2. Two Kinds of Two-Way Power

Two-way powers theorists advance two distinct views about the nature of the opposed results that agents are capable of achieving. In tracing these ideas back to Aristotle, these theorists accordingly emphasize different aspects of Aristotle’s thought. According to one prominent view, agency is a two-way power to act or refrain from acting.Footnote ³ For instance, Maria Alvarez, writes ‘if acting is a two way power’ then ‘whenever it is in our power to act it is also in our power not to act’ (Alvarez, Reference Alvarez2013, p. 110). Helen Steward similarly identifies two-way powers as those that ‘leave open two ways in which the agent might proceed: to act or not to act’ (Steward, Reference Steward2020, p. 354).Footnote ⁴ On this view, the two ‘ways’ in which a power manifests are opposed as contradictories;Footnote ⁵ such a power is

A second proposal by Kim Frost attempts to remedy what he regards as the shortcomings of the first proposal. According to Frost, the two exercises of Powers for Contradictories are too different to assign to a single power. For there should be a unified description of the distinct manifestations of such a power, if it is a single power; if there is not, then we are likely dealing with a plurality powers. However, as is reflected in the first view’s tendency to specify the two ‘ways’ disjunctively, there does not appear to be a unified description of the manifestation types (Frost, Reference Frost2020, p. 1142 & p. 1149). In this respect, Aristotelian rational powers are more promising candidates for being two-way powers. For Frost, an Aristotelian rational power is

Both exercises of a Power for Contraries are cases of bringing about a change: medicine, for example, is a power to bring about health and disease in a patient (Frost, Reference Frost2013, pp. 613–14, Reference Frost2020, p. 1148). However, these two changes are unified in a way that justifies assigning them to a single power. Frost, summarizing Aristotle’s remarks at Θ.2, 1046b7–15 writes ‘the healer knows how to produce health, but such knowledge consists in grasp of a logos or rational account, and so is by implication knowledge of the privation of health, and so can also be employed to that contrary end’ (Frost, Reference Frost2020, p. 1148). The view that rational powers are Powers for Contraries is also widespread in the scholarship on Metaphysics Θ. According to Charlotte Witt, rational powers ‘are capable of two contrary effects’ (Witt, Reference Witt2003, p. 65). Stephen Makin similarly writes that ‘a rational capacity to bring about φ [is also] the capacity to bring about the privation of φ’ (Makin, Reference Makin2006, p. 47). And Jonathan Beere glosses a rational power as ‘a single power to bring about either of two [sc. contrary] results’ (Beere, Reference Beere2009, p. 78). Because the locutions ‘capable of’ and ‘power to’ are typically used to identify the exercises of a power (what the power in question brings about), these scholars seem to attribute to Aristotle the view that the exercises of a single rational power are both cases of producing or bringing about one or another contrary change. I will consider this second proposal in §§4–7. In the following section, however, I first evaluate the contention that Aristotle countenances Powers for Contradictories.

3. Powers for Contradictories

Alvarez finds Aristotle’s commitment to a Power for Contradictories in a remark that Ross’s translation renders: ‘where it is in our power to act it is also in our power not to act’ (NE III.5, 1113b7–8; Alvarez, Reference Alvarez2013, p. 108).Footnote ⁶ The expression translated ‘in our power’ (ἐφ᾽ ἡμῖν) is not, however, connected to Aristotle’s word for ‘power’ (δύναμισ) and in what follows I will instead translate it ‘up to us.’ But even apart from this linguistic consideration, context does not support understanding this statement as a claim about a single power with two opposed manifestations. This claim supports Aristotle’s conclusion that vice is up to us (ἐφ᾽ ἡμῖν, 1113b6). Before arguing that vice is up to us, Aristotle first shows that virtue is (III.5, 1113b3–6). This first argument begins with the previously-established conclusion that

Aristotle adds that

Apart from these two explicit premises, Aristotle also implicitly relies on his earlier conclusion that

Aristotle’s argument for (3) clarifies that the expression ‘the things up to us’ denotes one of four domains of things caused: in addition to things caused by nature, necessity and chance, Aristotle maintains that human beings are causes of a distinct class of effects, which he calls ‘the things up to us’ (1112a30–33). Because we do not deliberate about or choose things caused by nature, necessity or chance, Aristotle maintains that deliberation and choice exclusively take this fourth class of effects as their objects (1112a21–34; cp. 1113a2–4). From (1) and (3) it follows that

And from (2) and (4) Aristotle can infer that

That is, virtuous actions belong to the class of things of which we, rather than nature, necessity or chance, are the cause.

Having shown that virtuous actions are up to us, Aristotle adds: ‘and in a similar way vice [is] also [up to us]; for

The purpose of (6) – which Alvarez cites as support for Aristotle’s commitment to a Power for Contradictories – is to justify including vicious actions among the things up to us. Aristotle additionally relies on the premise that

From (5)–(7) Aristotle can infer that vicious actions are up to us (1113b7–8). The argument does not require that we understand (6) as a claim that acting and not acting are manifestations of a single power. For when Aristotle claims that action A and its omission are up to us, he means only that we (rather than necessity, nature or chance) are the cause of A and its omission. The idea that we are also the cause of what we fail to do seems to be a version of Aristotle’s idea that the absence or omission of the cause of E is the cause of the opposite of E; for instance, the absence of a ship captain who would have steered the ship to safety is the cause of shipwreck (Metaph. Δ.2, 1013b11–16; Phys. II.3, 195a11–14). Aristotle’s commitment to 6 suggests that he would also assign the shipwreck to a captain who is on board and could have steered the ship to safety but does not do so. Based on this principle, Aristotle can similarly assign the virtuous failure to act on an unjust command to Socrates, who could have also acted viciously by obeying it. The idea that (6) we are the cause of not acting when acting is up to us – together with (7) – is all that Aristotle needs to reach his desired conclusion: that we (rather than to necessity, nature or chance) are the causes of virtuous (or vicious) omissions, in addition to vicious (or virtuous) actions. We need not additionally attribute to Aristotle the idea that act and omission are exercises of one and the same power.Footnote ⁸

Steward has recently argued that Aristotle evinces a commitment to Powers for Contradictories in the following passage:Footnote ⁹

Steward suggests that this passage focuses on a different distinction than that between non-rational one-way powers and two-way rational powers.Footnote ¹¹ According to Steward, here Aristotle distinguishes between one-way powers and ‘another interesting set of powers which he takes to be manifested more widely than merely in the human case, cases in which what he calls “desire” is operative’ (Steward, Reference Steward2020, p. 352 – italics Steward’s); these latter powers, moreover, are those that ‘leave open two ways in which the agent might proceed: to act or not to act’ (p. 354). Steward bases this suggestion on three considerations. The first is the idea, which is reflected in Steward’s square-bracket insertions into Ross’ translation, that ‘the latter kind’ (τοιαύτασ, a5) refers to powers that are present in both living and non-living things and ‘the former kind’ (ἐκείνασ, a7) to those present only in living things. The ‘former’ powers – which Steward takes to be those present only in living things – leave open whether one acts or not (a11–13). As further evidence for this proposal, Steward writes that choice – the decisive factor in whether ‘the former’ powers get exercised – is ‘a capacity Aristotle standardly permits to some nonhuman animals’ (Steward, Reference Steward2020, p. 352). Hence, these powers, unlike rational powers, are not restricted to human beings. Finally, Steward draws attention to the fact that ‘it is “the animal” (and not only the human being) who does whatever it desires decisively’ (Steward, Reference Steward2020, p. 352).

Steward’s understanding of ‘the former’ and ‘the latter’ is based on the lines that immediately precede the quoted passage, where Aristotle writes, ‘it is necessary that the former are in living beings and the latter are in both [sc. living and non-living beings]’ (1048a4–5). However, if we read back one clause further, it becomes clear that ‘the former’ and ‘the latter’ here refer to rational and non-rational powers respectively: ‘some things [the antecedent of ‘the former’] are capable of producing change on the basis of an account and their powers are rational, others [the antecedent of ‘the latter’] are non-rational and their powers are non-rational’ (1048a2–4). Secondly, Aristotle explicitly restricts choice (προαίρεσισ) to adult human beings (EE II.10 1226b16–22; NE III.2, 1111b8–10). And although desire (ὀρέξισ) is a generic notion that includes both rational and non-rational forms of motivation (EE II.7, 1223b26–27), Aristotle is clear in the lines that immediately follow the above passage that ‘desire or choice’ is a condition for the exercise of rational powers: ‘therefore, it is necessary that whenever anything that is capable on the basis of an account [i.e., any possessor of a rational power, cp. a2–3] desires that for which it has the power and it is in the appropriate condition, it does this’ (a13–15). Finally, although Ross’ translation includes a reference to ‘the animal’ in the translation of a11–13, Aristotle omits the subject of the verbs ‘desire’ and ‘do,’ writing literally ‘whichever of two [sc. contrary changes] [it] desires decisively, [it] will do this’ (a11–12). However, the conclusion that immediately follows this remark (a13–15, quoted in the sentence before last) clarifies that the subject of these verbs is a person that is ‘capable on the basis of an account,’ that is, the possessor of a rational power (cp. 1048a2–3).

Θ.5, therefore, does not introduce a distinction between non-rational powers and the powers of living beings that ‘leave open two ways in which the agent might proceed: to act or not to act’ (Steward, Reference Steward2020, p. 354). Here Aristotle instead argues that the possessor of a rational power can only bring about one or the other of two contrary changes that they are capable of producing if they desire or choose to do so. One might think that, even on this reading, desire and choice are Powers for Contradictories. For these appear to be the cause of whether or not one brings about one or the other contrary change. We cannot, however, infer from this passage that Aristotle is committed to the idea that desire or choice is a single power to φ or not to φ. For, according to this argument, whether one’s desire to φ results in φ-ing depends not only on one’s desire to φ, but also on one’s antecedent possession of a rational power that enables one to φ (or ψ). Desire and choice, according to Θ.5, are thus not single powers to φ (or ψ) or not; they are instead partial causes of φ-ing (or ψ-ing), which require the cooperation of a rational power also acting as a partial cause in order for φ-ing (or ψ-ing) to occur.

There are further passages that one might take to evince Aristotle’s commitment to Powers for Contradictories.Footnote ¹² But none of these passages convincingly shows that the power to act or not to act is distinctive kind of causal power. For Aristotle presents his official study of causal powers, or ‘origins of change in something else’ (Θ.1, 1046a10–11), in Metaphysics Θ.1–5. And according to these chapters, causal powers come in just two kinds: non-rational powers that bring about one change only, and rational powers, which are ‘of contraries’ (Θ.2, 1046b4–7). The absence of Powers for Contradictories from these chapters makes it unlikely that Aristotle recognizes them as a distinctive kind of causal power. Rational powers, on the other hand, are both ‘origins of change in something else’ and ‘of contraries.’ They appear to be more likely Aristotelian candidates for a two-way power: a single power whose exercises are two contrary changes. In the remaining sections of the paper, however, I argue that rational powers are not Powers for Contraries.

4. Rational vs. Non-Rational Powers (Θ.2, 1046b4–15)

In Metaphysics Θ.2, Aristotle distinguishes between rational and non-rational powers as follows (1046b4–15):

According to this passage, rational powers differ from non-rational ones in that the former are ‘of contraries’ whereas the latter are ‘of one thing only’ (b4–7). Aristotle supports this distinction by noting that knowledge is an account (λόγοσ) and that the same account discloses both an object (F⁺) and its privation (F^–) (b7–9). From this fact about knowledge in general it follows that ‘kinds of knowledge of this sort’, i.e., rational powers, are also ‘of contraries’ (b10–11). Throughout this passage, Aristotle repeatedly adds the qualification that knowledge in general (b9–10) and rational powers in particular (a11–12) do not relate to a thing and its privation in the same way (b9–10; 11–12). In b12–15 Aristotle explains this qualification in greater detail: the account of F⁺ accidentally (κατὰ συμβεβήκοσ) discloses F^–, by which he appears to mean that this account discloses F^– indirectly, or by way of negation.Footnote ¹⁴

Scholars typically take the claim that rational powers are ‘of contraries’ to describe the relation between a power and two contrary changes.Footnote ¹⁵ According to this interpretation, Aristotle maintains that a rational power has as its manifestations two distinct and contrary changes: making something else F⁺ and making it F^–. Three considerations support this reading. First, rational powers are ‘origins of change in something else’ (Θ.1, 1046a10–11; Θ.2, 1048a36–b4). Secondly, Aristotle’s example of a non-rational power – ‘the hot thing is of heating’ – clearly has a change as its exercise. This makes it natural to read Aristotle’s example ‘medicine is of health and disease’ analogously. Third, when Aristotle rehearses his distinction between one- and two-way powers in Θ.5 he claims that all non-rational powers are such that ‘a single power is productive (ποιητική) of a single thing, whereas [rational powers] are of opposites’ (1048a8–9). The second clause might simply elide the word ‘productive,’Footnote ¹⁶ in which case this would amount to a direct assertion that rational powers are Powers for Contraries.

Despite the attractions of this reading, I will argue that the pertinent remark instead concerns the relation between knowledge and its objects. The way that Aristotle explains the distinction between one way and two-way powers provides initial support for this suggestion. Rational powers are ‘of contraries’ because knowledge is an account, and the account of F⁺ (e.g., health) also discloses (δηλοῖ, b8; b14), indirectly, F^– (e.g., disease) (b7–9). From this fact about knowledge in general, it follows that ‘kinds of knowledge of this sort’, i.e. rational powers, are also of contraries (b10–11). These remarks suggest that when Aristotle claims that a single rational power is ‘of contraries’ he means only that knowledge of F⁺ discloses both F⁺ and F^–. Thus, by contrast with non-rational powers, where the power relates directly to a single change, the manifestations of a rational power are both cases of knowing or contemplating one or the other of two distinct and opposed objects.Footnote ¹⁷ According to this reading, a rational power is not an immediate origin of change in something else. The production of contrary changes is a more complicated process involving a plurality of powers. Aristotle explains this process in the passage that immediately follows T1 (Θ.2, 1046b15–24). But before discussing that passage (in §7), I first defend attributing this reading to Aristotle (§5) and explain why he endorses this view (§6).

5. ‘Knowledge of Counterparts is the Same’

In Θ.2, Aristotle explains his claim that rational powers are ‘of contraries’ by appealing to the fact that knowledge (ἐπιστήμη) is an account that discloses both a thing and its privation (1046b7–10). This is one instance of an idea to which Aristotle frequently appeals under the slogan ‘Knowledge of Counterparts (ἀντικείμενα) is the Same’ (henceforth: KCS).Footnote ¹⁸ Importantly for present purposes, KCS is a view about the relationship between knowledge and the counterparts that it enables one to contemplate. Because contemplative thought, for Aristotle, does not initiate actions (DA III.10, 432b26–9; PA 1.1, 641b4–8; NE VI.2, 1139a35–6; see §7 below), KCS is not a thesis about the relation between knowledge and action; it is instead the view that the knowledge by which one contemplates one member of a pair of counterparts also enables one to contemplate the other.

I begin with Metaphysics Γ.2, where Aristotle invokes KCS in his discussion of first philosophy, the purely theoretical science of being qua being. Having first argued that essential attributes of being such as sameness, similarity and equality are among the objects of first philosophy (1003b32–1004a2), Aristotle adduces KCS in order to show that this science also studies the counterparts of those attributes:

As in Θ.2, Aristotle here proclaims that because knowledge in general is related to two distinct and opposed objects (a9–10) – a claim that he defends in a long digression (a10–16) – the same is true of the specific kind of knowledge in which he is interested (a17–18; cp. Θ.2, 1046b7–12). Here, however, Aristotle’s target is a purely theoretical form of knowledge: it is a hallmark of the science of first philosophy that it has nothing to do with production (Metaph. A.2, 982b11–28). This form of knowledge relates to its objects by contemplating (θεωρῆσαι, a10) or by being well acquainted with (γνωρίζειν, a20) them.Footnote ²⁰ KCS – the thesis to which Aristotle appeals at the beginning of this passage (a9–10) – thus appears to describe acts of contemplation: what knowledge does is contemplate, or make one well acquainted with, counterparts (ἀντικείμενα).

Aristotle uses the term ‘counterparts’ to describe a generic class of items that includes contraries, contradictories and relatives (Cat. 10, 11b17–19; Metaph. Δ.10, 1018a20–22).Footnote ²¹ The long digression in the middle of T2 (a10–16) explains the distinct ways in which a single knowledge relates to two kinds of counterparts, namely, contradictories and contraries. Aristotle appears to have the latter in mind when he speaks about a thing and its privation (στέρησισ), (a10–12; a15–16). For Aristotle takes one member of a pair of contraries to be the privation of its counterpart within a determinate genus (Metaph. Γ.6, 1011b18–20; Ι.4, 1055b3–9). For instance, disease is the privation of health in a living body. And Aristotle likely adverts to such an underlying subject when he distinguishes privation from denial by claiming that the former additionally involves a genus (a13; cp. Γ.6, 1011b18–20), differentia (a14; cp. Ι.5, 1056a35–b2) or nature (a16; Ι.4, 1055b3–6) of which the privation is asserted. One contrary, then, is a qualified negation of its counterpart: it is the denial of the other within a specific genus. By contrast, when Aristotle speaks about a thing and its denial (ἀπόφασισ) he refers to contradictories: a thing and its unqualified negation. According to this explanatory digression, then, KCS applies to both kinds of counterpart: a single knowledge contemplates not only a thing and its qualified negation, or contrary; knowledge of one member of a contradictory pair also makes one well acquainted with its denial, or unqualified negation.

Although it is relatively clear why the account of one contrary discloses its qualified negation, the proposal that KCS applies to contradictories presents difficulties. Consider for instance the contraries health and disease and suppose that health is the due measure (συμμετρία) of hot and wet elements in a living being (Ps. Alex. In Ar. Metaph. 569, 27–28). Unqualified negation would yield the wholly uninformative account that disease is not a due measure of hot and wet elements in a living being. This is likely the implication that Aristotle seeks to avoid when he writes in Θ.2 that the account of one contrary ‘discloses its contrary by denial i.e. by privative negation (ἀποφορᾷ)’ (13–14). This use of the term, which usually means tax or rent, is otherwise unattested. But Aristotle’s concern with contraries in this passage suggests that ἀποφορά denotes the specific kind of term-negation that he takes to be a linguistic marker for privations (Metaph. Δ.22, 1022b32–3). This kind of negation delivers a more promising account, according to which disease is a lack of due measure (ἀσυμμετρία) of hot and wet elements in a living being (Ps. Alex. In Ar. Metaph. 569, 28–29).

Because unqualified negation does not, in general, illuminate a thing’s negative counterpart, I suggest that Aristotle restricts the class of items whose counterparts are illuminated through unqualified negation to items in the science of being qua being. In T2 Aristotle deploys KCS in connection with three essential attributes of being: sameness, similarly and equality (1004a17–20). For Aristotle, these attributes and their contraries are features of specific categories of being: sameness and difference hold of things whose substance is, or is not, one (Metaph. Δ.15, 1021a11), being similar and dissimilar apply to a different category of beings, namely qualities (Cat. 8, 11a15–19) and quantities alone admit of being equal and unequal (Cat. 6, 6a26–35). Because first philosophy studies each of these categories of being (in so far as they are beings), it studies each of these attributes and their contraries. However, because first philosophy concerns itself with all beings, it will also consider the contradictory opposites of these attributes. For the fact that the unqualified negations of sameness, similarity and equality hold of all items in categories other than the one mentioned in its definition also falls within the purview of first philosophy. Aristotelian metaphysicians know, for instance, that things are equal iff their quantity is one (Metaph. Δ.15, 1021a12). The qualified negation of this definition – negation within the genus in question – discloses inequality: things are unequal iff they are not-one in quantity. But the unqualified negation of this definition, according to which things are not equal iff they are not one-in-quantity discloses an important fact about all beings in non-qualitative categories: ‘equal’ and ‘unequal’ do not apply to them.

I thus propose that the long digression in T2 (a10–16) treats two versions of KCS, which apply to two distinct kinds of counterparts. Special sciences like medicine concern a specific genus of being (Metaph. Ε.1, 1025b4–13) and hence study contraries such as health and its privation. But first philosophy studies all beings just insofar as they are beings, and hence studies not only attributes and their contraries, which apply to specific categories of being, but also concerns itself with the unqualified negations of those attributes, which apply to all beings outside of the genus in question. But most important for present purposes is the fact that both versions of KCS in T2 concern acts of contemplation: if one possesses the account of one contrary or contradictory, then one can also contemplate, or be well acquainted with, the corresponding privation or unqualified negation, respectively.

Aristotle’s applications of KCS to relatives likewise confirm that it is a thesis about acts of contemplation. One clear commitment to this idea occurs in Parts of Animals 1.1, where Aristotle addresses general systematic questions about natural philosophy, the science that studies things with ‘natures’, or internal origins of change and stability (Phys. II.1, 192b8–15). Here Aristotle notes that the natural philosopher must study soul, since soul is an internal origin of change and stability for living beings. There is, however, a puzzle about the extent to which the natural philosopher must discuss soul:

The result that natural science concerns everything is supposed to be absurd because Aristotle wants to deny that the objects of natural science exhaust reality (Metaph. Γ.3, 1005a29–b5). In order to derive this absurd result, Aristotle assumes

and

If we then assume for reductio that

then the absurdity in question follows, so long as we supply the suppressed premise that

What is important for present purposes is that Aristotle does not solve this alleged problem by denying (8) or (9). Instead, in what follows, he denies (10): unlike the nutritive and perceptive parts of the soul, the rational part of the soul is not an origin of motion (κινήσεωσ ἀρχή, 461b4–10; I will return to this idea in §7 below).Footnote ²² Hence, reason does not fall within the purview of the study of nature. And the idea that there is a single science that studies each pair of relatives remains standing.

The idea that a single kind of scientific knowledge contemplates both a psychological activity and its correlative objects is in fact a central organizing principle of Aristotle’s investigation of the soul in De Anima. The following programmatic passage sets out Aristotle’s proposed method for studying each kind of soul (DA II.4, 415a14–22; cf. DA I.1, 402b10–16):

This passage sets the program for Aristotle’s treatment of the various powers of the soul. In order to know (e.g.) what the nutritive power is – in order to grasp its account – one must first understand the activity of nutrition. For nutrition is ‘prior in account’ to the nutritive capacity (DA II.4, 416b3–19). Thus, in order to know what the nutritive capacity is, we must first understand nutrition. However, we must additionally contemplate the ‘counterpart’ of nutrition – the objects of nutrition – before we can grasp the account of nutrition itself (a20–21). This, Aristotle claims, follows for the same reason (a21–2), namely, that the objects of nutrition are ‘prior in account’ to the activity of nutrition. In the following chapters, Aristotle follows this agenda in defining the major kinds of soul: first for the nutritive soul (DA II.4), then for the various perceptual powers that comprise the perceptual soul (II.7–11), and finally for the rational soul (DE III.4).Footnote ²³

As we saw in T2, in T3–4 KCS is clearly a commitment about knowledge and the objects with which it makes one well acquainted: knowledge of the object of a given perceptual activity enables one to grasp the account of the perceptual activity in question. When Aristotle invokes a version of KCS in Θ.2 in order to support his claim that a single rational power is ‘of contraries,’ we should similarly understand the latter as a claim about acts of contemplation: the account that makes one well acquainted with one contrary also enables one to contemplate its counterpart.

6. Counterparts and Priority in Account

Aristotle’s thesis that rational powers are ‘of contraries’ does not reflect a fact about human agency: it is not because of any special fact about humans that knowledge relates to these two distinct objects. Instead, knowledge of one counterpart (C1) discloses the other (C2) because the account of C2 contains the account of C1.

According to T4, there is a hierarchy of priority relations between (a) a perceptual power, (b) a perceptual activity and (c) the corresponding object of perception. Aristotle starts with the idea that b is ‘prior in account’ to a (415a18–20). On Aristotle’s view, x is prior in account to y just in case the account of x does not include the account of y but the account of y does include the account of x (Met. M.2, 1077a36–b4; cp. Z.1, 1028a34–6; Θ.8, 1049b10–17) (see Annas, Reference Annas1976, pp. 146–7, and Beere, Reference Beere2009, pp. 286–7). And activity in general is accordingly prior to power because ‘it is necessary for the account of the one to be there before the account of the other, and the good acquaintance (γνῶσιν) with the one before the good acquaintance with the other’ (Metaph. Θ.8, 1049b16–17). The fact that Aristotle here takes priority in account to entail priority in knowledge reveals that the relevant accounts here are those the possession of which constitutes knowledge: the accounts stating what the relevant power is (see Beere, Reference Beere2009, p. 287). According to Aristotle, then, the account of the exercise of a power is part of what the power in question is. This explains why one must know what the activity is before one can know the power.

Aristotle goes on to claim in T4 that ‘for the same reason’ we must grasp the counterparts of nutrition, perceiving and thinking, that is, nourishing, perceptible and intelligible objects (415a21–2). This is an instance of a general commitment about a certain kind of relative. According to Aristotle ‘relative counterparts (ὅσα… ὡσ τὰ πρόσ τι ἀντίκειται) are things said to be just what they are of their counterparts (ἅπερ ἐστι τῶν ἀντικειμένων) or in some other way in relation to them’ (Cat. 10, 11b24–5; cp. 6a36–7).Footnote ²⁴ Aristotle uses the locution ‘x is just what (ὅπερ) y is’ in order to indicate that y is part of the account of x (Top. 4.1, 120b21–6; Metaph. Γ.4, 1007^a33).Footnote ²⁵ In the case of most relatives, this ‘just what it is’ relation is symmetric: R1 is just what it is in relation to R2 and R2 is just what it is in relation to R1 (e.g., the greater/the lesser, Cat. 7, 6a36–b2). Perception and thought, however, are exceptions that exhibit definitional asymmetry: the activities of perception and thought are ‘called just what they are’ in relation to their counterpart objects, however, perceptible objects and intelligible objects are not just what they are relative to perception and thought (Metaph. Δ.15, 1020b30–2 & 1021a26–9).Footnote ²⁶ According to this distinction, then, the accounts of perceptible objects are part of the accounts of the various kinds of perception, but not vice versa (see Everson, Reference Everson1995). It is thus clear why Aristotle endorses KCS in the case of definitionally asymmetric relatives. It is by knowing what the proper perceptible object (e.g., sound) of a given perception (e.g., hearing) is that one is able to grasp what that sense is: ‘for the essence (οὐσία) of each sense is by nature relative to [its proper object]’ (DA II.6, 418a23–5).

The other two counterparts to which Aristotle applies KCS are also definitionally asymmetric. I suggested above that, within first philosophy, KCS applies not only to contraries like similar and dissimilar, but also to contradictories such as similar and its unqualified negation, not-similar. According to Aristotle, things are similar just in case they are one in quality (Metaph. Δ.15, 1021a11–12). If one grasps the account of similarity, one is also automatically in a position to understand what it is for things to be not-similar. For the latter is the unqualified negation of the former: things are not similar just in case they are not one in quality. In the case of contradictories, then, a thing (e.g., being similar) is clearly prior in account to its unqualified negation (e.g., being not similar): because the ways of failing to be similar are manifold and disunified, one can only provide a uniform account that captures everything that fails to be similar by stating what they all lack.

Aristotle understands contraries in a similar way. As we saw above, one member of a pair of contraries (F^–) is the denial of its counterpart (F⁺) within a determinate genus (Metaph. Γ.2, 1004a12–3; 15–6; Γ.6, 1011b18–20). For instance, disease is the lack of health in a living being. Disease is accordingly posterior in account to health. Health, by contrast, is a unity that admits of a direct determination: according to the consensus among ancient medial writers and philosophers, health is a correct balance or equilibrium of different elements in a body (Hip. De prisc. Med. 19.36–42; De nat. hom. 4.2–3; Pl. Resp. IV. 444d; Tim. 82a).Footnote ²⁷ Given this conception of health, it is reasonable to think that knowledge of health discloses both health and disease. For disease is manifold: there are innumerable ways in which a living being can fall short of the standard of health. The only thing that these deficient bodily conditions all have in common is their failure to be a correct balance of bodily elements.

Aristotle takes the definitional asymmetry exhibited in the case of health and disease to hold for contraries in general. For he takes one member of a contrary pair (F⁺) to be a unity and the other (F^–) a lack of that unity. Thus, while we can directly define F⁺, we can only capture F^– in terms of the account of the unity it lacks. There appears to be a Pythagorean background to this idea about contraries. In Metaphysics A.5, Aristotle mentions a Pythagorean ‘table’ of contrary principles arranged in two columns (τὰσ κατὰ συστοιχίαν λεγομένασ), which includes (986a22–26):

Limit Unlimited

Even Odd

Unity Plurality

Rest Rest

Good Bad

Although Aristotle rejects the Pythagorean view that these contraries are principles, he endorses the ordering of contraries into such ‘positive’ and ‘negative’ counterparts along these lines. In Metaphysics Γ.2, for instance, he avers that ‘among contraries, the one column (συστοιχία) is a privation, and all [contraries] are explained in terms of being and non–being and in terms of unity and plurality’ (1004b27–9). He similarly contends in his treatment of contraries in Metaphysics I.4 that one member of a contrary pair is always a privation, that unity and plurality are the genera of contraries, and that each contrary pair is explained in terms of, or reduced to (εἰσ ταῦτα ἀνάγεται), these genera (1055b26–9).Footnote ²⁸

This background theory about contrariety is the key to understanding Aristotle’s contention in Θ.2 that a rational power relates to the positively existing thing (τοῦ ὑπάρχοντοσ) in its own right (καθ᾽ αὐτό) and to its privation only accidentally (κατὰ συμβεβηκόσ; 1046b9–10; 11–12; 12–13). It is not an arbitrary matter, for Aristotle, which contrary is primary. It is of course reasonable to think, for example, that humans’ primary purpose in developing, say, medicine was to produce health and that the rational power of medicine is therefore primarily concerned with grasping health. But Aristotle’s point is a deeper one.Footnote ²⁹ Among contraries, one is a unity that we can directly disclose by giving an account of it. The other member of the pair, however, is disunified: we can only offer an account of F^– by appropriately negating F⁺.

I have argued that we should understand Aristotle’s claim that a rational power is ‘of contraries’ as the claim that the exercises of a rational power are acts of contemplation: the account of F⁺ makes one well acquainted with both F⁺ and F^– because the latter is posterior in account to the former. But Aristotle nevertheless maintains that these accounts are origins of change in something else (Θ.2, 1046b2–4). In the final section of this paper, I consider Aristotle’s account of the way in which rational powers serve as origins of the changes that their possessors produce.

7. Rational Powers as Origins of Change in Something Else (Θ.2, 1046b15–22)

After having distinguished between rational and non-rational powers (Θ.2, 1046b4–15 = T1), Aristotle explains how possessors of rational powers (henceforth: ‘experts’) bring about contrary changes as follows:

Aristotle initially supports the claim that the expert (‘the one with knowledge’) produces contrary changes (b19–20) by appealing to three premises (b15–17):

Aristotle does not explain why (12) is pertinent.Footnote ³⁰ But he plausibly invokes this principle in order to justify this passage’s implicit contention that the power/change relationship is different in the rational and non-rational cases. As Aristotle’s examples ‘the hot thing’ and ‘the thing good for health’ clarify, possessors of non-rational powers make other things F by being, in some sense, F themselves.Footnote ³¹ The hot thing makes other things hot by being (synonymously) hot and green tea makes humans healthy by being ‘healthy’ in a different but related sense, that is, by promoting the health of that which is called ‘healthy’ in the primary sense of the term, namely bodies (cp. Metaph. Γ.2, 1003a34–b4). However, because experts are capable of bringing about contrary changes, contraries would occur in the same thing if rational powers made other things F⁺ and F^– by themselves being F⁺ and F^–. However, according to 12, medical knowledge cannot be simultaneously ‘healthy’ – i.e., productive of health – and ‘diseased’ – i.e., productive of disease. 13 likely clarifies why this absurd result does not obtain. Knowledge is a power by having an account of F⁺, which discloses both F⁺ and F^–. The origin of the change that (partially) explains why experts can make something else F⁺ or F^– is thus a single unitary thing: the account of F⁺.

Aristotle’s premise that (14) the soul has an origin of motion does not repeat his declaration that (13) the expert possesses a rational power (pace Makin, Reference Makin2006, p. 57). For Aristotle repeatedly cites the expert’s account and the soul’s origin of motion as though they were separate items (b16–17; b20–21). Moreover, Aristotle labelled rational powers origins of change (ἀρχαὶ μεταβλητικαί, Θ.2, 1046b3–4) whereas in 14 he mentions an origin of motion (κινήσεωσ ἀρχήν, b17; b21). The distinction that Aristotle may intend to mark with this terminological difference becomes apparent when we recall that the exercises of rational powers are acts of contemplation: the account of F⁺ enables one to contemplate, or makes one well acquainted with, both F⁺ and F^–; that knowledge is not responsible for making a patient F⁺ or F^–. For those changes result from the expert acting on the patient in which those changes occur. And Aristotle denies that the rational soul – in which rational powers are located (Θ.2, 1046a36–b2) – moves human beings to action:

Although rational powers are clearly, in some sense, origins of change (ἀρχαὶ μεταβλητικαί, Θ.2, 1046b3–4), passages such as this make it clear that they cannot be what initiates the motion (ὁ κινῶν) by which an expert brings about those changes. For, as we have seen in §4, the claim that rational powers are ‘of contraries’ is an instance of KCS, which is itself a claim about how forms of knowledge relate to the counterparts that they contemplate. Because rational powers relate to their objects by disclosing or contemplating them, these forms of knowledge cannot be what is responsible for pursuing some course of action. Indeed, shortly after T6 Aristotle writes ‘we generally see that one who has medical knowledge does not heal, on grounds that something else – not knowledge (ἐπιστήμησ) – is in control (κυρίου) of one’s acting on the basis of that knowledge’ (DA III.9, 433a4–6). Instead, as Aristotle goes on to conclude in DA III.10, ‘what is called desire (ὀρέξισ) is the sort of power in the soul that initiates motion (κινεῖ)’ (433a30–1). I accordingly suggest that the soul’s origin of motion in 14 is the desiderative power. This reading gains further support from Aristotle’s declaration in Θ.5 that desire (ὀρέξισ), or the specific form of desire that is choice (προαίρεσισ), is in control (κύριον) of whether an expert brings about the change that they are capable of producing (1048a10–11).Footnote ³² Two kinds of origin are thus involved in the production of contrary changes: the rational power itself, which makes one well acquainted with the definitions of two contraries and the desiderative power, which is the origin of the motions that bring about the relevant changes.Footnote ³³

In the second half of T5, Aristotle tersely explains how these two origins work together so as to produce change: ‘from the same origin [sc. the expert, cp. b19–20] will produce both (ἄμφω…κινήσει) [sc. kinds of contrary change] by having connected [them] to the same thing’ (b20–2). The origin from which the changes proceed is presumably the expert’s origin of motion, or the desiderative power. Aristotle likely refers to the account when he speaks of ‘the same thing’ to which these changes get connected. And the expert plausibly connects their account to the desired change through deliberative reasoning about means. Indeed, Aristotle frequently uses the verb that I translate ‘connect’ (συνάπτειν) for the process of seeking the middle term(s) that connect(s) two further terms in a deductive chain (An. pr. 1.23, 41a1; 12; 19) (see Ross, Reference Ross1924, p. 243). The fact that Aristotle envisions reasoning about such means as part of the exercise of a rational power makes it attractive to think that choice is the specific form of desire that impels the expert to produce the changes of which their rational power makes them capable.Footnote ³⁴ For choice is a form of rational desire that results from having deliberated about the means of achieving a desired end (1226^b19–20; NE III.3, 1113^a9–12). If this is right, then the proximate origin of the changes that the expert produces is the desire to pursue a specific course of action that they form as the result of having deliberated about how the account of F⁺ relates to the relevant patient in the circumstances in question.Footnote ³⁵

Two distinct origins are at work when experts produce change. Choice is an origin of motion (ἀρχὴ κινήσεωσ, 1046b17) that impels one to produce the desired change, once deliberation has forged a connection between the account and the relevant change though an appropriate actionable means.Footnote ³⁶ And the account of F⁺, although not an origin of motion, nevertheless qualifies as a power, or origin of change (ἀρχὴ μεταβολῆσ, Θ.2, 1046b3–4), in that it constitutes a starting point for the expert’s deliberation about how to make any given patient either F⁺ or F^–.

8. Conclusion

Two-way powers are sometimes thought to be marks of agency: a single power that has two distinct and opposed manifestations, at least one of which is an action. I have argued, on the one hand, that Aristotle does not recognize causal powers whose manifestations are to act or not to act (§2). Two-way powers theorists have rightly noted that choice is a plausible candidate for such a power. But as we have seen, choice is not a single origin of action or omission: for choice plays the subordinate role of determining whether and how a further source of change – a rational power – gets exercised. On the other hand, neither do rational powers have two contrary changes as their exercises (§§3–6). Rational powers are ‘of contraries’ because they are kinds of knowledge and the knowledge that makes one well acquainted with one member of a pair of counterparts also makes one well acquainted with the other (§4). And the fact that an expert is well acquainted with two counterparts is not grounded in a special feature about experts; a priority in account between the relevant counterparts instead explains the fact that knowledge of counterparts is the same (§5). A rational power is thus an origin of change in the sense that the account F⁺ is the starting point for deliberations about how to make something else F⁺ or F^–; but this rational power itself – just as the knower’s choice, on which the occurrence of change also depends – is not a single source of two distinct and opposed changes (§6)Footnote ³⁷

Sigla: