many of the operations of list B can take as objects abstract entities such as propositions, concepts, rules, the meanings of words. Such things may be known, learnt, understood, etc.; they can even be loved. But abstractions cannot be said to be subject to change, and therefore not to affection. So to know, to learn, to understand, etc. is not to affect. Further, positions in space and time such as point A14 on the grid and midnight 3/11/1941 are a kind of abstraction: yet they can be preceded and approached. Hence these verbs too are not affective.’ The main weakness of this argument of course is its reliance on ‘abstract entities’. Whether or not there are such things, our ideas of them are not solid enough to build on. Is it right, for instance, to lump together apparently quite different things such as concepts, dates, and legal systems without more ado? And the assertion that abstractions cannot change itself needs as much proof and explanation as the conclusion it is supposed to prove and explain, namely that ‘to learn’ and so on are not affective verbs. As it stands, this argument is of little use.

Before going further, we should note the fact that the same verb may be affective and non-affective in different contexts. For instance, ‘carve’ is surely affective in ‘he is carving a stone’, but not in ‘he is carving a statue’, for here the carving creates the statue not affects it. ‘He threw a grenade’ implies that the grenade's position was changed, but ‘he threw a lob’ does not imply that he affected, rather that he produced a lob. Again, there is ‘x cured this man’ and ‘x calmed this stream’, as opposed to ‘x cured this man's arthritis’ and ‘x calmed a storm’. In the two latter, what was cured or calmed was thereby destroyed. A different type of case is ‘he is carving a battle-scene’, as opposed to ‘he is carving a stone’. The former may be true although there is no answer to the question ‘Which battle-scene?’ or to, say, the question ‘Is it set in Germany or not?’ Anselm Müller pointed out to me another sort of case: ‘x has swollen the number of cars on the road’, ‘x has lowered the price of oil’, as opposed to sentences like ‘x has swollen the balloon’, ‘. . . has lowered the bucket’, where the verbs are certainly affective. In the first pair of sentences, they are not, by (c) above. For suppose that the price of oil = $35 a barrel (I am assuming that this will pass as an identity statement). ‘He lowered the price of oil’ does not imply ‘he lowered $35 a barrel’. And if the price of oil = the price of tea, the first sentence does not imply ‘he lowered the price of tea’. Similar examples are: ‘x changed, or affected, the colour of the walls’, ‘x changed the direction of the telescope’. Here even ‘change’ and ‘affect’ turn out to be non-affective. If there is any force in the argument of (d) above, it can be applied here too: the verbs in ‘he lowered the price’, ‘he changed the colour’ are non-affective because the grammatical objects designate universals, abstractions. (d) would also show that ‘twist’ is non-affective in ‘he twists the meaning of “republic”’; whereas it is affective in ‘he twists a bit of putty’.

III AFFECTIONS

If a verb ‘φ’ is affective, there corresponds to it a property or set of properties with which things are affected when they are φ-ed. Let us call these affections. Examples of affections are: straightness, degrees of curvature, degrees of length, motion, colours. I shall now set out some logical features of affections. In this way we reach a necessary condition for being an affective verb: a verb is affective only if there corresponds to it a property or set of properties with these features.

(i) An affective verb is necessarily connected with a certain affection or set of affections. For example, if x straightens y, necessarily y is straight. If x heats y, necessarily y has one of a set of different possible temperatures. If x sets y in motion, y is, necessarily, moving in relation to something. This is not to say, of course, that there is any one referent in relation to which y must move, if it is set moving.

(ii) If ‘φ’ is affective, and the corresponding affection is F-ness, then if x φs y so that y is F, y would not have been F if, under the circumstances, it had not been φ-ed. Affecting a thing makes it different in some way: not necessarily different from what it was before, since not all affecting is change, but different from what, under the circumstances, it would have been had it not been affected. If something is changed to being F, it was not-F before and would have remained not-F if it had not been changed. And if it is kept F, that is, prevented from becoming not-F, it follows that it was going to become not-F and would have done so but for being prevented. Just as x must be not-F if it is to be changed to F, so it must be going to be not-F if it is to be kept F.

For instance, when the affective verb corresponds to a single affection, as ‘cure’ does to health: if x cures y, y is healthy, and would not have been healthy if, under the same circumstances, it had not been cured. If x holds y rigid, then y is rigid and would not have been if not so held. Now for cases where there is a set of several possible affections: if x warms y, y is warm to some degree, and although it might have still been warm if, ceteris paribus, nothing had warmed it, it would not have been so to the same degree. Or if it would have been so to the same degree, it would not have been to the same extent: less of it would have been that warm. Again, if x sets y moving, y is moving relatively to something, and would otherwise have been at rest in relation to that thing, although it might still have been in motion relatively to something else.

Let us put (i) and (ii) together. If ‘φ’ is affective, there is some property F-ness such that necessarily: if x φs y then y is F or is F to some degree or extent or relatively to something, etc.; and if under the same conditions y had not been φ-ed at all, it would have been not-F, or F but not to the same degree, extent or relatively to the same thing, etc. Let us call this proposition S1. It is important that the counterfactual supposition in S1 is ‘if y had not been φ-ed at all’, not ‘if y had not been φ-ed by x ’. If the latter were the supposition, S1 would be false. For suppose that x ≠ z and that x and z both simultaneously give y a push, setting it in motion, relatively to m, say. Suppose, as is possible, that the action of each by itself would have been sufficient to move y. Then x sets it moving: but if, ceteris paribus, x had not done so, still y would have been moving, relatively to m, through the action of z – not with the same velocity, but still moving. Only if nothing had moved y, would it have been at rest.

Now let us see whether all the affective verbs in list A above meet this requirement S1, and then whether S1 is strong enough to rule out non-affective verbs. For ‘to stretch’, ‘to bend’, ‘to compress’ in A, there are degrees of length, curvature, pressure. For ‘to move’ there are different positions or degrees of place. ‘To shake’ can be explicated in terms of degrees of vibration. For ‘to calm’ and ‘to melt’, there are calmness, softness and liquidity, which admit of rough degrees. There are even degrees of bruises. If y is bruised, it has more or worse bruises than it would have had otherwise. Again, if y is held back, necessarily it is in some position which it would not have been in had it not been held back. To take some more examples of affective verbs: ‘erode’, ‘push’, ‘cut up’, ‘cook’, ‘distort’. The object eroded is a certain shape and size and would not have been that shape and/or size if it had not been eroded. That which is pushed is either in motion or under pressure, and would otherwise have been at rest or under less pressure. That which is cut up is now in more pieces than it would have been: that which is cooked is now less raw than it would have been. There is no positive name for the affection things get through being cooked, apart from the participle ‘cooked’: however, this affection is a distinguishable, identifiable state, and if we coined a word for it, the word could be defined ostensively. The same is true of ‘distort’, which is affective in certain contexts;³ there is no name for the affection, apart from ‘distortedness’. But again the affection can be distinguished and pointed out. Doing this involves knowing not only what the object is like now, but also what its natural or original shape was or should be. Distortedness is in this respect unlike health or lengths or degrees of pressure.

But S1 as it stands is also satisfied by non-affective verbs. For instance, if x knows or loves y, then necessarily y is familiar or dear to x, and would not have been if nothing had known or loved it. If x approaches y, y is some distance from x, and if, ceteris paribus, nothing, and hence not x, had approached it, it would have been a different distance from x. But if ‘being dear to x’, ‘being three miles from x’, etc. are names for affections which y gets through being loved or approached, we shall have to conclude that ‘love’, ‘approach’, etc. are affective.

Again, for many values of ‘φ’, affective or not, ‘x φs y’ necessarily implies ‘y is being φ-ed’. For example, if something heats y, it is being heated; if something learns it, it is being learnt. And if nothing had heated or learnt y, y would not have been being heated, learnt.

Again, for many values of ‘φ’, affective or not, ‘x φs y’ necessarily implies ‘y has been φ-ed’.⁴ For example, if something moves y, it has been moved, if something observes or thinks about y, it has been observed, etc. And, it would seem, if nothing had moved or observed it, it would not be something which has been moved, observed: the properties of having been moved, having been observed, would not belong to it.

Someone may want to say that being learnt, or heated, having been moved or observed, being dear to x, etc., are not genuine properties: S1, one might say, must be interpreted as referring only to real properties, and then these ones present no difficulty. But of course this is no solution unless we are also told the difference between ‘real’ properties and others. I am using ‘property’ to mean anything designated, indicated, denoted, by a predicate expression. In this sense these are perfectly good properties, and show S1 as it stands to be too weak.

I shall answer these difficulties, in the reverse order, by presenting some further logical features of affections.

(iii) An affection, it seems to me, is a property which something can lose once it has it. This of course, says more than that it is a property which something has contingently. For example, it is only contingent that Tom is a war veteran: it is by no logical necessity that he went to the war and survived it. But having done so he is, necessarily, a war veteran for the rest of his days. Affections, on the other hand, can cease to belong. For instance, something is kept rigid say for ten minutes: there is no logical reason why it should not have been kept rigid for only five; in which case it would, ceteris paribus, have become bent, thus losing the affection. And a change, I think, can always be repeated; which implies that if y is once changed to being F, it can return to being not-F.

But the property of having been φ-ed, whatever the value of ‘φ’, is like being a war veteran or being Tom's widow: once something has it, it must stick. If x moves y, then y is, necessarily, something which has been moved, and cannot cease to be so. The property is fixed, just as the past is fixed.

If x φs y, y has the property of having been φ-ed. But it is only by an equivocation that we can think that y would not have had this property if nothing had φ-ed it. It is true that if nothing had ever φ-ed it, it would not have had the property. But it might have had the property even if nothing had φ-ed it now. If it had ever been φ-ed before, it would still have the property now, whether or not anything φs it now. But in S1 above, what is meant is of course that (for affective ‘φ’) if x φs y, y has some property now which it would not have had now if nothing had φ-ed it now, i.e. when x does.

Thus, even when ‘φ’ happens to be affective, the property of having been φ-ed is not an affection. So although, e.g., ‘to learn’ carries having been learnt with it, this does not mean in the least that ‘to learn’ is affective.

(iv) An affection is not tied to any single operation. For instance, if y is straight, this may be because at that instant something straightens it, or because something is keeping it straight. If it is 30° hot, this could be because it is being kept at 30° or because it is being heated and has just reached that point, or because it is being cooled. In short, if F-ness is an affection, there is no single operation φ such that ‘Fy’ necessarily implies ‘something φs y’. It is clear, then, that the property of being φ-ed, whatever ‘φ’ may be, is not an affection: for if y is being φ-ed, necessarily something φs it.

Here I would add that as well as being independent of any one particular operation, affections, or at least some of them, are independent of affective operations altogether. That is, something can have an affection F although nothing affects it so that it is F. A branch grows curved – do we want to say that something is changing its shape to that curve, or keeping it that way? These phrases are appropriate not when it follows its natural bent, but when, say, a weight pulls it into a distorted shape, or when a rigid frame holds it. Again, a moving body in free space continues its motion, although nothing moves it. Its motion stays the same, although nothing keeps it the same, and its position changes although it is not being changed.

In saying that something may be F although nothing makes or keeps it so, I do not mean that it is F independently of all conditions. There must be causally related conditions without which the body would not continue to move, etc. But these are not agents that move it. This conflicts with Hume's remark, Treatise Book I, part III, sect. XIV, that there is no logical difference between efficient and other causes. For if not, we ought to be willing to say of the free space, without which the body would not move, that it moves or keeps it moving, i.e. is an efficient cause of its motion – in the same sense in which one says that a horse keeps a cart moving.

(v) Finally, if one subject φs y so that y has a certain affection, y would have had the same affection if a different subject had φ-ed it, under the same circumstances, and in the same way. For example, if x heats y or raises y in relation to A, y has a certain temperature, is a certain distance above A. And if z, not x, had heated and raised y in the same way, the same amount in relation to the same thing, y would have had exactly the same temperature and position. This principle, that the same affection can arise through the action of different subjects, or agents, is like, or perhaps in the end the same as, the principle that an effect is logically independent of its cause.

But suppose x loves or knows y, so that y is dear or familiar to x, or rates y high relatively to A, so that y is worth more to x than A. If z, not x, had been the subject and had loved, known y equally much, rated it relatively to the same thing, y not only would not but could not have had the properties it gets when x is the subject. When x is the subject, y is dear to x; when z is, y is dear to z. Again, if x approaches y, y is a certain distance from x, which it would not have been otherwise ceteris paribus. But if, ceteris paribus, z, not x, had approached y, from the same point, the same amount, etc., y would not have been that same distance from x. Hence, being dear to x, and so on, are not affections of y, although y necessarily gets these properties if it is loved, etc.

It is true of course that whether x or z loves y, y has the property of being dear to something. But the point is that it would not be the same something. Contrast this with genuine affections, which can also be described indefinitely. X heats y so that y has some degree of heat, and if z, not x, had done what x does, y would again have had some degree of heat: the same one.

If we add these last points to S1, the result is:

If ‘φ’ is affective, there is some property F-ness, such that:

1. necessarily, if x φs y then y is F or F to some degree, etc. and would not have been so then if, ceteris paribus, nothing had φ-ed it then;
2. once something is F or F to some degree, etc., it can cease to be so;
3. ‘y is F’ or ‘y is F to some degree, etc.’ does not logically imply ‘something φs y’;
4. if y is F or F to some degree, etc. when x φs it, it would have had the same property if z, not x, had φ-ed it as x φs it, conditions being the same.

Let us call this S2.

I think it is clear that each verb in the original list A is associated with a property such as S2 lays down. To take only one example (as most of the cases seem obvious), if x holds y back, necessarily y is in a place which it would not have been in otherwise, ceteris paribus; once y is in this place, it can leave it; that y is in it does not entail that anything is holding it back; if z had held it back, not x, ceteris paribus, it would have been in the same place. Other intuitively affective verbs such as ‘cleanse’, ‘injure’, ‘push’, ‘discolour’, ‘starve’, ‘moisten’ also satisfy S2. So do the psychological verbs ‘depress’, ‘reassure’, ‘frighten’ when they are so used that the subject refers to the cause of fear, depression, etc., not to its object, as in ‘Tom depressed me with remarks about the coming revolution’ and not as in ‘The coming revolution depressed me’. For we may treat the properties: being in low spirits about A, . . . about B, . . . about C, etc., analogously to the properties: being in motion relatively to A, . . . to B, . . . to C, etc. Just as the latter set is necessarily connected with the verb ‘set moving’, so the former set is with the verb ‘depress’. Now if x depresses y, y is low-spirited about something, and to some degree, and would not have been low about the same thing, to the same degree, if nothing had depressed him then. Once low in this way, y can cease to be so. Y can be thus low although nothing at the moment depresses him, causes the depression: for instance if something has earlier depressed him, and he is still under the effects. Finally, if z, not x, had depressed y about the same thing, to the same extent, y would have been in the same state.

---

© Cambridge University Press

Table of Contents

1. Affecting and being affected; 2. Backwards causation and continuing; 3. From necessity to fate: an inevitable step?; 4. Alternative world-histories; 5. A contemporary look at Aristotle's changing Now; 6. Nature and craft in Aristotelian teleology; 7. Soul and body in Plato and Descartes; 8. Aristotle and contemporary ethics; 9. The idea of the summum bonum; 10. What should we mean by 'the highest good'?; 11. The good of practical beings: Aristotelian perspectives; 12. Taking stock of leisure.

Editorial Reviews

'I commend this stimulating collection of essays. It will certainly repay the close attention of all who are interested in Aristotle and beyond.' British Journal for the History of Philosophy

From the Publisher