The existence of powers.

Johnston, Rebekah

Introduction

Aristotle relies on and uses the concept of '[TEXT NOT REPRODUCIBLE IN ASCII.]', which I will translate throughout this paper as 'power', in a wide variety of philosophical discussions. In Metaphysics IX 1-5, Aristotle provides a detailed account of powers. The discussions in Metaphysics IX 1, IX 2, and IX 5 focus primarily on providing an account of what it is to be a power. Chapter 1 establishes the range of things that count as powers in the most proper sense. (1) The primary referent of 'power' is '[TEXT NOT REPRODUCIBLE IN ASCII.]' (1046a10-11), i.e., a principle of change in another or qua other. In addition to this primary referent Aristotle identifies that in virtue of which some item can be acted on and changed by another (1046a11-12) and that in virtue of which some item is insusceptible to being changed for the worse by another (1046a13-14). Chapters 2 and 5 further divide powers into rational and non-rational.

In addition to explaining what powers are and delineating the different sorts of powers, Aristotle argues in Metaphysics IX 3 for the existence of powers. Here, through a series of four arguments against the Megarics, Aristotle establishes that inactive powers must exist. There are, however, two important questions to which Aristotle's answers are unclear: 1) what does it mean to say that powers exist? and 2) how can it be determined that a subject does in fact possess a particular inactive power?

There are two main ways in which commentators have sought to remove this obscurity. Some argue that the 'criterion of the possible' serves as a test for determining when some subject possesses a power. (2) Others argue that the existence of powers is best understood through a dispositional analysis. I argue, however, that neither account is sufficient and that instead powers must be understood as one in number with but different in essence from various categorical features of substances. Understanding powers in this way reveals both what it means to say that inactive powers exist and when some subject has or lacks a power.

1 The Existence of Powers and the Criterion of the Possible

Although Aristotle is committed to the existence of powers, he is not clear about what it means to say that a power exists or about how to determine whether some subject possesses a power. Some interpreters attempt to clarify Aristotle's position about the latter issue by considering Aristotle's claims about the relationships between the capable and possessing a power and between the capable and the possible.

In Metaphysics V 12, Aristotle explains three different ways in which the term '[TEXT NOT REPRODUCIBLE IN ASCII.]' is used. In one sense, a thing is [TEXT NOT REPRODUCIBLE IN ASCII.] if it possesses a power (1019a33-b15). For example, a log is [TEXT NOT REPRODUCIBLE IN ASCII.] with respect to being burned if it possesses the power for being burned and fire is [TEXT NOT REPRODUCIBLE IN ASCII.] with respect to burning if it possesses the power to burn. I will translate this sense of '[TEXT NOT REPRODUCIBLE IN ASCII.] as 'capable'. In other sense, '[TEXT NOT REPRODUCIBLE IN ASCII.]' specifies 'the not necessarily false' (1019b28-9) or 'that which is not necessary but, being assumed, results in nothing impossible' (Prior Analytics I 13, 32a17-19). In this sense '[TEXT NOT REPRODUCIBLE IN ASCII.]' means 'possible'.

Some interpreters, such as Witt and Ide, argue that Aristotle's discussion in Metaphysics IX 3-4 establishes that there is a relation of mutual implication between the two senses of '[TEXT NOT REPRODUCIBLE IN ASCII.]', i.e., between the capable and the possible. (3) If there is a relation of mutual implication between the capable and the possible, then Aristotle is committed to the claims (1) if it is possible that Lucy swims, then Lucy must be capable of swimming and (2) if Lucy is capable of swimming, then it is possible that Lucy swims.

Given this relationship of mutual implication, in conjunction with the meaning of being 'capable', some claim that we can use this aspect of Aristotle's position as a test for when some subject possesses a power. (4) More specifically, if I want to know whether X possesses the power to [phi], I could determine the case as follows. Since it is the case that (1) if X possesses the power to [phi], X is capable of [phi]-ing and (2) there is a relation of mutual implication between the capable and the possible, I can determine whether X has the power to [phi] by determining whether it is possible for X to [phi].

Although I think that Witt and Ide are correct about the mutual implication claim, I do not think that this relation of mutual implication is useful for determining whether a subject possesses a particular power. The reason it is not useful is that in order to determine whether or not it is possible for X to [phi] one needs already to know whether X possesses the power to [phi]. My evidence for this claim comes from Aristotle's final argument against the Megaric claim that inactive powers do not exist. In Metaphysics IX 3 at 1047a10-20, Aristotle offers an argument for the existence of inactive powers which reveals, additionally, how he conceives of the connection between possibilities and powers. This connection shows that in order to judge whether it is possible for X to [phi] it is necessary already to know whether X possesses the power to [phi].

Aristotle says:

 [TEXT NOT REPRODUCIBLE IN ASCII.]
 (1047a10-20).

 Further, (A) if the thing that is deprived of a power is incapable,
 then (B) the thing that is not coming to be will be ([TEXT NOT
 REPRODUCIBLE IN ASCII.]) of coming to be. And (C&D) the one who
 says that the [TEXT NOT REPRODUCIBLE IN ASCII.] is coming to be or
 will come to be, will speak falsely (for this signified [TEXT NOT
 REPRODUCIBLE IN ASCII.]), (E) so that these arguments destroy both
 motion and generation. For the thing that is standing will always
 stand and the thing that is sitting will always sit; for if it sits
 it will not get up; for it, what at any rate is not capable of
 getting up, will be [TEXT NOT REPRODUCIBLE IN ASCII.] of getting
 up. (F) Therefore, if it is not possible to say these things, then
 it is evident that power and actuality are distinct (but these
 thinkers make power and actuality the same thing, on account of
 which they seek to destroy no small thing). (5)

I take the explicit claims in the argument to be the following:

A. If X does not have the [TEXT NOT REPRODUCIBLE IN ASCII.] to [phi], then X is incapable ([TEXT NOT REPRODUCIBLE IN ASCII.]) of [phi]-ing.

B. If X is not [phi]-ing, then X is [TEXT NOT REPRODUCIBLE IN ASCII.] with respect to [phi]-ing.

C. 'X is [TEXT NOT REPRODUCIBLE IN ASCII.] with respect to [phi]-ing, but is [phi]-ing' is false. If X is [TEXT NOT REPRODUCIBLE IN ASCII.] with respect to [phi]-ing, then X is not [phi]-ing.

D. 'X is [TEXT NOT REPRODUCIBLE IN ASCII.] with respect to [phi]-ing, but will [phi]' is false. If X is [TEXT NOT REPRODUCIBLE IN ASCII.] with respect to [phi]-ing, then X will not [phi].

E. Therefore motion and coming to be are done away with.

F. Since motion and coming to be exist, power and actuality are not the same, they are different. (6)

Claim A is an account of what it means to be 'incapable'. As we have seen, '[TEXT NOT REPRODUCIBLE IN ASCII.]', in one sense, means 'capable'. A subject is capable of [phi]-ing if that subject has a power for [phi]-ing. Claim A asserts a parallel sense of '[TEXT NOT REPRODUCIBLE IN ASCII.]'. (7) A subject is [TEXT NOT REPRODUCIBLE IN ASCII.] or incapable of [phi]-ing if it does not have a power for [phi]-ing. Given the way the Megaric position is presented in the opening lines of IX 3, we can assume that Aristotle takes claim A as uncontroversial. For both the Megarics and for Aristotle himself it is appropriate to claim that some subject, X, is capable of [phi]-ing if that subject has a power for [phi]-ing and it is appropriate to claim that some subject, X, is incapable of [phi]-ing if that subject does not have a power for [phi]-ing. The disagreement is not about what it means to be incapable; it is about when some subject has or lacks a power.

Claim B is more puzzling. Is it a premise in the argument or is it a conclusion? In my view claim B must be taken as a sub-conclusion which is derived from A and several unstated premises. (8) If we fail to take claim B as a conclusion, then the structure of the argument is unclear. For it is not immediately apparent why we should accept claim B. We ought, then, to take claim B as a sub-conclusion that relies on claim A and one or more unstated premises.

But what is needed to get from claim A to claim B? This question intersects with a second difficulty with claim B. If '[TEXT NOT REPRODUCIBLE IN ASCII.]' means 'incapable' in claim B as it does in claim A, then the missing premise is easily identified. Since the Megarics hold that there are no inactive powers, the argument for claim B is:

A. If X does not have the power to [phi], then X is incapable ([TEXT NOT REPRODUCIBLE IN ASCII.]) of [phi]-ing.

Implicit Premise. X has the power to [phi] iff X is [phi]-ing.

Therefore, B. If X is not [phi]-ing, then X is incapable ([TEXT NOT REPRODUCIBLE IN ASCII.]) of [phi]-ing.

On this reading, the conclusion, B, follows from claim A and an implicit premise that simply states the Megaric position concerning when X has the power to [phi]. This, however, will not do. The coherence of the passage can only be preserved if '[TEXT NOT REPRODUCIBLE IN ASCII.]' in claim B means 'impossible' rather than 'incapable'. (9) Claim D demands that we understand '[TEXT NOT REPRODUCIBLE IN ASCII.]' to mean 'impossible' in claim B.

If claim B fails to introduce 'the impossible' rather than just 'the incapable', then claim D is false. If a subject presently lacks the power to [phi] and thus is incapable of [phi]-ing, it does not follow that one must be speaking falsely if one asserts that X [phi]-s in the future. X may very well gain the power, become capable, and [phi]. If so, then claim D is false because it is based on an inappropriate move from incapable now to always incapable. If, however, claim D is false, then Aristotle is not entitled to the crucial claim, claim E, that on the Megaric view motion and coming to be are done away with.

We must, then, read claim B as introducing the sense of '[TEXT NOT REPRODUCIBLE IN ASCII.]' that means impossible. The path from claim A to claim B, therefore, must involve more than the implicit premise stated above. What is needed is an account of why the Megaric cannot respond by saying that although X is not now capable of [phi]-ing, X may become capable of [phi]-ing and [phi] at that time. The Megaric, then, may refute claim B if there is a way to establish, on the Megaric view, that (1) it is possible that even though X is not now [phi]-ing X may [phi] in the future and (2) it is possible that even though X is not now capable of [phi]-ing X may become capable of [phi]-ing. Because the Megaric holds that X is capable of [phi]-ing iff X is [phi]-ing, claims one and two amount to the same claim for the Megaric. We must consider, then, whether it is possible that some subject, X, change from not [phi]-ing/incapable of [phi]-ing to [phi]-ing/capable of [phi]-ing given the Megaric view of when something has a power.

Aristotle does not explain how he gets from claim A to the conclusion B. (10) If however, we take into account the context of the discussion we can fill in the path from A to B using Aristotelian commitments. Aristotle is concerned, in IX 3, to explore the implications of limiting the class of the capable to the active. The loss of inactive powers and thus the loss of a sense in which some X is capable of what X is not doing are immediately relevant to Aristotle's general concern to defend change against the Parmenidean attempt to eliminate change and becoming. Aristotle can claim that, given the Megaric position concerning when X is capable of [phi]-ing, if X is not capable of [phi]-ing (now) it is impossible that X will become capable of [phi]-ing and thus it is impossible that X [phi]-s because the Megarics have no way of overcoming the Parmenidean dilemma concerning change and becoming. (11)

In Physics I 8 and Generation and Corruption I 3, Aristotle explains and responds to the Parmenidean position that there is no coming to be. Parmenides holds that things cannot come to be because they must come to be either (1) from nothing or (2) from being. Neither option, however, is acceptable. Clearly something cannot come to be from nothing. But neither can something come to be from being since being already is. Given that coming to be cannot happen either from nothing or from being, Parmenides holds that coming to be is impossible.

Aristotle solves this dilemma, in Generation and Corruption, by introducing a third option. He says: '[i]n one sense things come-to-be out of that which has no being without qualification; yet in another sense they come-to-be always out of what is. For there must pre-exist something which potentially is, but actually is not: and this something is spoken of both as being and as not-being' (317b15-17). Aristotle's solution to the difficulty depends on the distinction between being actually [phi] and being potentially [phi]. This distinction depends on the existence of inactive powers. Some subject, X, is potentially but not actually [phi] if X has the inactive power for [phi]. Because Aristotle accepts inactive powers as real/existent items change and becoming are preserved.

The Megarics, however, deny that inactive powers exist and thus they cannot accept the distinction between [phi]-ing potentially and [phi]-ing actually. By eliminating inactive powers, the Megaric eliminates the distinction between [phi]-ing potentially and [phi]-ing actually and thereby eliminates coming to be. Consider the following example. In order to avoid claim B, the Megaric asserts that although Lucy is not now building = Lucy is incapable of building = Lucy lacks the power to build, Lucy may change such that she is building = is capable of building = has the power to build. If the Parmenidean challenges the Megaric to explain from what the change came to be, the only answers the Megaric has available are from nothing or from being. Because the Megaric does not allow for the existence of inactive powers and thus cannot distinguish potentially building from actually building, the Megaric must falter on the Parmenidean challenge.

Aristotle, therefore, has good reason to conclude that for the Megarics, if X is not [phi]-ing, then it is impossible for X to [phi]. The Megaric rejection of inactive powers and the attendant restriction of the capable to the active make it the case that it is impossible for anything to change.

Claims C and D represent an unacceptable objection to claim B. The Megarics may try to salvage their position by claiming that although X is neither capable of [phi]-ing nor actually [phi]-ing X may nevertheless [phi] in the future. This objection is plausible if one fails to note the introduction of 'impossible' into claim B. Aristotle, however, uses claims C and D to draw attention to this point. Aristotle says that 'and the one who says that the impossible [TEXT NOT REPRODUCIBLE IN ASCII.] is coming to be or will come to be, will speak falsely (for [TEXT NOT REPRODUCIBLE IN ASCII.] signified this)' (1047a12-14). While one would not speak falsely in such a case if B lacked the modal implication of impossibility, Aristotle draws attention to this with the claim that 'this is what [TEXT NOT REPRODUCIBLE IN ASCII.] meant.' (12) Claims C and D, then, represent an objection that has no force and thus E follows. Claim E is not a different claim from claim B. It is simply a restatement of claim B in more explicit terms.

Claim E, however, is unacceptable for there is motion and change. One may object that perhaps the Megarics do not think that motion and change exist. Aristotle, however, does think that change and becoming exist and thus he is entitled to reject the Megaric view because it eliminates change and becoming. Furthermore, McClelland points out that the initial arguments against the Megarics proceed on the assumption that the Megarics accept change and becoming. Building occurs and seeing and hearing and tasting occur as well. (13) So, the Megarics themselves, at least according to Aristotle's report, accept change and becoming yet cannot account for it given their position on the capable.

Aristotle's argument, here, reveals why the mutual implication between the capable and the possible cannot serve as a test for determining whether a subject possesses a power. In the refutation of the Megaric position, Aristotle shows that without inactive powers motion and change are eliminated because it is impossible for a change to happen without an inactive power. Since, in general, motion and change are impossible without inactive powers, it is reasonable that a particular motion or change is also impossible without the relevant power. The significance of this claim is that if one tries to apply the possibility test in order to determine whether a subject possesses a power one must already know whether the subject possesses the power. For instance, if I want to know whether it is possible for Lucy to build a house, I need to consider whether anything impossible results if I assume that Lucy builds a house. One impossibility I need to be concerned about is whether Lucy builds without possessing the building power. So, since I cannot determine whether it is possible for Lucy to build unless I already know whether she possesses the building power the possibility test does not provide an answer to whether or not Lucy has the power to build.

2 How to Understand the Existence of Powers

Another way in which interpreters try to clarify Aristotle's position about the existence of powers is according to a dispositional analysis. In 2.1 I argue that although powers can be understood according to a dispositional analysis this method is insufficient as a theory of powers and is not exhaustive of Aristotle's position on powers. In section 2.2 I argue that Aristotelian powers are one in number with but different in essence from categorical features of the world.

2.1 Dispositional analysis

Some interpreters take Aristotle's discussion of the activization conditions of powers in Metaphysics IX 5 to reveal not only Aristotle's position on how to specify a power but also when some X has a power. Beere says with reference to desire and the appropriate conditions that 'these two criteria are formulated as a specification of when an ability is necessarily exercised. But they turn out to be criteria for when something has an ability at all.' (14) According to Beere, '[s]omething has the non-rational ability to [phi] if, given that the relevant conditions obtain, it necessarily [phi]-s. Something has the rational ability to [phi] if, given that the relevant conditions obtain and its decisive desire is to [phi], it [phi]-s.' (15) On this view, powers are understood as dispositions of substances. Witt, as well, takes powers to be dispositions of substances. She says that 'Aristotle's discussion of agent and passive powers strongly suggests that they can be given a dispositional analysis.' (16) On her view '[t]o say that fire has the agent power of heating is to say that under certain conditions (which can be given a general or lawlike specification) fire will heat another object.' (17)

While both Witt and Beere take dispositional analyses to reveal when X has a power and what it is to be a power, there are two questions that I wish to address: (1) does Aristotle's treatment of powers lend itself to a dispositional analysis? and (2) is this sort of analysis (a) a sufficient way to analyze powers? and (b) exhaustive of Aristotle's treatment of powers? Aristotle's treatment of powers does lend itself to a dispositional analysis. This sort of analysis, however, does not exhaust Aristotle's treatment of powers and it is not a sufficient analysis of powers. This sort of analysis makes powers mysterious items in Aristotle's ontology and it is not sufficient for the task that Beere and Witt attribute to it. It cannot tell us, in controversial cases, when X has a power to [phi] and when X lacks a power to [phi].

2.1a Aristotelian powers can be given a dispositional analysis. Aristotle's discussions of agent and patient powers in Metaphysics book IX proceeds primarily (but not exclusively) by way of three sorts of claims. First, Aristotle describes powers in terms of the activities they are powers for. He provides numerous examples--powers may be for heating or being heated, cooling or being cooled, burning or being burned, crushing or being crushed, curing or being cured, harming or being harmed, building or being built.

Second, Aristotle describes powers in terms of the substances to which they belong. For the most part, Aristotle uses the preposition '[TEXT NOT REPRODUCIBLE IN ASCII.]' (in) to describe the relation between a power and the subject to which it belongs. This occurs at 1046a12 and 1046a22 to describe the relation between patient powers and their subjects and it occurs at 1046a26-7 to describe the relation between both rational and non-rational agent powers and their subjects. It occurs as a general claim about powers at 1048a3-5. Aristotle makes a similar claim at 1046a36-b2, but he replaces the preposition with '[TEXT NOT REPRODUCIBLE IN ASCII.]'.

Finally, Aristotle describes powers in terms of their activization conditions. In Metaphysics IX 5, he tells us that in the case of non-rational powers, when agent and patient meet in the appropriate way, the one must act and the other must be acted on (1048a5-8). For instance, when fire comes into contact with cotton, the fire must burn the cotton and the cotton must be burned by the fire. The case of rational powers is more complicated. Rational powers, unlike non-rational powers, are for contraries. The doctor, in virtue of the medical art, can both heal and harm a patient. Because rational powers are for contrary effects, the mere meeting of agent and patient cannot necessitate the activization of the power. Because it is for contraries the agent would, in such a case, produce both contraries in the same subject at the same time. And this is impossible. In the case of rational powers, an additional factor, desire or choice, is needed in order for the rational power to be activated. When a rational agent power meets a patient power in the appropriate way and desires one contrary rather than the other, then the rational agent must act and the patient must be acted on. Although the conditions required for each power to manifest in the particular activity it is a power for are different, both sorts of powers are discussed in terms of their activization conditions.

These three aspects of Aristotle's account of powers, when combined, reveal that Aristotelian powers can be given a dispositional analysis. According to this analysis, to say that a certain substance, X, has a power to [phi] is to say that 'X will [phi] if the appropriate conditions are met'. The appropriate conditions are different for rational powers as opposed to non-rational powers, but the basic structure of the claim is the same for both.

2.1b Is a dispositional analysis sufficient as an explanation of the existence of powers? A dispositional analysis of this sort, while not incorrect, is somewhat uninformative. On this view, Aristotelian powers remain mysterious. Powers, for Aristotle, are taken to exist both when they are inactive and when they are active. The builder has the building power both when she is building and when she is sleeping. The tree has the power to be burned both when it is being burned and when it is blooming in the garden. If, however, we give a dispositional analysis of powers in terms of what a substance does in certain circumstances, these powers, especially when they are inactive or unmanifested, remain mysterious. The conditional statement does not reveal in virtue of what the substance [phi]-s in the appropriate circumstances. (18)

2.2 Dispositional analysis does not exhaust Aristotle's treatment of powers

In order to remove the obscurity surrounding powers, we need to consider whether Aristotle's theory of powers can be developed in another direction. In Metaphysics IX, in addition to explaining powers in terms of their subjects, activities, and activization conditions, Aristotle makes claims of the form 'X is a power'. (19) In IX 2, at 1046b2-3, he says 'all the arts, i.e. the productive sciences are powers; for they are principles of change in another or qua other.' At various other places in Metaphysics IX he replaces 'power' with a particular example. For instance at both 1046a26 and at 1046b5 Aristotle gives heat and the art of building as examples of agent powers and at 1047a4-5 he gives, as examples of powers, perceptible qualities in general and in particular cold, hot, and sweet. Such claims can be found outside of the Metaphysics as well. For instance, in Meteorology IV 1, 378b30, he calls hot, cold, moist, and dry powers.

So, in addition to a dispositional analysis constructed from the subject to which a power belongs, the activity it is for, and its activization conditions, we can also detect, in Aristotle's discussion of powers, claims of the form 'X is a power'. These claims are promising insofar as they focus our attention on the powers rather than on the subjects that have them. 'Is' claims in Aristotle, however, have many meanings. (20) The claim 'X is (a) Y' could mean that X is the possessor of the attribute Y as in 'the apple is red' or 'the wagon is dirty'. Alternatively, 'X is (a) Y' could mean that Y is the kind (genus or species) to which X belongs as in 'Socrates is a human' or 'Felix is a cat'. Similarly, 'X is (a) Y' could mean that Y is the category to which X belongs as in 'red is a quality' or 'Rover is a substance'. We need, then, to figure out what sort of claim Aristotle is making when he says 'X is a power'. In what follows, I argue that we should not understand claims of the form 'X is a power' as expressing the relationship of subject to attribute. Moreover, I will argue that, although the structure of the claim 'X is a power' is similar to claims like 'Socrates is human' or 'Rover is a substance', neither of these options can capture the correct relation that holds between 'X' and 'power' in the phrase 'X is a power'. We should, instead, understand the claim 'X is a power' as analogous to claims such as 'fire is an element', 'Socrates is one', 'the art of building is a cause', and 'the road from Thebes to Athens is the road from Athens to Thebes'--that is, we should understand powers as one in number with, but different in essence/ being from that of which they are predicated. (21)

In all these cases, cases which I think are analogous to cases like 'hot is a power', 'the art of building is a power', and 'dry is a power', the 'is' expresses numerical identity but essential difference. The road from Thebes to Athens is not a numerically different road from the road from Athens to Thebes, but to be the latter is not the same in essence as to be the former. Likewise, a power is not numerically distinct from the item that fills in the placeholder X in the claim 'X is a power' but to be a power is not the same as it is to be X.

In order to see why we should understand claims of the form 'X is a power' as expressing the 'one in number but different in essence' relationship it is useful to consider both elements of the locution. That is, 1) what evidence is there that Aristotle takes both X and power in the claim 'X is a power' to be numerically the same 2) what evidence is there that Aristotle takes X and power to be different in essence.

Aristotle's manner of arguing against the Megaric position in Metaphysics IX 3 provides evidence for the numerical identity claim. Aristotle argues that the Megaric position is absurd because it requires that arts, such as the art of building, are acquired in a moment (when one starts building) and are lost in a moment (when one stops building) and this is not how arts are gained and lost (1046b34-7a4). What is important about this argument is that Aristotle treats the issue of the acquisition and loss of a power as answerable through and thus somehow equivalent to issues about the acquisition and loss of an art. Since he makes his argument about the acquisition and loss of powers by means of a discussion of the acquisition and loss of an art, it must be the case that he considers the art and the power to be, at least numerically, the same item.

One might object to this analysis on the grounds that the numerical identity claim is too strong. Could it not be the case that Aristotle treats the question of the acquisition or loss of a power through the acquisition or loss of an art not because the art and the power are numerically the same item but because the power is an essential attribute of the art? In this case the power is not the same item as the art but it must nevertheless accompany the art.

If this suggestion is correct, then the claim 'X is a power' should be understood as expressing the relation of subject to attribute. But the 'X' in the claim 'X is a power' is not a proper subject--it is, in each case, something that belongs to a subject, i.e., an attribute of a substance (e.g. hot, cold, dry, art). To treat these items as subjects that can bear accidents is equivalent to claiming that properties can bear accidents. Aristotle's analysis, in Metaphysics V 7, of such claims as 'just is musical' or 'the musical is white', and his discussion in Posterior Analytics I 22, reveal that he is skeptical about this position. To say that the musical is just is not to say that the musical has the property just--rather it is to say that the musical and the just are both accidents of the same substance. If X and power are numerically distinct items then we will be forced either to take X as a subject that can bear accidents which is problematic or we will be forced to treat X and power as two unrelated attributes of the same substance, like just and musical, in which case Aristotle couldn't use the arguments he uses in IX 3 against the Megarics.

There are, then, reasons to think that 'X' and 'power' are numerically the same item. Just as the road from Athens to Thebes and the road from Thebes to Athens is numerically the same road so too, for example, the art of building and the power to build are numerically the same item and the quality hot and the power to heat are numerically the same item. Why, then, must we take them as different rather than the same in essence?

It is clear, as I argued above, that the phrase 'the art of building is a power' cannot be interpreted as employing the subject-attribute relation. Why, though, must it be interpreted according to the one in number but different in essence relation rather than according to the genus/species to particular relation or according to the category to particular relation. In other words, why not interpret the claim 'X is a power' as expressing a claim either like 'Socrates is a man' or like 'Rover is a substance' where the predicate tells us what kind of thing the subject is? For instance, why not take 'power' as specifying the class to which hot, dry, and art belong?

There is some evidence that makes the 'class' interpretation plausible. Hot and cold are described in Generation and Corruption at II 1, 329b24-5 and in Meteorology IV 1 at 379b13 as [TEXT NOT REPRODUCIBLE IN ASCII.] (capable of acting). Dry and moist, on the other hand, are described in these passages as [TEXT NOT REPRODUCIBLE IN ASCII.] (capable of being acted on). Furthermore, in Nicomachean Ethics VI 4, at 1140a6-8, Aristotle describes an art as 'a reasoned state of capacity to make', i.e., as [TEXT NOT REPRODUCIBLE IN ASCII.]. These characterizations of hot, cold, and art as what can act and of dry and moist as what can be acted on, are not haphazard descriptions. In the Meteorology passage, at 378b20, Aristotle says that [TEXT NOT REPRODUCIBLE IN ASCII.] and [TEXT NOT REPRODUCIBLE IN ASCII.] are the accounts we give when we define ([TEXT NOT REPRODUCIBLE IN ASCII.]) their natures ([TEXT NOT REPRODUCIBLE IN ASCII.]). Furthermore, in the Nicomachean Ethics passage, Aristotle indicates that [TEXT NOT REPRODUCIBLE IN ASCII.] captures 'what it is' ([TEXT NOT REPRODUCIBLE IN ASCII.]) to be an art.

Aristotle's characterization of these items as [TEXT NOT REPRODUCIBLE IN ASCII.] and [TEXT NOT REPRODUCIBLE IN ASCII.] makes it tempting and at least plausible to claim that these items are, essentially, powers, i.e., that being a power is the essence of hot, cold, dry, moist, art, etc. There are, however, serious problems with both the genus/species interpretation and the category interpretation.

First, a power, by definition, is a principle of change. Its status as a principle depends on the existence of that of which it is a principle, i.e., change. The items that Aristotle identifies as powers, i.e., hot, cold, sweet, art, etc., do not depend on change. When Aristotle speaks of such items in, for instance, Metaphysics VII 1, he treats them as dependent on the substances to which they belong but he does not cast them as dependent on the existence of change.

Since being a power is dependent on the existence of other activities, not just on the existence of substances as possessors, it is conceivable that hot could remain what it is, i.e., a quality of a body, without being a power. Aristotle's analysis of perceptibles, one sort of power, in Metaphysics IV 5 at 1010b30-11a2 confirms this. Without ensouled beings perception does not exist. And without perception there are no perceptibles--but the substrata--the items that give rise to perception and would be perceptibles if perception existed, do continue to exist. This analysis implies that although the property that we call visible is a power, that property, as substrate, could remain what it is even though perception and thus that quality's status as a power is compromised. If the subject can remain what it is while no longer being described by the predicate, then it cannot be the case that the predicate picks out the genus or species of the subject. Since the substratum remains what it is even if it were to fail to be a power, then it cannot be the case that 'power' specifies the genus or species of that substratum.

Second, there is more than one way of saying 'what' something is. Sometimes when Aristotle speaks of saying 'what' something is he means that we should say what category it belongs to, i.e. substance or quality or quantity. In other places, however, when he speaks of saying 'what' something is he means that we should give the definition, i.e., two-footed rational creature or three-sided plane figure. If one says that X is a power, this is more like saying that it is a quality or quantity than it is like giving the definition. 'Power', however, like 'element', 'cause', and 'one', the items Aristotle identifies as being one in number with but different in essence from the subjects of which they are predicated, is not one of the categories Aristotle identifies.

As I suggested above, I think that we should understand powers as being one in number with but different in essence from that of which they are predicated. On this view a power is numerically the same as some other item, but nevertheless different in being/essence from that item. In Metaphysics X 1, after setting out the various meanings of the word 'one', Aristotle makes the following claim. He says: '[i]t is necessary to consider that one must not assume that to say what sort of thing is said to be one and what it is to be one and what its formula is are the same ... ' (1052b2-4). In this discussion he identifies several other items such as 'element' and 'cause' as items of this sort, items for which it is not the same thing to say what it is predicable of and what its definition is (1052b4-16).

According to this discussion, if I say that a man is one, I am saying: a) that there is only one item, a man, b) that to be that item is not the same as to be 'one', i.e. that the definition of 'man' and the definition of 'one' differ and c) that 'one' has an account of its own. Likewise with 'cause' or 'element'; if I say that fire is an element, I am saying a) that there is only one item, i.e., fire, b) that to be that item, i.e., to be fire is not the same in essence as being an element, and c) that 'element' has an account/essence of its own. While Aristotle does not specifically mention powers as items of this sort, he indicates, at 1052b16, that this list (cause, element, one) is not exhaustive of terms of this sort.

Since the claim 'X is a power' cannot be interpreted as expressing the relation of subject to attribute or particular to species/genus or particular to category, and since 'power' like 'cause', 'element', and 'one' function like class terms but are not identified by Aristotle as amongst the categories, it seems that we should understand the claim 'X is a power' as expressing the 'one in number but different in essence' relation. According to this analysis if I say that 'X is a power', I am saying that a) there is only one item, X, b) that to be X is not the same in essence as to be a power, and c) that 'power' has an account/essence of its own. This analysis allows one to identify the items in the world that are powers without reducing what it is to be a power to what it is to be these items and thereby eliminating powers from Aristotle's ontology.

Understanding a power as one in number with but different in essence from some quality such as hot, dry, art, etc., allows us to focus specifically on the item the substance has rather than on hypothetical situations which may tell us that something has a power but cannot tell us, specifically, what item 'power' picks out. To say that the builder has the building power when she is sleeping is to say that she has the art of building. The art of building is a science and as such is a quality of a substance. This item, however, is also properly described as a power, i.e., as a principle of change in another or qua other.

3 Controversial cases of unmanifested powers

I suggested above that the dispositional analysis Beere and Witt provide cannot, as they claim, be the criterion for determining when some X has a power. While this analysis cannot reveal when some X has an unmanifested power, understanding powers in terms of 'is' claims can solve this difficulty.

Although, according to the discussion in Metaphysics IX 5, when agent and patient meet in the appropriate way the one must act and the other must be acted on, it is not always that case that the appropriate circumstances are present. This is the meaning of Aristotle's commitment to the claim that inactive powers exist. Since there will be times when some subject does not manifest its power, how will we distinguish those cases from cases where the subject simply doesn't have the power? The dispositional analysis that Beere and Witt provide will only work if we ignore this distinction. Once we accept this distinction it is insufficient for telling us, in hard cases, when some X has a power and when it lacks that power. For example, we may ask whether wood at the bottom of the ocean has or lacks the ability to be burned by fire. Clearly, wood at the bottom of the ocean cannot be burned by fire right now but does it have or lack the inactive power to be burned?

A dispositional analysis constructed from the subject, the activity, and the actualization conditions cannot provide an answer to our question. Start with the claim: 'Wood has the power to be burned = under certain conditions (which can be given a law-like specification) wood will be burned'. In order to fill in the conditions we can consider cases where wood is burned and cases where it isn't. So, for instance, wood doesn't burn when it is wet but it does burn when it is dry, wood doesn't burn when it is not in contact with fire and it does burn when it is in contact with fire and wood doesn't burn where there is insufficient oxygen and wood does burn where sufficient oxygen is present. We can now fill in the relevant conditions and generate the following conditional statement: wood has the power to be burned: if under certain conditions (i.e., the wood is dry, the wood is in contact with fire, and there is sufficient oxygen), wood burns. To determine whether wood at the bottom of the ocean has the power to be burned we must ask: would the wood burn if it was dry and came into contact with fire and there was sufficient oxygen? If yes, then the wood at the bottom of the ocean has the power to be burned: if 'no', then the wood at the bottom of the ocean does not have the power to be burned. The problem with this is clear. It can reveal nothing at all about the wood at the bottom of the ocean--for one of the conditions is the removal of the wood from the bottom of the ocean.

If it were the case that all the powers substances have belong to them necessarily or always, just in virtue of what they are, then the conditional analysis would tell us something about the wood at the bottom of the ocean. But Aristotle makes it clear, in both Metaphysics IX 3 and IX 5, that many powers can be gained and lost. Removing the wood from the bottom of the ocean, therefore, is problematic, because that process may cause the acquisition of the power about which we are concerned.

The question: 'does wood at the bottom of the ocean have the power to be burned?' can be answered in an alternative manner if we take seriously Aristotle's claims about certain properties being powers. That is, to ask whether the wood at the bottom of the ocean has the power to be burned is to ask: does the wood at the bottom of the ocean have the specific property which is one in number with the wood's power to be burned? That is, we can ask: is the wood dry? The wood is not dry, so it does not have the power to be burned. This manner of questioning does not require that we take subjects as having their powers always and forever and it allows us to give different answers about the same subject in cases that may seem, initially, to be structurally similar. For instance, we can answer 'no' to the question 'does the wood at the bottom of the ocean have the power to be burned' but we can answer 'yes' to the question 'does wood locked in a metal box have the power to be burned'. In both cases, the wood will not be burned right now--but in the former case it is because the wood lacks the power (dryness) and in the other case it is because the conditions are not right. If we switch our emphasis from talking about mysterious possessions certain subjects have to talking about specific, categorical, actual features of those subjects, then we can determine whether a subject has a power at a certain time without relying on possibility claims or on conditional statements.

Conclusion

In Metaphysics IX 3 and IX 4, Aristotle continues his consideration of powers in the strict sense. Although he provides details, in IX 1, IX 2, and IX 5, about the scope of power in the strict sense and about the various referents of power he does not, in these chapters, argue specifically for the existence of powers. In IX 3, he undertakes the task of showing, not only that powers exist but that inactive powers exist. The existence of powers is important; without powers, Aristotle argues, nothing is capable of doing anything other than what it is presently doing. Change depends on the existence of inactive powers. Aristotle's position, then, is that inactive powers exist and thus that substances, as possessors of powers are capable of doing things they are not now doing.

Although Aristotle is clearly committed to the existence of powers, he does not provide a detailed account of how we should understand the claim. He does not specifically address the question of what it means to say that a power exists and he does not specifically address how it can be determined whether a subject possesses or lacks an inactive power. The criterion of the possible and dispositional analyses are not adequate for the tasks of specifying what it means to say that powers exist and of determining whether or not a subject possesses an inactive power. The first is unsatisfactory because in order to use the test one must already know whether the subject has or lacks the power under consideration. The second is unsatisfactory because it focuses on the subjects that possess the powers rather than on the powers themselves and because it cannot be used to distinguish cases where a subject lacks a power from cases where the conditions for manifestation simply have not been met.

Although these two strategies are unsatisfactory, Aristotle also makes claims of the form 'X is a power' where X picks out a specific categorical feature such as hot, sweet, or art. By understanding powers as one in number with but different in essence from these items it becomes clear both what it means to say that some subject possesses a power and how to tell when the subject does in fact possess this power. (22)

Bibliography

Aristotle. The Complete Works of Aristotle. Vol. I and II. Ed. Jonathan Barnes. Princeton, NJ: Princeton University Press 1991.

Aristotle. Analytica Priora Et Posterioria. Ed. W.D. Ross. Oxford: Clarendon Press 1964.

Aristotle. Aristotle's Metaphysics: A Revised Text with Introduction and Commentary. Vol. 1 and II. Ed. W.D. Ross. Oxford: Clarendon Press 1924.

Aristotle. De Generatione et Corruptione: A Revised Text with Introduction and Commentary. H.H. Joachim. Oxford: Clarendon Press 1922.

Aristotle. Ethica Nicomachea. Ed. I. Bywater. Oxford: Clarendon Press 1894.

Aristotle. Physica. Ed. W.D. Ross. Oxford: Clarendon Press 1950.

Beere, J. The Priority of Active Being: an interpretation of Aristotle's Metaphysics. Dissertation. Princeton University 2003.

Cleary, J.J. 'Powers That Be: The Conception of Potency in Plato and Aristotle'. Methexis XI (1998) 19-64.

Hintikka, J. Time and Necessity. Oxford: Clarendon Press 1973.

Ide, H. 'Dunamis in Metaphysics IX'. Apeiron XXV (1992) 1-26.

McClelland, R.T. 'Time and Modality in Aristotle, Metaphysics IX.3-4'. Archiv fur Geschichte der Philosophie 63 (1981) 130-149.

Molnar, G. Powers: A Study in Metaphysics. Oxford: Oxford University Press 2003.

Witt, C. Ways of being: Potentiality and Actuality in Aristotle's Metaphysics. Ithaca, NY: Cornell University Press 2003.

(1) Part of Aristotle's project in Metaphysics IX is to develop a new sense of power, a sense in addition to its proper sense as a principle of change. I will not be concerned, here, with the development of this new sense.

(2) The 'criterion of the possible' states that X is possible if, when it is assumed to be actual, no impossibility results. See Prior Analytics I 13, 32a17-19, Metaphysics IX 4, 1047a24-6.

(3) For the arguments in support of this claim see Witt (2003, 30-5) and Ide (1992).

(4) Witt (2003, 25) says that '[t]here is little doubt that Aristotle has here adapted the principle of possibility to serve as a rule for determining whether a substance has a power.' Cleary (1998) does not explicitly make this point but it follows from his claim that Aristotle's use of '[TEXT NOT REPRODUCIBLE IN ASCII.]' at 1047a24 is 'a very general definition of potency as what something has if there is nothing impossible in its attaining the activity of which it is said to have the power' (27).

(5) The first instance of '[TEXT NOT REPRODUCIBLE IN ASCII.]' clearly means incapable. I have, however, left the other instances of '[TEXT NOT REPRODUCIBLE IN ASCII.]' untranslated because it is unclear whether they should be translated as incapable or impossible. I argue below that they must be translated as impossible.

(6) I take Aristotle to be committed here to claims A, C, D, and F. Claims B and E are the problematic results of the Megaric position that inactive powers do not exist.

(7) See Metaphysics V 12 for this sense of '[TEXT NOT REPRODUCIBLE IN ASCII.]'.

(8) Beere (2003, 125-9) takes claim B as a conclusion. I discuss his interpretation below. See note 10.

(9) Most commentators agree.

(10) Beere (2003, 127-8) says that in the background, as support for claim B, are the claims 1) that [TEXT NOT REPRODUCIBLE IN ASCII.] are intrinsic features of their possessors, 2) that change depends on these features, and 3) that the objects that possess these [TEXT NOT REPRODUCIBLE IN ASCII.] are 'completely ready' to bring about or undergo changes before they occur. I agree with Beere's claims, but I do not think that they are sufficient to show why Aristotle is entitled to claim B. In order to establish claim B an account of why the intrinsic properties are required for change is needed. My explanation attempts to clarify this point by suggesting that on the Megaric view, the distinction between [phi]-ing potentially and [phi]-ing actually is lost. Witt (2003, 28-30) rightly claims that in claim B '[TEXT NOT REPRODUCIBLE IN ASCII.]' must mean impossible because the argument demands it. It is unclear, however, what Witt takes the path from claim A to claim B to be.

(11) One may object that the Megarics have no interest in overcoming this dilemma, that they support the elimination of motion and change.

(12) Hintikka (1973, 104) takes this claim as evidence that Aristotle holds the principle of plenitude. In particular, that Aristotle defines the impossible as what never is. Hintikka, however, notes that '... the passage is perhaps somewhat inconclusive, for [TEXT NOT REPRODUCIBLE IN ASCII.] might possibly be a weak term here, to be translated in terms of "indicating" rather than "meaning" ' (104). In my view Aristotle is not giving a definition of the impossible as 'what never is'. Rather, he is pointing out that in claim B '[TEXT NOT REPRODUCIBLE IN ASCII.]' means impossible. Witt (2003, 30) points out that '[t]he use of the imperfect tense in the phrase "for that is what adunatos meant" (at 1047a13) indicates that Aristotle thinks he is stating a generally accepted truth, and the two meanings of adunatos (incapable and impossible) have that status in ordinary Greek.' Aristotle is not, Witt claims, relying on the controversial position that the principle of plenitude is the basis of the modal concepts.

(13) See McClelland (1981, 136).

(14) Beere (2003, 101)

(15) Beere (2003, 101)

(16) Witt (2003, 42)

(17) Witt (2003, 42)

(18) Molnar (2003, 84-9) calls conditional analyses of this sort 'naive'. They are problematic because they reveal absolutely nothing about what the object has that makes the response follow the stimulus. If this is all there is to Aristotelian powers, then Aristotle's account is vulnerable to this sort of objection.

(19) Beere and Witt both recognize that Aristotle makes claims of the form 'X is a power' but they do not develop this angle of analysis.

(20) I wish to thank an anonymous reviewer from Apeiron for insightful comments on an earlier version of this section which assisted me greatly in clarifying my position.

(21) See Metaphysics X 1 and Physics III 3.

(22) I wish to thank Kara Richardson, Lloyd Gerson, Brad Inwood, Jennifer Whiting, Marguerite Deslauriers and an anonymous reviewer at Apeiron for comments on an earlier draft of this paper.

Rebekah Johnston

Department of Philosophy

Wilfrid Laurier University

75 University Avenue West

Waterloo, ON

N2L 3C5

rjohnston@wlu.ca