Concocting Teleology in Aristotle’s Meteorology 4 and Generation of Animals

1 Introduction

In Aristotle’s view, the generation of an animal is a realization of two sorts of capacities: active ones provided by the father, who supplies the efficient cause and the form, and passive ones provided by the mother, who supplies the matter on which the efficient cause acts (GA 1.2, 716a4–7; GA 2.4, 740b18–25; GA 1.20, 729a20–33, 729b12–14). The process is teleological: its efficient cause – the nutritive soul or “nature,”^[1] first in the father (GA 1.20, 729a9–10; 1.21, 729a34–b21; 2.1, 735a12–26; cf. 2.3, 736b8–13) and then in the developing animal (2.4, 739b33–740a24; 2.1, 735a12–26) – acts for the sake of something (2.6, 743a37–743b17; cf. 742a16–742b17).^[2] In doing so, Aristotle claims in GA 2.6, this cause acts like a housekeeper:

Each of the other parts comes to be out of the nourishment, those that are most honorable and that share in the most authoritative principle out of the concocted and purest and first nourishment, but the parts that are necessary and for the sake of these out of the worse [nourishment] and the remainders and residues.^[3] For just like a good housekeeper, nature too is accustomed to throw out nothing out of which it is possible to make something useful (χρηστόν). And in housekeeping arrangements, the best of the nourishment that comes to be has been assigned to the free people, but the worse and the residue of this to the slaves, and they give the worst bits, in fact, to the animals that are nourished alongside them. So just as the mind from without does these things for growth, in this way in the things that are coming to be themselves nature organizes flesh and the bodies of the other sense organs out of the purest matter, but bones and sinews and hair out of the residues, and further nails and hoofs and all such things; this is why these get their organization last, once a residue of nature comes to be. (744b11–27)^[4]

This passage distinguishes two groups of parts: the “most honorable” ones, which “share in the most authoritative principle” (which I will call ‘privileged parts’) and those “that are necessary and for the sake of these” (‘less privileged parts’) (744b12–15). It also describes the processes (γίγνεται, 744b11; συνίστησιν, b24; λαμβάνει τὴν σύστασιν, b26) by which their efficient cause – “nature” – produces them.^[5]

What are these processes like? In making the less privileged parts, “nature” exhibits her tendency “to throw out nothing out of which it is possible to make something useful” (744b16–17). In context, this is a significant claim. Earlier in GA 2.6, Aristotle argued that being “useful,” “what the end uses,” and “tool-like […] for various uses” was one way of being for the sake of something (742a22–23, 32, 36; cf. 742b3–6).^[6] He then explained how “nature uses” heat and cold (743a26–27) – unpacking his earlier claim that it “uses” them as “tools” (740b31–32). Using and usefulness, moreover, also feature in the Parts of Animals’ account of the generation of the less privileged parts (what the housekeeper analogy calls “something useful”). Horns, for instance, come to be when nature “uses” residue for a purpose (PA 3.2, 663b31–33).^[7] Their generation – a case of “mak[ing] something useful,” according to GA 2.6 – is thus a case of using matter, as is that of other such parts (GA 2.4, 738b1, cf. PA 2.15, 658b24). This all tends to suggest that in GA 2.6 and in discussions of the less privileged parts elsewhere, using, usefulness, and making useful are important causal notions – ones that can apply to the causal processes by which parts come to be.

What, then, are such processes like? What happens when nature exhibits her tendency “to throw out nothing out of which it is possible to make something useful”?^[8] Leunissen has famously argued that such processes differ from those that produce the more privileged parts insofar as they exhibit different kinds of end-directed efficient causation (what she calls “causal patterns”): “primary” and “secondary” teleology.^[9] These patterns differ in where they locate the “causal connection”: in secondary teleology, it is “between the available material potentials” – in my terminology, capacities – “and the function or use to which the materials with those potentials are put”; in primary teleology, it is “between a potential for form and the realized potential, which constitutes the final cause.”^[10] Leunissen develops this view by expanding on three differences mentioned in the housekeeper analogy: in (1) the role of the product in an animal life^[11] (e. g., “most honorable” or “necessary”; 744b12, 14), (2) the matter it is made from^[12] (“purest” nourishment versus “residues”; 744b13–16), and (3) whether the efficient cause is realizing a “preexisting” capacity “for form”^[13] or using and co-opting available materials (“mak[ing] something useful”; 744b17).^[14] As she sometimes puts it, parts produced by secondary teleology – unlike those produced by primary teleology – “come to be of material necessity, but are present for a function.”^[15]

This way of putting it suggests that Aristotle’s view is not merely that nature acts for the sake of something in two distinctive scenarios, differing in the materials they make available, how far generation has already progressed, and the function of the part to be produced (cf. Leunissen’s (1) and (2)). Clearly, Aristotle does think – to adapt an example from Henry – that nature, like a builder, might sometimes find that she “discovers that she has leftover materials that were not specified in the original blueprints.”^[16] But there is also a stronger claim in view: that once nature is in a distinctive scenario like this one, it acts in a different way – exhibiting a different sort of end-directed agency, subject to different sorts of (end-directed efficient) causal explanation (Leunissen’s different “patterns”). This stronger claim, moreover, is a very natural reading of the housekeeper analogy. In describing that image, after all, Aristotle is extending his discussion of the causal work that using and usefulness can do.^[17]

How, then, does nature act when it “make[s] something useful”? This turns out to be a difficult question to answer. The problem is that if we are inquiring not into the scenario itself but into what nature does in it, it is hard to say what might be distinctive about its actions. As Henry puts it, when the builder “uses that extra lumber to build window boxes, we would still say,” contra Leunissen’s (3), “that those window boxes came to be for the sake of holding flowers.”^[18] In this scenario, it can seem, what nature does is much the same as it did in earlier scenarios: it makes something for the sake of an end – previously, the walls and roof; now, window boxes.

What is needed is thus a clearer account of what happens in the second scenario: of what nature does in “using” the matter to make something “useful.”^[19] This paper takes some initial steps towards that account.^[20] I defend two main claims. The first is that Meteor. 4.2 offers an underappreciated source of evidence. In its account of “concoction” – a key process type in animal generation – it distinguishes between cases where the “end” is “nature,” understood as “form and substantial being,” and cases where the “end” is “useful” and “in regard to a certain underlying character” (379b25–29). My first claim is thus that this distinction lies behind the housekeeper analogy: that it spells out one way – there may be others, perhaps connected to this one, which I hope to explore elsewhere – in which the processes distinguished by that analogy differ.

My second claim concerns the nature of Meteor. 4.2’s two sorts of process and the ends they realize. Both ends are, of course, the ends of their respective types of concoction: they explain, in the way Aristotelian ends do, how processes of those kinds (non-accidentally) unfold (cf. Phys. 2.8, 199a8–12) and in that way make them the processes they are.^[21] Still, they differ in that some essential (process-defining) features of the second end – but not the first – are supplied by what the capacities of the “underlying” patient (on which the agent acts) are for. (My interpretation thus makes a distinction between what a process and the capacities realized in it are for.)

Suppose, for instance, that a cook roasts a cauliflower, exercising her capacity to make a delicious meal.^[22] (This is what her capacity aims at.) When she does so, that roasting process happens for the sake of a meal that is not just delicious in any old way but specifically in the way cauliflower is delicious. (This process will fail not only if it produces something that is not delicious, but also something that doesn’t taste like a cauliflower.) The reason is that that process’s end is specified in part with respect to what the patient’s (passive) capacities are for (e. g., a certain sort of flavor).^[23] It is also, of course, specified with respect to what the agent’s (active) capacities are for: because she is exercising a capacity for making a delicious meal, the process’s end is not just a cauliflower flavor, but a delicious one. What lies behind this complexity in the process’s end, moreover, is that the agent’s capacity is of an interesting sort: a good cook knows there is no one way food must taste to be delicious – and so she lets the flavors of her ingredients shine through. This is the sort of end – open to the flavors of the ingredients – that defines her agency and her characteristic capacity. (Her capacity is thus what I have elsewhere called a ‘flexible’” capacity, and – like others I have considered elsewhere – one that is exercised as it is in part in virtue of the circumstances it is exercised in.^[24]) This is why the resulting process has the complex end it does: because the end of the agent’s active capacities does not specify the specific flavor, something else can.

Aristotle’s point in Meteor. 4.2 is that this differs from another sort of case, where the “end” is a “form and substantial being.” There, I will argue, the process’s end is fully specified by the end the agent’s capacity is for; the patient (as patients do) just makes it possible for it to come to be. The difference between the two cases, then, is in what specifies the process’s end – and thereby in what the agent and patient do in bringing it about. And this, I argue, is one – the one I focus on here, as a first step towards a full account – of the points the housekeeper analogy draws our attention to.

Below, I set out the case for this view. Section 2 introduces the Meteor. 4.2 passage and sketches its relationship to GA 2.6’s housekeeper analogy. Sections 3 and 4 present the evidence for my account of the core distinction. Section 5 returns to the analogy. Section 6 considers one way of making my view precise: using the machinery of determinables, determinants, and determinates.

2 Meteorology 4.2 and the Good Housekeeper

Meteor. 4 occupies an important place in Aristotle’s scientific program, setting out many of the causal notions he puts to work in his biology.^[25] It begins by introducing certain active and passive causes: agents and patients. The “active” ones are the hot and cold; the “passive” are the dry and moist (4.1, 378b12–13). The work’s first two tasks are to “grasp the workings by which the active ones are at work, and the forms of the passive ones” (378b26–28). And it soon becomes clear that the active and passive causes are capacities (δυνάμεις, 378b32–34) or, perhaps, what has them (4.2, 379b11–12).^[26]

Capacities, of course, are key notions in Aristotle’s account of efficient causation.^[27] According to Physics 3.1–3, a causal process (κίνησις) – what an efficient cause is responsible for (2.3, 195a8, 11, 23; 2.6, 198a3) – is the realization of two of them: the passive capacity of the patient (say, bronze, or a learner) and the active capacity of the agent (a sculptor, a teacher) (3.3, 202b26–27).^[28] It is the production of a statue, or the process of teaching and learning. Its efficient cause is the active capacity or the agent that has it: the building craft (a capacity), or the builder (2.3, 194b19–32, 195b21–25). Sometimes, this cause aims at (is or has a capacity for bringing about) an end (e. g., a beautiful statue), and the process is end-directed: it happens for the sake of something, non-accidentally, insofar as it is a realization of that capacity (cf. 2.8, 199a8–12).

This picture of efficient causation is in the background of Meteor. 4.^[29] In inquiring into “the workings (ἐργασίας) by which the active ones are at work (ἐργάζονται)” (4.1, 378b27–28), it is inquiring into the realization (“workings”) of the active capacities (cf. 4.5, 382a28–29, 31). At a general level, Aristotle’s proposal is that these “workings,” when successful, involve “mastering” the “passive” capacities (4.1, 378b34) and “matter” (ὕλη) (379a1, 378b33; 4.2, 379b33). 4.1–3 articulate this proposal, with 4.1 offering a more “general” (as Lennox puts it) account of mastering in natural change and generation (and its failure in cases of destruction) and 4.2–3 turning in detail to the “workings” (cf. 4.1, 378b27–28) of the active causes.^[30] 4.2 thus begins by promising to set out the “resulting forms, as many as the capacities mentioned work up (ἐργάζονται) out of underlying things already organized by nature (ἐξ ὑποκειμένων τῶν φύσει συνεστώτων ἤδη)” (379b10–12).^[31] These turn out to be kinds of heating: “concoction” (πέψις) and “inconcoction” (ἀπεψία).^[32] With this work in place, 4.3 turns to their subtypes. The distinction I am interested in occurs in 4.2:

And the end for some things (τοῖς μὲν) is the nature, and nature is what we mean in the sense of form and substantial being;^[33] for others (τοῖς δὲ), the end of the concoction is in regard to (εἰς) a certain underlying character (ὑποκειμένην τινὰ μορφὴν), when what is moist comes to be such-and-such and so-much when it is either roasted or boiled or putrefied or heated in some other way; for then it is useful and we say it has been concocted, like new wine and the things that are organized in growths, when pus comes to be, and a tear, when rheum comes to be; and similarly also in the other cases. (379b25–32)^[34]

This is a distinction between two sorts of end (τέλος) a concoction process can realize, in two different cases (τοῖς μὲν, τοῖς δὲ). I will refer to the process types that realize them as concoction-1 and -2.

My first main claim is that this distinction lies behind GA 2.6’s housekeeper analogy: Aristotle is putting its ideas to work to address the issues he is interested in there. (I am not claiming that the overall point of the two passages is the same; each is motivated by questions arising in its own context.) Three preliminary reasons favor this suggestion. The first and most general is the fact that Meteor. 4.2’s distinction is between kinds of concoction – perhaps the key process in animal generation.^[35] It would therefore be unsurprising to find that a distinction among processes involved in animal generation picks up a distinction between two sorts of concoction.

The second – more specific – reason has to do with the analogy’s privileged parts: the “most honorable” parts that “share in the most authoritative principle” (GA 2.6, 744b12–13). In context, this is the “principle” that belongs to the “most authoritative part” of “the whole and the end”: the one that “has the principle and end of the entire nature” (742a33–35, cf. b17; 742b1). Aristotle’s view is that this part (the heart: 2.4, 740a3–4; 2.5, 741b15) and the other “most honorable” parts together form a “whole” that is the animal’s “end.” They are “parts of the end” (2.6, 742b17), and the other parts are for their sake (742a36, 742b3–8).^[36] What I want to argue is that this privileged collection of parts is closely connected with what Meteor. 4.2 calls a thing’s “nature […] in the sense of form and substantial being” (379b25–26).^[37]

We can start by observing that one part is for the sake of another when its defining action is for the sake of – and “prior to” – the other’s (PA 1.5, 645b29–32). The fact that the privileged parts are the end of the less privileged parts (GA 2.6, 744b14–15) thus provides some evidence that they perform an action that is the end of the less privileged parts’ actions: this would be why the less privileged parts are for their sake. What action might be this be? Insofar as the privileged parts are “the whole and the end” (742a35, cf. 742b17), it is presumably the “complete action” that defines the “entire body” (PA 1.5, 645b16–17).^[38] Importantly, such an action defines the body in much the way that sawing defines a saw (645b14–19). And in Aristotle’s view, a thing is defined with respect to its substantial being (Metaph. Z.4, 1030b4–6; Z.13, 1039a19–20). This gives us another reason to associate these parts with Meteor. 4.2’s “nature […] in the sense of form and substantial being” (379b25–26): they are the parts required for being able to perform the actions that define a thing’s nature.^[39]

The third reason picks up the less privileged parts: those in whose production nature exhibits her tendency “to throw out nothing out of which it is possible to make something useful (χρηστόν)” (GA 2.6, 744b16–17). When nature does this, the resulting products are in fact notable for their usefulness. They are, after all, parts “for the sake of” the privileged parts (744b14, cf. 742a22–23) – where these just are the “whole and the end” (742a35) that “uses” (742a32) them and to which they are “useful” (χρησίμου) (742a32). This is as in concoction-2, where “the end […] is in regard to a certain underlying character, when what is moist comes to be […] useful” (χρήσιμόν) (Meteor. 4.2, 379b26–29).

There are three preliminary reasons, then, to think the distinction between concoction-1 and -2 lies behind the housekeeper analogy: that concoction-1 produces one group of parts and concoction-2 the other. (As my argument progresses, I will offer two more.)

What, then, are concoction-1 and -2 like? In Meteor. 4.2 Aristotle focuses on their characteristic “ends” (τέλη, 379b25, 27). In doing so, he fills out and further develops his earlier claim that “concoction is a perfection (τελείωσις) by the natural and proper heat out of the opposed passive (ἐκ τῶν ἀντικειμένων παθητικῶν) [capacities^[40]],” where “these are the proper matter (ὕλη) for each thing” (379b18–20; cf. 379b20–21). The τέλη in question – both of them (379b25, 27) – are thus τέλη of τελειώσεις.^[41] How are we to understand Aristotle’s distinction among such τέλη?

There is, as it happens, a tricky question about just what sorts of ends Meteor. 4 is concerned with.^[42] For my purposes, however, a fairly minimal notion will be enough.^[43] It arises from Metaph. Θ.8’s remark that “the work (ἔργον) is an end (τέλος), and the activity (ἐνέργεια) is the work, which is why the name ‘activity’ is said in accord with the work and extends to the actuality (ἐντελέχειαν)” (1050a21–23).^[44] In context (as Charles has observed), Aristotle is appealing to the fact that both the results in cases of acting (the “work,” e. g., playing the flute) and actualities beyond the action (e. g., houses) function as ends. He then suggests that the same name can apply to both (cf. 1050a23–28). This points, Charles argues, to a way of thinking about actuality: in an end-like way, as the “success condition for the realisation” of a given capacity. The capacity to be a house, for instance, is successfully realized when a house has come to be. A house is the capacity’s end: it is the standard by which we can say that the capacity has been successfully realized. It is for this reason, Charles is suggesting, that the house counts as an actuality.^[45] An ἐντελέχεια is, after all, (as he puts it) “what has the goal within it.”^[46] This account is also a good fit with Aristotle’s claim (also emphasized by Charles) that in defining a capacity, we do so with reference to the “end” (ὁρίζεσθαι πρὸς τὸ τέλος) – which is the “maximum” that results when all goes well (e. g., walking a hundred stades; De Caelo 1.11, 281a1–27).^[47] Thus Aristotelian ends are distinguished from other “last” things in being “best” (Phys. 2.2, 194a32–3).

My proposal, then, is that we can start by treating the “ends” distinguished in Meteor. 4.2 as things that set the standard for the success of the processes they characterize. They will define those processes and so explain, in the way success conditions do, how those processes go: the way they must if they are to succeed (cf. Phys. 2.8, 199a8–12, PA 1.1, 639b11–19, 640a15–19; 2.1, 646a24–28).

3 The End of Concoction-2

How, then, do those standards differ? And when does a process have one rather than the other? We may start by noticing that in context, Aristotle has several resources to work with. He is developing an account of certain active and passive capacities, and of how they are realized in causal processes. I will argue that these are precisely the resources he uses to unpack the difference between his two standards. The success conditions of concoction-1 and -2 differ in whether they are set only by the active capacity (more precisely, by what it is a capacity for realizing) or whether they are also set by the passive one (more precisely, by what it is a capacity for realizing).

I will start with concoction-2. Its “end […] is in regard to (εἰς) a certain underlying character, when what is moist comes to be such-and-such and so-much when it is either roasted or boiled or putrefied or heated in some other way; for then it is useful and we say it has been concocted” (379b26–30). This remark points to an interestingly complex end (success condition). Aristotle’s point is not just that when moist stuff acquires some qualitative or quantitative “character” (“such-and-such and so-much”), the end of the process of producing it is or is specified with regard to (εἰς) to that character. In acquiring this character, the stuff must also become^[48] (γένηται, τότε, πεπέφθαι) useful – a fact that helps explain why (γὰρ, 379b29) the process’s success condition (its “end”) is specified in the way it is, and so also why it is a case of concoction (and, indeed, concoction-2) specifically.^[49] This is a complex success condition, appealing to an underlying character and to usefulness.

The following chapter, 4.3, reinforces this complexity. It revisits one of 4.2’s examples of concoction-2, the concoction of new wine (γλεῦκος, 379b29–30), and adds a new example:

We also say that moist things are boiled, for instance, milk and new wine, when the flavor in the moist changes into a certain form (εἰς εἶδός τι) by the agency of the surrounding external fire, as it is heating […] (But the end is not the same for all things, neither things boiled nor things concocted, but for some it is with respect to (πρὸς) eating, for others towards drinking, and for others towards another use (χρείαν), since we also say that we boil drugs.) (380b31–381a4)

This passage revisits (and confirms) both aspects of the complex success condition. Here, the “such-and-such” or “underlying character” is a form of flavor: “a certain form” into which “the flavor in the moist changes” (380b32–33).^[50] But, once again, this is not quite enough. The parenthetical remark makes the key move: “the end is not the same for all things” (381a1). Why might this be? Aristotle’s idea seems to be – as it was in 4.2 – that just pointing to (εἰς, 379b26^[51]) the underlying character does not yield a full specification of the process’s “end.” We also need to point to (πρὸς, 381a2, 3) a “use.” This “use” likewise looks back to Meteor. 4.2’s talk of what is useful: for such things to be made useful is for them to be made to contribute to a further end: a use.^[52] Aristotle’s point, in both passages, is thus that to understand a case of concoction-2 we need to identify that use: in the boiling of wine, the end is not just a flavor, but one useful for medicine, or in other cases, drinking (cf. PA 2.15, 658b24).^[53] Thus “the end is not the same for all things” (381a1) – because the use is not the same.

Aristotle’s view thus appears to be that the end of concoction-2 – its success condition – must be specified with respect to (εἰς, πρὸς) both the “underlying character” and also the “use” for which that character, when it comes to be, is “useful.” This is a complex success condition, specified with reference to two things. What defines concoction-2 – what explains, in the way an end does, how it goes if it is to succeed (cf., again, Phys. 2.8, 199a8–12) – is complex.

How exactly are we to understand this (complex) end? I will begin with Aristotle’s claim that the character with respect to which it is specified is “underlying” (ὑποκειμένην). This is an important clue. Phys. 1.7 famously suggests that what “underlies” (ὑποκεῖσθαι, 190a15; cf. 34–36, 190b20) is the patient of a change: the human who becomes musical (190a16–18, 190b14), or the bronze that becomes a statue (190b16–17). Meteor. 4, importantly, is (as we have seen) concerned with such patients, and it describes them as “underlying” – and as “matter” (and so as the sort of underlying patient that matter is).^[54] 4.1’s general account of natural change and generation, for instance, claims that it takes place “out of (ἐκ) the underlying matter (ὑποκειμένης ὕλης) for each nature” (378b32–34) – where “these are the capacities called passive” (378b34). And this claim puts Aristotle in a position to claim that “the hot and the cold generate by mastering the matter (τῆς ὕλης)” (379a1). Similarly, at the start of 4.2, Aristotle begins to inquire into the “forms” produced by the active capacities, “out of (ἐξ) the underlying (ὑποκειμένων) things already organized by nature” (379b10–12) – where these underlying things (the things “out of” (ἐκ) which) turn out to be “the opposed passive [capacities]” that are “the proper matter for each thing” (ἡ οἰκεία ἑκαστῳ ὕλη, 379b19–21). What counts as “underlying” in 4.2’s discussion of concoction-2 should therefore be what these chapters call “matter” (cf. 379b33, 380a9; 4.5, 382a29) and treat as “passive” patients or capacities.^[55]

Is an underlying character, then, a character closely associated – in a way still in need of explanation – with such patients? Aristotle’s later remarks suggest it is.^[56] Once again, the case of new wine (γλεῦκος) provides the important clue. Aristotle returns to that example not only in 4.3 (see above), but also in 4.7, and 4.10. Those three discussions, I will now argue, begin to spell out the patient’s relationship to the “underlying character.” They do so by (1) identifying the sub-type of concoction-2 involved (boiling), (2) explaining what the patient of a process of that sub-type is like (very earthy and so boilable), thereby (3) pointing to how the patient contributes, in virtue of its earthiness, to the successful realization of concoction-2’s end.

I begin with (1). Meteor. 4.3 makes clear (in the passage quoted above: 380b31–381a4), that this is a case where γλεῦκος is the patient of a process of boiling: one of 4.3’s three sub-types of concoction. 4.2, of course, also associated boiling with concoction-2 (379b27–29). This offers us an important clue: we can inquire into why the process of boiling γλεῦκος produces an underlying character – and so into what that character is like. After 4.3, 381b31–381a4’s parenthetical remark, Aristotle adds: “the result is that all those things” – including γλεῦκος – “can be boiled that can come to be thicker or smaller or heavier, or some bits of them such and others the opposite, on account of the fact that in being divided some parts are thickened and others thinned, just as milk into whey and rennet” (381a4–8). 4.10 explains why this is so:

And as many [moist things] as are thickened by heat are combinations (and someone might be at a loss regarding wine, among the moist things; for this both could evaporate, and it is thickened, as new wine (ὁ νέος) is; and the cause is that wine is not said “in one form,” and that different [wines] are disposed differently; for new wine is composed more of earth than old wine is; for this reason it is also thickened by the hot most of all and is solidified less by the cold; for it has much heat and is of earth^[57] […]; then if every wine has sediment, it is composed of each one – earth or water – in accord with the amount it has of this); and as many as are thickened by cold are composed of earth; and as many as are [thickened] by both are combinations of more [than one], for instance olive oil and honey and sweet wine. (388a32–b10)

The explanation is that bodies – like γλεῦκος or “new wine” – can become “thicker or smaller or heaver” when they contain a lot of earth. This makes them susceptible to thickening, and this (recall 4.3, 381a4–8) makes them susceptible to boiling. Similarly 4.7: “as many things as are mixed of water and earth should be spoken of in accord with the amount of each; for a kind of wine is both solidified and boiled, such as γλεῦκος” (384a3–5). Γλεῦκος, it seems, differs from other wines in containing lots of earth (cf. 384a8).^[58]

So far, the lesson is that 4.3, 4.7, and 4.10 together indicate (and this is claim (2)) that γλεῦκος is very earthy, which makes it an appropriate patient for boiling. But this is not the only reason its earthiness matters. When γλεῦκος is boiled, its flavor “changes into a certain form” (4.3, 380b32–33). It does so, I will now argue, because γλεῦκος’ earthiness not only makes it boilable, but also (due to the earth’s qualities) makes it boilable in a specific way: they fix what sort of “form” (380b32–33) it acquires when (successfully) boiled. This is claim (3). The evidence comes from De Sensu 4:

[W]e see flavors changing by the agency of the hot when fruits are picked^[59] and treated with fire, by reason of the fact that they [the flavors] are coming to be of such a sort not by pulling it out of the water, but by changing in the fruit itself, and [we see] them by losing their moisture and lying there coming to be harsh and bitter and of all sorts from [being] sweet, on account of the time, and by being boiled changing into practically all sorts of flavors. (441a11–17)

This passage describes a sort of flavor change caused by boiling (among other things). Aristotle’s explanation of it appeals to the earthy quality of what is heated. His view is that flavors change when heat moves moisture through dry, earthy stuff, helping (cf. 441a29) make the moisture thicken and take on the quality (ποιόν τι) of that stuff – bitter where the earth is ashy, salty where it is salty, etc. (441a20–b20). Thus “this is flavor: the affection that comes to be by the agency of the aforementioned dry in the moist” (441b19–20). This also happens when fruits are heated (441b7–8) – including when they are “boiled”: when they change “into practically every sort of flavor” (441a11–17).

This allows us to fill out Meteor. 4.2–3’s picture. Insofar as γλεῦκος acquires a new flavor when it is boiled, its flavor change (cf. 441a11–17) is like what happens when “nature […] by filtering and moving the moist through the dry and earthy by means of the hot makes it of a certain quality” (441b17–19).^[60] Here, then, is the deeper reason 4.3, 4.7, and 4.10 emphasize γλεῦκος’ earthiness: this not only makes it susceptible to boiling, but also explains what the process of boiling is like.^[61] In boiling, the heat helps bring out the flavor of the earth in it – at least when the process is successful: we are not considering a case of burning the wine, but of bringing out its flavor.^[62]

At a higher level of generality, the point is not just about earthiness, but about the earthy thing: the matter that is the patient the efficient cause acts on. Γλεῦκος is a patient: it is what is boiled – in virtue of its earthy, watery nature.^[63] That the patient should play such a role is, of course, what Phys. 1.7 led us to expect.^[64] The character that is part of the success condition is “underlying” in that it is a feature of the patient: more precisely, a feature the patient (4.2’s “what is moist,” the “matter”) will acquire or retain when heated. It is what the patient, insofar as it is the earthy, watery matter it is, has a capacity (cf. 4.1, 378b26–28, 378b32–34) to acquire or to retain^[65]: the success condition for its passive capacity.

This result also helps show that claim (3) is not an isolated observation about an isolated example. The “form” “into” which γλεῦκος’ flavor changes is just one example of the “underlying character” with regard to which concoction-2’s complex “end” – its success condition – is partly specified. I am thus arguing more generally that concoction-2’s success condition is specified with respect to the success conditions of the patient’s – the matter’s – passive capacity. In cases like these, the patient does not just make it possible to achieve some end (as patients do), but also helps set the process’s success condition. If I boil wine to make a sauce, and the sauce tastes like something other than that wine, that boiling process has failed. That process has the end it does in part because of the passive capacities it realizes, and it brings out a feature of the patient because this is what its success in part consists in. (This, I am suggesting, is what saying that concoction-2’s success condition is ‘specified’ in part with respect to the success conditions of the patient’s passive capacity comes to.)

This, then, is how I propose to understand the notion of an underlying character.^[66] But that character, of course, is only half the story. The end of concoction-2 is complex: it is also specified with respect to a use for which that character is useful. My next question is thus: what are these uses like? Meteor. 4.3 offers some examples: the uses crafts and craftspeople make things useful for. Doctors boil drugs and make them useful for medicine (381a3–4); cooks boil things and make them useful for eating or drinking (381a2–3).^[67] Phys. 2.2 makes a similar point also dealing with “craft”: that “we too are in a way an end” (194a33–35). These examples provide a clue: the relevant uses are uses aimed at by agents and their active capacities – here, craftspeople and crafts.^[68] Importantly, however, not all such agents and capacities will be craftspeople and crafts; there are also natural cases of concoction-2 (Meteor. 4.3, 381a10–12), natural uses (GA 2.6, 742a26), and useful natural parts (742a32, 744b17).^[69] There are also natural agents, including the nature that acts like a good housekeeper (744b16–17) – for what “uses” useful animal parts for certain “uses” is not just an end, but also an efficient cause (742a32–b3). And when nature “uses” the matter to make certain parts, it does so with a view to an end: a “use” (e. g., PA 658b24).^[70]

The suggestion, then, is that for a concoction-2 process to succeed, it needs to realize both a character specified with respect to the patient (more precisely, with respect to the “form” it can acquire, etc.), and a usefulness specified with respect to the agent (more precisely, with respect to the use at which it aims and for which the product is useful – where this just is the active capacity’s success condition). Both the character and usefulness thus matter to its success: when a doctor boils wine, the process’s end is not just an underlying character that just so happens to be brought about in circumstances where their results are useful. Instead, it is a process that happens non-accidentally for the sake of a medically useful particular flavor.^[71] It is, after all, a standard case of craft production, which happens as it does for the sake of its useful result, and not by accident (cf. Phys. 2.8, 199a17–20).^[72]

That the complexity of concoction-2’s success condition takes this form is, in a way, not a surprise. Meteor. 4 is concerned with agents (and their active capacities) and patients (and their passive capacities) and how they are realized together. In offering his sketch of concoction-2, Aristotle is just outlining a process type whose success condition is set by the success conditions of both capacities: because these capacities are exercised, this is the success condition the process will have.

4 The Agents of Concoction

There is an implied contrast here to concoction-1: if concoction-2’s success condition is set by the success conditions of both the active and passive capacities realized, concoction-1’s should be different. Is it? In this section, I argue concoction-1’s success condition is set just by the success condition of its agent’s active capacity. This is not, of course, to deny that the patient and its passive capacity make it possible to achieve the process’s success condition – to contribute, in the way patients always do, to the realization of that end. It is just to say that in concoction-1, the process’s end is set only by what the agent’s capacity aims at: it has the end it does just in virtue of exercising that active capacity.

My argument requires a preliminary step. Meteor. 4.2 associates three heating types with concoction-2: roasting, boiling, and putrefying (379b28–29). This list includes only two of the three sub-types of concoction Aristotle will go on to distinguish in 4.3 (roasting and boiling) – and excludes another (ripening). Why might this be? The answer, I will argue, is that ripening is a paradigmatic case of concoction-1, not -2.^[73]

Now, according to 4.3, ripening is “the concoction of the nourishment in fruit casings” (380a11–12). Its efficient cause is a thing’s nature (380a25–26; cf. 4.2, 379b18–20): its “natural heat and cold” (380a22). It “is perfect at the time when the seeds in the fruit casing are able to bring to perfection another like it” (380a13–15). Ripening is thus a process in which a plant’s nature perfects the seeds, by concocting the nourishment in the fruit (cf. DA 2.4, 416b2–23). In doing so, it is making something like itself: making the seeds be of a sort that will eventually develop into another mature plant with a nature of that very same sort (Meteor. 4.3, 380a12–15). It does all this by integrating the “undetermined moisture” (380a29) with itself: it “pulls some things into itself^[74] […] and others it tosses out” (380a25–26).^[75] All this is typical of reproductive activity, including seed production.^[76] In reproduction, an organism “remains not itself but like itself, not one in number but one in form” (DA 2.4, 415b6–7). This is why reproduction and nutrition are activities of the same capacity (415a25–26, 416a19–20): they aim at the same end. This end, importantly, is a thing’s “substantial being”: for while nutrition^[77] “preserves the substantial being” (416b14),^[78] in reproduction that same capacity is “productive of generation, not of the thing nourished, but like the thing nourished” (416b15–16) – in other words, the same sort of substantial being, just in another organism.

The end of reproduction generally – and so seed production specifically – is thus “form” (415b6–7) and “substantial being” (416b14). But this is just to say they have the end typical of concoction-1: “nature […] in the sense of form and substantial being” (Meteor. 4.2, 379b25–26). The upshot is that ripening (seed production) is a paradigmatic case of concoction-1 – which, in turn, explains why it is absent from the list of heating types 4.2 associates with concoction-2. The upshot of the preliminary step is thus that ripening offers a case study of concoction-1 (much as boiling new wine did for concoction-2).

In the remainder of this section, I build on this upshot, sketching its lessons for our understanding of the active capacities exercised in concoction-1 and their role in setting its success condition: a thing’s “form and substantial being.” The structure of my argument is as follows. First, I argue that ripening (which I will call R) is a case of “purification” – and that this is what makes it a paradigmatic case of concoction-1. It is not alone in this respect: the process in which the male acts on the female principle in generation (or G) is also a paradigmatic case of purification – and, for this reason, a case of concoction-1. Second, I note that Aristotle contrasts G, in some respects, with the process in which fig juice and rennet act on milk (or F). I argue that he thinks they differ in the extent to which the agent’s active capacity aims at and thereby causes successful products to have certain specific qualitative features. (Their products have those features because having them is required for being the result of a successful exercise of that capacity.) Now, F is a paradigmatic case of concoction-2 – and so in distinguishing it from G, Aristotle is contrasting paradigmatic cases of concoction-1 and concoction-2. Moreover, he is doing so with regard to the activity capacity’s end: in concoction-2, if the product of a successful process characteristically has certain qualitative and quantitative features, it is not because they are required for being the result of a successful exercise of that active capacity. This difference, I will argue, captures key ideas in Meteor. 4.2’s distinction between the success conditions for the processes of concoction-1 and -2.

The first step is thus to notice that R is a case of purification.^[79] The evidence comes from GA 1.20, where Aristotle compares G to R:

[T]he menses are seed that is not pure but in need of work, just as in the generation of fruits, when it is not yet sifted,^[80] there is nourishment in it, but it is in need of work as regards purification. This is why the former, when it is mixed with semen, generates, and the latter, when it is mixed with pure nourishment, nourishes. (728a26–30)

In both G and R, a purer and a less pure version of the same stuff come into contact (cf. GA 1.19, 726b30–727a2; 1.20, 728a17–20; 2.4, 739b25–26); the former works on – purifies – the latter. In R (“the case of fruit”), “sift[ing]” occurs; as the Problems puts it, “foreign” (ἀλλότριον) moisture is eliminated and the fruit takes on a “proper” (οἰκείαν) smell (20.30, 926a27–29). Similarly, “when the σπέρμα from the male […] arrives, it organizes the purest bit of the residue – for the majority of it is useless” (GA 2.4, 739a6–8). GA 4.1 adds:

[J]ust as in the work to do with fruits the small useful bit is separated off from the first nourishment – from much of it – and at last the ultimate bit is nothing compared to the initial amount, in this way again also the parts, taking [nourishment] up in turn in the body, in the work the final bit comes to be altogether small, from the whole of the nourishment. (765b28–34)

These remarks echo Meteor. 4.3’s claim that in R, “nature pulls some things into itself, and others it tosses out” (380a25–26) – where, we have seen, this just is how nature goes about making the nourishment like it and thereby realizing the fruit’s “nature […] in the sense of form and substantial being” (4.2, 379b25–26). This particular process of pulling in and tossing out, in other words, is what makes R a case of concoction-1. That pulling and tossing, however, just is the process the GA describes as one of working on, organizing, sifting, separating – purifying – some matter. To purify the matter, in this way, just is to realize a thing’s form and substantial being in the way characteristic of concoction-1. The first upshot is that R, like G, is a process of purification, where this makes it a paradigmatic case of concoction-1.

G too, in fact – and this is the second upshot – is a paradigmatic case of concoction-1. For while GA 1.20 and 4.1 put less emphasis on G’s end, the wider context makes clear that in acting on the female’s matter in animal generation in a purifying way, the male principle is acting for the sake of something – something (as in R) with the form and substantial being of the relevant organism. It is, after all, the basic form of reproduction, which (we have seen) aims at just that end.^[81] Thus G, too, is a case of concoction-1 – for, like R, it reproduces the form and substantial being in just this way.

My proposal is thus that if we figure out what is distinctive about the active capacities of purification and their ends – what it is to pull and toss in such a way as to bring about a form and substantial being – we will better understand concoction-1. GA 2.6’s housekeeper analogy, moreover, supports this approach. It claims that in making the first (privileged) group of parts, “nature organizes flesh and the bodies of the other sense organs out of the purest matter” (744b22–24). I have already argued that these parts ought to be produced by concoction-1. Now, we can see the significance of the fact that they are concocted (744b13) out of the purest matter, in a process that separates pure stuff from leftovers (744b15–16, 19–20, 24, 27). They are produced by a purifying process: concoction-1.

What, then, is characteristic of that process and the capacities realized in it? My strategy will be to consider the difference between two processes to which Aristotle compares G. One, as we have seen, is R, which G resembles insofar as it is a case of purification. Another, however, is F: the action of rennet and fig juice on milk (1.20, 729a11–14; 2.4, 739b20–28). While F has something in common with G, it is also importantly different. Aristotle discusses this difference in GA 4.4, in a difficult passage concerned with why some animals produce not a single large offspring out of a large amount of material, but instead several smaller ones (771b14–18, 24). One of Aristotle’s steps in answering that question is to contrast F (771b25–27, 772a22–25), G (772a2–12, 17–22, 25–30), and a further case, H: the action of heat on water (772a12–17).^[82] His idea is that thinking about how these processes differ points to the features of G that explain the phenomenon of multiple offspring. For my purposes, the important point is that his answer appeals to the capacities (772a8, 10, 20, 29) realized in those processes. My argument will be that whereas G is a paradigmatic case of concoction-1, F is a paradigmatic case of concoction-2. Moreover, the very features that distinguish F from G are the ones that make F satisfy Meteor. 4.2’s account of concoction-2 – where it is highly plausible to think G’s contrasting features make it a case of purification and so concoction-1.

Aristotle starts by asking why G is not like F: “on account of what cause does it not [in G] bring to perfection one animal of great size, just as in the other case [F] the fig juice does not get divided in organizing a certain quantity but rather what is solidified is greater by the quantity by which it [the juice] comes into more [i. e., more milk] and is more [i. e., is a greater amount of fig juice]?” (771b24–27). F is a case where the agent somehow brings about (a) a particular quantity, (b) at least in part in virtue of its own quantity and that of the patient, (c) in such a way that a greater quantity of agent and patient results in a greater degree of effect on the patient. This sets the stage for the discussion to come: what is at stake is exactly what the agents of F, G, and H produce – and why.

How exactly G differs from F becomes clearer when Aristotle considers H, which apparently has more in common with G than F – in certain respects, at least^[83] – and so provides a clue to G’s nature. In H, “neither does the fire heat the water more by whatever quantity by which it is more, but rather there is a certain determinate limit (ὅρος) of the heat, which when it is present, if someone increases the fire, the water no longer comes to be more hot but evaporates more and in the end disappears” (772a12–16). H is thus like F with regard to (a), but differs with regard to (b) and (c): it is not the case that a greater quantity (of the agent) results in a greater (quantitative) effect. The reason is that the agent – the heat – produces the effect in virtue of having a certain ὅρος – by implication not primarily in virtue of its own quantity, etc. This ὅρος, Aristotle’s later remarks on G suggest, is of the heat’s capacity (772a8, 772a20, and 772a29) – the active capacity of the agent. His idea is that this heat is (or has) a capacity not just for heating water, but rather for doing so to a certain degree.^[84] When this is so, as in H, having more of the stuff that has the capacity does not increase the effect. When it is not, as in F, it does.

This discussion of H helps Aristotle understand G: just as H’s agent’s capacity has a determinate limit, so too does G’s. Thus only “as many [animals] as are determined (ὥρισται) by the sizes that are fitting” can come to be, “and neither the seed of the male nor the capacity in the seed will organize anything more or less than what is natural” (772a6–9). Here too the active capacity has a determinate limit: it is for producing animals of a certain size. This, in turn, has consequences for the amount of matter needed to make an animal of that size: this is “not unlimited” (772a3). Aristotle sums this up by noting that “the capacity of what is acted on and of the heat that acts are determinately limited (ὥρισται)” (772a28–30; cf. 772a20).^[85] The upshot is that G and H differ from F according to (c) and (b).^[86]

Aristotle’s point is that it is because the active capacity has this ὅρος that the processes in which it is realized characteristically – that is, successfully – produce animals of that size. (Here too, his focus is not on cases where things go awry, but on how animals are “perfected,” 771b33.) That ὅρος both identifies the size that counts as success in the relevant process – a process of generation for animals of a species – and also is realized in the process of producing them. As in H, it is a ὅρος of the active capacity: the relevant heat.

This is enough for Aristotle to answer his question about multiple offspring (772a25–30). Still, he adds a further point, noting another difference between G and F: “the milk example mentioned is not similar; for the heat of the seed not only organizes a quantity but also a certain quality, but that [heat] in the fig juice and the rennet only the quantity” (772a22–25).^[87] The upshot is that G also differs from F – and presumably H – with respect to (a): the agent produces not only a certain quantity (cf. 771b25), but also a quality. The three cases thus differ as follows:

1. Agent produces quantity + greater quantity of agent or patient results in greater effect + due to there being no (quantitative) determinate limit to what the agent’s capacity is for

2. Agent produces quantity and quality + greater quantity of agent or patient does not result in greater effect + due to a (quantitative and qualitative) determinate limit to what the agent’s capacity is for

3. Agent produces quantity + greater quantity of agent or patient does not result in greater effect + due to a (quantitative) determinate limit to what the agent’s capacity is for

G thus differs from F in that its agent produces a specific quantity and quality, due to a capacity for it.

To say F’s capacity is not like this, importantly, need not mean that what is produced by its activity – cheese curds – lacks such a quantity. GA 4.4 is explicit that the heat produces a specific quantity (771b25), and Meteor. 4.12 talks of the shapes (σχήματα) of curds (390a24–b2); indeed, any concoction that goes as it should results in a “determinate body” (ὡρισμένον σῶμα, 382a2; cf. 4, 22, 26). Insofar as these cases are at all good comparanda for G, these should be results those concoctions produce non-accidentally. Aristotle’s point is just that in cases like F, unlike in G, something other than what the active capacity is for determines the exact quantity and quality. That quantity is a function of the amount of agent (and patient) present, and the quality presumably of something else – perhaps, as with γλεῦκος, what the passive capacity is for. After all, the process is an exercise of that capacity too.

We are almost in a position to see how exactly these distinctions bear on the distinction between concoction-1 and -2. The final thing we need to do is to recognize that F is a paradigmatic case of concoction-2. This comes out in Meteor. 4.7, which observes that “the earthy bit [in milk] is organized also by the fig juice, if somehow someone boils it, as doctors do when they curdle it.” This is how “the whey and the cheese are separated” (384a20–22). Several signs suggest this a typical case of concoction-2. For one thing, the agent is not just (the heat of) fig juice or rennet, but also (the heat of) a doctor. Doctors are among the agents in view in 4.3’s discussion of boiling’s “uses,” which mentions the use for which “drugs” are “boiled” (381a3–4) – where these “uses” are characteristic ends of concoction-2. Moreover, these examples of boiling for a use include both my central example of concoction-2 (new wine) and also – right alongside – the boiling of milk, which is likewise a case of flavor change: “moist things are boiled, for instance, milk and new wine, when the flavor in the moist changes into a certain form” (4.3, 380b31–33). The upshot is that insofar as the boiling of milk, with the addition of fig juice or rennet (cf. 4.7, 384a20–22), is (a) a case of boiling where the end is usefulness for a use and (b) a case of flavor change – and like the γλεῦκος case in these respects – we should take it to be a case of concoction-2.

The connection is closer still: F is not just a paradigmatic case of concoction-2 – but is so in virtue of being the exercise of the active capacity GA 4.4 attributes to its agent. Notice what Meteor. 4 suggests about the patient’s role. 4.3 emphasizes that when “someone boils it,” it is the milk’s “earthy bit” that is “organized also by the fig juice” (384a20–21); thus “if some milk does not have cheese [in it] or has little, this is more of water,” which burns off (384a23–25). This observation aligns F with the γλεῦκος case: the implication is that it is the milk (the earthy bit in it) and not the heat that ensures that the process produces something cheesy. In F as in the γλεῦκος case, we can think of the earthy qualities of the patient as supplying certain quantitative and qualitative features of the process’s end. These features, Meteor. 4.2 suggests, constitute the “underlying character” to which the process’s success condition refers. Aristotle’s view is thus that when the active capacity is not for a product with certain determinately limited qualitative and quantitative features, there is ‘room’ for something else – the patient – to determine which features the process in which they are realized characteristically produces: its success condition.^[88] (This ‘room’ thus just is the non-determinately limited aspects of the agent’s end; I adopt the metaphor for ease of expression. See also Section 6 for further articulation.) But this active capacity, we have seen, just is the active capacity of an agent that aims at making things useful for a use. This is the sort of active capacity GA 4.4 is describing: it is for something – which Meteor. 4 suggests is usefulness for a use – that is not determinately limited in this way.

Meteor. 4.2 is suggesting that when that capacity is realized together with the patient’s capacity, the resulting process has as its end (its success condition) not just usefulness for a given use, but usefulness that is also specified with respect to certain determinate qualitative and quantitative features set by the patient’s passive capacity. What this means is that the very features of the active capacity that distinguish F from G are features that help to make it a case of concoction-2 – and also explain why it has the features highlighted in Meteor. 4.2–3.

This account of concoction-2’s active capacities is, moreover, independently plausible. Consider a case. Like many cases of concoction-2, this is a craft case.^[89] Suppose I rummage through the kitchen, finding radishes, squash, and onions. I roast them at 425 °, trying to make them suitable for eating (among the “uses” in 4.3). To succeed, I (and my capacity) only need to make them acquire a flavor “useful” for eating; I do not need them all to taste like a radish. Why might this be? The active capacity here is the cookery craft (cf. Metaph. Θ.2, 1046b2–3); a flavor useful for eating is an end I aim at insofar as I am a cook. And it would be a strange conception of that craft that posited separate crafts for making radish and onion flavors.^[90] A craft is a sort of knowledge, which prepares one to deal with a variety of situations (Metaph. A) – perhaps, the knowledge of how to make a variety of things useful for eating.^[91] In dealing with this variety, I must be flexible about exactly what counts as success (for me, as a possessor of my craft): if all I have is an onion, I will fail if my craft’s aim is to make a radish flavor – but if it is to make something useful for eating, there are many ways for me to succeed, and to do so insofar as I have the cookery craft. But what it is for me to succeed as a cook may well come apart from what it is for the processes I cause to succeed: when I am roasting a radish, that roasting process will fail if the result tastes like a squash.^[92]

All this suggests, in turn, that the features GA 4.4 presents as distinctive of G’s active capacity also help make it a case of purification and so concoction-1. This too is independently plausible: for the purificatory activity in G to be non-accidentally caused, as reproduction is, it should be the exercise of a capacity that is for some such activity. GA 4.4 is just telling us what this (active) capacity is like. The male undertakes its purificatory activity in virtue of its active capacity for a specific quantity and quality; this just is the active capacity that causes purification. This account of purification’s active capacities also yields an independently plausible account of R. Ripening’s end is the same in form as the agent that produces it, where this is a natural thing’s formal “nature […] in the sense of form and substantial being.” In other words, the agents of such processes tend to be formal natures, and the ends they aim at and that are their success conditions are likewise formal natures – of the same sort. And those forms ought to be specified in specific qualitative or quantitative ways; otherwise, a ripe ‘apple’ that smelled like a grape and was the size of a melon would count as a successful product of the active capacity exercised in an apple’s ripening.^[93] Taking this result together with Meteor. 4.2’s account of concoction-1’s end suggests that what distinguishes concoction-1 is that it aims at an end the specific features of which are set by what its active capacity is for. The process has that success condition because the active capacity does.

I thus conclude that the active capacities of concoction-2 – those that aim at making things useful for a use – are not for a specific quantity and quality. They leave ‘room’ for something else to determine such features – as when “the end […] is in regard to a certain underlying character, when what is moist comes to be such-and-such and so-much” (4.2, 379b25–28). This something else just is the success condition of the patient’s passive capacity: the features it supplies are not mere accidents of the context, but features with a good claim, insofar as they are in that capacity’s success condition, to being essential to the nature of the process in which it is actualized. The active capacities for concoction-1 (those that aim at form and substantial being), by contrast, do aim at specific quantities and qualities, and so do not leave ‘room’ of that sort.

5 The Housekeeper’s Concoction

One of my main claims has been that this distinction lies behind GA 2.6’s housekeeper analogy.^[94] Now, I want to revisit that analogy, asking (in a preliminary way) why it might matter in that context. Roughly, the structure of GA 2.6 is this: it introduces a question about the order in which the parts come to be (741b25–742a16); sketches a method for answering it (742a16–743a1); considers the generation of the uniform parts (743a1–b18, cf. Meteor. 388a16–17), including the role of heat and coldness (743a26–b5); and returns, more explicitly, to its question about order (743b18–745b22). The housekeeper analogy comes in this last section.

In context, that analogy’s role is to apply the lessons of Aristotle’s method and his account of the uniform parts to the order in which certain parts come to be.^[95] Here’s how. Aristotle’s method involves distinguishing the “useful” parts (742a32) from the rest – and the housekeeper analogy characterizes the generation of the less privileged parts as a case of “mak[ing] something useful” (744b17). It also claims that they come to be later (744b26–27), out of residues (744b15–16, 24, 27) – and here it also picks up the discussion of the uniform parts. There, Aristotle claimed that nails, horns, hoofs, and beaks come to be when “as many of the things rising up as are very earthy, having little moisture and heat, and these, as they cool, when the moist is being evaporated off with the heat, come to be hard and earthy in their character (τὴν μορφήν), such as nails and horns and hoofs and beaks” (743a11–15, cf. DA 1.5, 410b1; 3.13, 435a24).^[96] While this remark does not mention residues by name, it is about the parts the housekeeper analogy will present as made of residues (744b25). And the “very earthy” materials “rising up” (743a11–12) are surely residues, which are very earthy (745b15–20; PA 3.2, 663b31–35).

In picking up these threads, I want to emphasize, the housekeeper analogy is also picking up typical features of concoction-2. In such processes, the end is “useful” and specified with regard to an “underlying character” – one, we have seen, often associated with earth, and which Aristotle here calls a “hard and earthy” “character” (τὴν μορφήν). Read against the background of Meteor. 4’s theory,^[97] then, part of the analogy’s point turns out to be that in the generation of the useful parts, that “very earthy” (743a12) stuff – the patient – plays a role in setting the end, alongside an agent’s “use.” That agent, of course, is the one that “make[s] something useful” out of the residues (744b17). Such agents, I have argued, exercise capacities for ends that are not determinately limited and so ‘leave room’ for the patient’s capacity’s success condition (a “hard and earthy” “character”) to supply part of the process’s success condition. Part of the point of the housekeeper analogy is thus to suggest that this just is the sort of agency that allows the “very earthy” stuff to play this end-setting role – that leaves ‘room’ for it to do so.

How, precisely, does it do so? In this section, I will (tentatively) suggest that Aristotle’s account of the details of the generation of hair – one less privileged part (744b25) – when read against the background of my interpretation of concoction-1 and -2, suggests an answer. Aristotle discusses this case in more detail in PA 2.14–15.^[98]

The first step is to notice that those remarks fill out GA 2.6’s talk of ‘making useful.’ This is most explicit in PA 2.15’s discussion of how eyelashes and eyebrows – examples of hair – come to be:

The eyebrows and eyelashes are both for the sake of protection. The eyebrows are for the sake of the moisture running downwards, so that like eaves they may shelter the eyes from the moisture from the head. The eyelashes are for the sake of things falling towards the eyes, like the palisades sometimes put up in front of walls. The eyebrows are at the conjunction of bones, which is also why in many cases they become bushy as people age, so that they need to be trimmed. The eyelids, however, are at the ends of small blood vessels; for where the skin terminates, the small blood vessels also reach their limit. The result is that because the moist secretions oozing out are bodily, it is necessary that – unless some function of nature impedes it for another use (χρῆσιν) – indeed owing to a cause such as this, hair from necessity comes to be in these locations. (658b14–27)

In suggesting that there is “another use” to which (πρὸς, 658b24) nature might have put the “secretions,” Aristotle implies that the “oozing” process by which they come to form eyelashes already has one: the “protection” eyelashes are “for the sake of” (658b14–16).^[99] That this is a “use” is reinforced by PA 3.1’s claim that nature gives parts for “protection” to those able to “use” them (661b28–30). Now, insofar as this is a case of concoction-2, this will be a “use” for which hair is “useful” – it is, after all, a “useful” part (GA 2.6, 744b17, 742a32) – and respect to which the “useful” end of its production process is specified. It should also be, in some way, a use at which the agent – nature (742a29–36, 744b16, 22) – aims, a fact PA 2.15 confirms (658b24–25). And this use – protection – is indeed some sort of end, as PA 2.14’s discussion of human head hair brings out:

with respect to the head, the human is the most hairy of animals from necessity, on account of the moistness of the brain and on account of the sutures (for where there is much moisture and heat there is necessarily much growth), and for the sake of protection, so that it may provide covering, warding off the extremes of both cold and heat. And since the human brain is the most moist, it is also most in need of this protection. (658b2–9, cf. HA 2.1, 498b19–20)

More generally, “hair is present in those that have it for the sake of covering” (2.14, 658a18–19). Protection (perhaps covering or shelter^[100]) is a use that is at least part of the (process’s) end.

The upshot of all this is that PA 2.14–15 fills out GA 2.6’s account of hair production as a case of “mak[ing] something useful.” What, then, does nature do in this process? PA 2.15’s slightly mysterious answer is that “because the moist secretions oozing out are bodily, it is necessary that – unless some function of nature impedes it for another use – indeed owing to a cause such as this, hair from necessity comes to be in these locations” (658b22–27). On any reading, this remark allows the patient an important role in explaining the process’s result; it also seems to limit the agent’s role in some way – it just lets the secretion happen.^[101] This limited role, it appears, is the one that Aristotle, in filling out the housekeeper analogy’s talk of “mak[ing] something useful,” is attributing to nature.

What is that role like? Nature, it seems, is declining to reorient the secretions to another location (e. g. a place where hair would be eyebrows instead of lashes, 2.15, 658b14–22); its primary concern is with picking a place that benefits the creature in question. (Humans, for instance, need protection on their more valuable parts and around their moist, chilly heads: 2.14, 658a23, b9–10^[102].) As Aristotle puts it elsewhere, “nature always provides each thing” – including all the parts that are “for protection” – “to those able to use it,” where this means putting larger and stronger such parts on the side where they will do more good (4.8, 684a26–32). And – in the generation of another useful part – the “residual surplus of this sort of [i. e., earthy, 3.2, 663b29] […] nature uses for protection and advantage, and the surplus, which flows of necessity to the upper region, in some cases it distributes to teeth and tusks, in other cases to horns” (3.2, 663b31–35, cf. 20–24).

All this is just what nature does in “us[ing]” the flowing materials – in acting in light of the “use” at which it aims. It is just the activity that counts as acting in light of that end. This end is therefore concerned with the location to which the surplus should be “distribute[d].” To make that surplus useful is to put it where it will be useful to the animal in question. But why is this end so particularly concerned with location? And how?^[103] Here my interpretation suggests the germ of an answer: nature is aiming at an end that, as in GA 4.4’s F, ‘leaves room’ for the patient’s capacity’s success condition to help set the process’ end. My suggestion is that if this is to happen, determining the location is what is ‘left’ for nature to do.^[104]

To see how this works, we need to look more closely at the patient’s success condition. The patient in question, GA 2.6 suggests, is an earthy residue: a typical concoction-2 patient. Indeed, it is because hair is made “of residue” that humans have the least hairy “body” – since “residue is what is unconcocted, and what is earthy is of all things in bodies the least concocted” (745b15–20). PA 2.15’s “oozing” “secretions” are “bodily” (658b23) – where earth is very bodily (784b8–10, 761a33–35).^[105] More generally, PA 2.14 emphasizes features of the patient (e. g., moisture; 658b9). And PA 2.19, in its discussion of parts that are “for the sake of protection” (655b5),^[106] claims that “an earthen and hard nature” “is the defensive capacity” (655b11–13, cf. 3.2, 663b31–35, 662b31–32).^[107]

My proposal is thus that what the “use” leaves ‘room’ for is this matter’s passive capacities’ success conditions. But what are they? GA 5.3’s account of hair’s differentiae provides a clue.^[108] It begins:

For the sake of what nature made the kind of hair for animals has been said earlier in the causes concerning the parts of animals; but to clarify what things being present and on account of what necessities each of these things occurs is the task of the present inquiry. (782a20–24)

The reference is to PA 2.15 and 2.14: hair is for protection, etc. Now, Aristotle is proposing to examine the underpinnings of that account, giving an explanation that follows GA 5’s typical pattern.^[109] Such explanations appeal to how something “naturally acts and suffers such-and-such or such-and-such (τοιονδὶ ἢ τοιονδὶ)” (778b18–19, cf. 778a35–b1, b14–15). They therefore promise to illustrate Meteor. 4.2’s claim that in concoction-2 “what is moist comes to be such-and-such and so-much (τοιονδὶ […] καὶ τοσονδὶ)” (379b27–28). In the case of hair, the relevant qualities and quantities are the διαφοραὶ (5.3, 782a1, cf. 5.1, 778a16, b15 and ἕκαστον, 5.3, 782a23) of hair: its variations in texture, length, etc. To the extent that hair generation follows the concoction-2 pattern, these will be the quantities and qualities (like the flavor of new wine) that PA 2.15’s “use” leaves room for.

And, in fact, they are among the qualities mentioned in PA 2.15. For in giving its account of these quantities and qualities, GA 5 just is filling out the case that was the particular focus of the PA 2.14 passage just quoted. As GA 5 formulates it, (a) “the hairs on the head are thickest in humans” (5.3, 782b9–10) and (b) “in humans the hairs on the head are longest” (782b16). Its explanation of (a) and (b) also picks up PA 2.14’s appeal to (1) “much moisture and heat” and (2) “protection” and PA 2.15’s appeal (in an analogous case) to (3) earthiness. It does so by applying its more general account of hair’s (a) thickness and thinness and (b) length and shortness. These variations result from the fact that hair is made of skin (782a30–33, cf. GA 2.6, 745a20–21 and HA 3.10–11). For instance, “of thickness and thinness, the skin is a cause most of all” – though “the differentiation of the moisture in it is also a co-cause” (5.3, 782a24–27). This fills out (3) earthiness – for the skin has an “earthy” nature, indeed, one that “underlies” (782a28–29). A greater “amount of earth” thus produces thicker hair (782a33–b1), which is why (a) “the hairs on the head are thickest in humans” (782b9–10). And in a parallel to (1), “the brain – being moist and cold – supplies much abundance of the moist” (782b17–18) – which is why (b) “in humans the hairs on the head are longest” (782b16). The point of (1) and (3), as spelled out by GA 5.3, is that a feature of the patient (skin) explains why the product is (a) thick and (b) long. These are the quantities and qualities this process, as a case of concoction-2, has in its success condition – by being the exercise of those passive capacities.

All this points to an explanation for nature’s concern with location: this is the sort of concern that allows it to leave room for the patient to supply these features to the process’s end. In ensuring that the long, thick stuff goes to the right place, it is putting it where it will be useful – and making it useful for a use at which it aims – but it is not aiming at thickness and length. If the process has those ends, it is not because its agent’s active capacity is for them; that capacity is for usefulness, and it is implemented by selecting the right location.

There is a deeper question here: why is a capacity for making something “useful” for a “use” the sort of capacity that does not aim at such specific qualities and quantities? A full answer will take us too far afield, but here are the seeds of one tentative hypothesis.^[110] It emerges from the fact that the only “end” explicitly mentioned in GA 2.6 (the housekeeper analogy’s context) is “the whole and the end” (742a35, cf. 742b17) that contains the privileged parts, but not the “useful” ones.^[111] To the extent that this is – in a chapter focused on what nature does in animal generation – the only “end” explicitly mentioned, it should be the end at which nature aims in animal generation. It is, moreover, what the “principle” (the efficient cause the housekeeper analogy will soon call “nature”) aims at and is part of (2.1, 732a22–27, 32–35); in context, the implication is that this is the end that principle aims at also when it causes generation (2.6, 742a24–25, b3–5).

But if this is the end nature aims at in animal generation, in making the parts of an animal it will be aiming at all and only the “parts” of that end (742b17), not – at least, not in just the same way – at the less privileged parts.^[112] At the same time, nature does make them, and not for no reason. It declines to redirect the oozing secretions due to its concern with the “use” for which they are “useful,” where that use is also what the relevant parts are for the sake of. Is there, then, some sense in which it does aim at them? And what is it?

The key is to note that the useful parts are “for the sake of” the parts in the end (744b14–15) – and that those (privileged) parts “use” them for “various uses” (742a32, 36). These “uses” are the actions that define the “useful” parts (cf. PA 1.5, 645b14–19). To the extent that the useful parts are for the sake of the privileged ones, their defining uses are likewise for the sake of the actions that define the privileged parts (PA 645b29–32). This, I suggest, allows us to identify a derivative way in which nature aims at the “useful” parts, insofar as it aims at the privileged parts in the end.^[113]

The idea – which builds on a proposal from Leunissen and Gotthelf^[114] – is that one and the same end (“the whole and the end”) sets the active capacity’s success condition, both when it makes the parts in the end and also when it makes the other parts. That end, in other words, determines what counts as useful.^[115] Here is one way of working this out. Aristotle’s view is that “what is capable is capable of something and at some time and in some way and as many other things as it is necessary that there be added in the definition” (Metaph. Θ.5, 1048a1–2). Nature, for instance, is or has an active capacity to make the privileged parts (to φ) in certain circumstances (C), for example, when confronted with pure matter. And this makes it derivatively able to, in other circumstances C*, make other parts that are “useful” for these (to φ*). The reason is that what those parts are useful for is an activity that defines the parts that are the aim of φ-ing: the capacity to φ* is derivative on the capacity to φ in that they, ultimately, share the same end.

Importantly – and here is the key point for my purposes – there are many things that are useful to an animal. Any one of the parts in its end might benefit from getting to do its task more efficiently, or with better results, or shielded from interference. And not all of them take an earthy form. Kidneys serve another part – the bladder (PA 3.7, 670b24–27) – but are exceptionally “fatty” (3.9, 672a1, 13), where this is “soft fat” rather than earthy “hard fat” (2.5, 651a26). In aiming at usefulness, then, nature must aim at an end that does not specify any one such set of quantities and qualities: that – to pick up the point I have developed elsewhere – is open to the available matter as to what form the usefulness takes.^[116] It must let that matter set those aspects of the success condition of the process in which it is exercised; if that matter is human skin, that process – when successful – will produce something long and thick.^[117] The long and thick way of being useful is, perhaps, the protective, covering sort.^[118] (Kidneys, by contrast, are not useful to the bladder by protecting it, but by filtering its residue, 3.9, 671b18–22.) And nature, aiming at making something useful, just puts in the right – useful – locations; this is the sort of concoction-2 agency it is exercising. But the process in which that agency is exercised aims not just at usefulness, but usefulness of the long, thick sort.

6 Concluding Remarks

One thing that distinguishes concoction-2, then, is the complexity of its success condition: it is specified with respect to the success conditions of both the active and passive capacities it realizes. This interpretation gives us the resources to explain one way – again, here I am only taking one step towards a full account, and there is more to do – the agency nature exhibits in “mak[ing] something useful” might be distinctive. It is true enough that on some level it does much the same thing as it does in concocting(-1) the more privileged parts: it acts for the sake of something. But this is a very general claim. When we look at the details, what it acts for the sake of – the activity that defines its capacity – is different: a kind of usefulness that leaves room for something else to determine the exact form it takes.

To return to Henry’s example, a builder who makes window boxes out of leftover materials exercises a capacity for making something useful for houses – not for making window boxes. She will succeed if she makes window boxes or adds a nice trim. But the process in which she exercises her capacity – which is also the exercise of a passive capacity – has a different success condition: it makes something useful for houses in a way that exploits the materials at hand. In this way, distinguishing between the ends of the process and the ends of the capacities exercised in it equips us to respond to Henry’s objection to interpretations (like Leunissen’s) that take there to be a real difference in how nature acts in making the less privileged parts – things like window boxes. Consider, for example, Leunissen’s claim that such things “come to be of material necessity, but are present for a function.”^[119] Henry’s concern, as we have seen, is this:

Even if we agree that the animal’s formal nature makes these parts using resources that are lying around thanks to material necessity, why should that fact entail that the resulting parts do not come to be for the sake of protection and advantage? Suppose that as a necessary consequence of how lumber is packaged a builder discovers that she has leftover materials that were not specified in the original blueprints. If she uses that extra lumber to build window boxes, we would still say that those window boxes came to be for the sake of holding flowers (that she made the boxes in order to hold flowers). (Henry 2019, 191)

What my interpretation reveals is that even if it is true to say that the window boxes (and eyelashes!) come to be for the sake of something, this need not mean that there is nothing distinctive about the process that produces them or the capacities exercised in it. In my view, after all, that process (the coming-to-be) does indeed have as its success condition something that is useful in the particular way in which window boxes are useful (the way that exploits the materials to hand). To this extent, we should accept that window boxes and eyelashes come to be for the sake of something, just as Henry claims: they are, after all, the results of a process that happens for the sake of such things. But to accept this is not to claim, with Henry, that there is nothing distinctive about that process, or the agency nature exhibits in causing it. Its success conditions, after all, are set in a distinctive way, in virtue of the distinctive sort of active capacity nature is exercising in causing it.^[120]

There is an interesting question about how to formulate this view. Is the process’s success condition just a combination of features, some taken from what the active capacity is for and some from what the passive capacity is for? In the cases I have considered, the relationship is closer: having such-and-such a tasty flavor is a way of being useful for eating, having such-and-such healing qualities is a way of being medically useful, and being thick and long (etc.) is one way of being useful for human life. These examples suggest a way of making the view precise: in terms of the machinery of determinables, determinates, and determinants. (This is certainly not the only way of spelling out the view, and the basic picture I have proposed can be considered independently of it.)

This machinery has, of course, good Aristotelian credentials, and many recent interpreters have found it useful in thinking about Aristotle’s accounts of matter, form, and – especially germane to my purposes – capacities.^[121] Indeed, Charles and Peramatzis have argued that the Metaphysics’ language of determinacy ought to be understood in just this way. Metaph. Z.11’s claim that matter is “indeterminate” (ἀόριστον, 1037a27), for instance, may be understood as claiming that it is determinable – because it needs to be made determinate by the form, functioning as a determinant (1049a27–b3).^[122] Charles has also argued that there are cases where the form is itself a determinate combination of a determinable and a determinant.^[123] And Salmieri has argued that we can think of animal generation as succeeding when it transmits certain determinable features.^[124] And I have argued elsewhere that some possess determinable capacities whose exercise is made determinate by the patient.^[125]

We also find the language of determination in passages I have considered here. Notice how GA 4.4 describes M: “as many as are determined (ὥρισται) by the sizes that are fitting” can come to be” (772a7–8) – where this is to say that “there is a certain determinate limit (ὅρος) of the heat” (772a14, cf. 772a28–30 with ὥρισται). The heat is or has the active capacity: it is a capacity for something with a determinate size. The point is that F’s active capacity’s success condition lacks a determinate limit. That is why the passive capacities’ success conditions can supply the process with a determinate end. It is easy enough to see how they might do so: that passive cause is dry and earthy, and the dry is determined with difficulty (δυσόριστον, Meteor. 381b29–31). All it needs to do is to supply those qualities where the “determining” heat (378b15, 379b34) allows it to.

My proposal, then, is that we may think of the process as having a determinate success condition – where the agent’s active capacity sometimes does and sometimes does not have this fully determinate end as its success condition. In the latter (concoction-2) case, the active capacity’s success condition is determinable in certain respects: about which qualitative and quantitative features are to be produced. The process in which it is exercised, however, has a determinate success condition because the patient’s passive capacity’s success condition functions as a determinant.

For instance, a doctor’s active capacity might aim at a determinable end: medical usefulness.^[126] When she exercises it on a particular wine, that wine’s passive capacity’s success condition functions as a determinant. For the resulting process to succeed, it must produce this determinate medically useful flavor. When a process with this success condition unfolds, this material (the patient – not in the sense of the person a doctor treats, but in the sense of the passive recipient of the doctor’s agency) not only contributes certain specific features to the result (as any patient does), but does so in ways that the agent’s end does not specifically require, and therefore not because it specifically requires them – but because the success condition for the process does. And that condition, I have argued, is specified with respect to the success conditions of both the active and passive capacities it realizes (Phys. 202b26–7) – related as determinable and determinant.