Jump to ContentJump to Main Navigation
The Powers of Aristotle's Soul$

Thomas Kjeller Johansen

Print publication date: 2012

Print ISBN-13: 9780199658435

Published to Oxford Scholarship Online: January 2013

DOI: 10.1093/acprof:oso/9780199658435.001.0001

Show Summary Details
Page of

PRINTED FROM OXFORD SCHOLARSHIP ONLINE (oxford.universitypressscholarship.com). (c) Copyright Oxford University Press, 2020. All Rights Reserved. An individual user may print out a PDF of a single chapter of a monograph in OSO for personal use.  Subscriber: null; date: 18 September 2020

Towards a Scientific Definition of the Soul: DA II.2

Towards a Scientific Definition of the Soul: DA II.2

(p.34) 2 Towards a Scientific Definition of the Soul: DA II.2
The Powers of Aristotle's Soul

Thomas Kjeller Johansen

Oxford University Press

Abstract and Keywords

Aristotle's argues in DA II.2 that a proper causal definition of the soul is required if we are to know what it is. The attempt to identify the essence of the soul relies on a common conception of it as the cause of life. Since life is said in many ways, the definition of soul will be differentiated accordingly. This approach should be seen in the context of Aristotle's theory of demonstration, as articulated in the APo. The definition of the soul will provide the explanatory principles for demonstrations of the characteristic attributions of soul, and, by extension, of whatever attributes living beings have qua ensouled. In accordance with the APo's view of demonstration in natural explanation, the soul's role in a demonstrative syllogism could be understood both as a formal, final, and efficient cause. The account of the soul in this way provides the explanatory basis for the biological works. The objections that have been raised against applying the model of demonstrative reasoning to the biological works do not count strongly against my interpretation of the DA.

Keywords:   definition, essence, life, demonstration, four causes

The most common account was, first of all, a response to the first two questions raised in DA I.1: to which of the ten categories does the soul belong? Is it a fulfilment or a capacity? It is clear that answering these two questions cannot be an attempt to define what the soul is, its essence, in any precise way. It is rather a preliminary to defining the soul in the sense that it tells us what kind of definition it is appropriate to demand for the soul. Similarly, if I ask you what colour is and I’m told that it is a quality of bodies, with the potentiality to cause vision, then this account may roughly pick out colours from other attributes, but it is not a proper definition of the essence of colour. I will have a rough idea of what you are talking about, at least enough to inquire further into the nature of colour. So also the common account of the soul gives out an outline, as Aristotle says, of the sort of thing the soul is, to be filled in by an account of its essence. An essential definition of the soul will have to comply with the outline account. As we shall see, Aristotle, once he has given such a definition, will check that it fits the outline account.

It is confirmed right from the start of DA II.2 that Aristotle does not think we have defined the soul properly yet:

Since clarity and what is better known in account (kata logon) comes about from things that are less clear but more obvious, we should try again to approach the soul in this way. For the defining account should not only show the fact [‘the that’], in the way the majority of definitions do, but also contain and present the cause. For as it is the propositions (logoi) of the definitions are like conclusions. For example, what is squaring? The construction of an equilateral rectangle equal in area to a given oblong. Such a definition is a proposition proper to a conclusion. But a definition which says that squaring is finding a mean states the cause of the matter. (413a11–20)

Aristotle here recommends a new approach. We shall start with what is more obvious to us about the soul and work towards what is most knowable in account. This will be the definition of the essence of soul. This definition will differ from those generally offered by showing not just what the object is but why it is. Aristotle is drawing on his distinction in APo II.10 between three kinds of definition. Scientific inquiry may be seen as proceeding in three stages. Given that we possess a nominal definition of a phenomenon, we can inquire into its existence and then into its essence. This is the (p.35) view of scientific inquiry argued for by David Charles,1 but as far as my argument is concerned nothing hangs on whether we accept a two or a three stage view, since it is assumed that we know that the soul exists. (None of the views of the soul canvassed in Book 1 is eliminativist.) The key point is that we may start a scientific inquiry into the essence of something because we already possess a definition or account which, while it falls short of a full definition of the essence, does indicate a feature that directs us towards the essence. Such a definition may state part of the essence or merely a necessary consequent of the essence, but in either case it will fall short of a full statement of the essence. Such a definition will often be a statement of what the phenomenon is generally taken to be. The example of thunder helps make the point. The nominal definition of thunder is a noise in the clouds, and the definition of its essence is a noise made by fire being extinguished in the clouds. Here ‘noise in the clouds’ is part of the essence of thunder but it is ‘noise from fire being extinguished in the clouds’ which states the full essence and the cause of the phenomenon. Because the essence plays the role of cause in relation to the other necessary features of thunder, we can use the definition of the essence to explain the nominal definition.

Aristotle takes the nominal definition to play the role of the conclusion in a demonstrative syllogism, while the essential definition provides the middle term through which the conclusion is demonstrated. Similarly, in the example Aristotle gives in DA II.2, the definition of squaring as ‘the construction of an equilateral rectangle equal in area to a given oblong’ states the meaning of ‘squaring’ and works as the conclusion of a demonstrative syllogism, for which the essential definition provides the middle term. Methodologically, Aristotle is recommending that we start with a nominal definition of soul and then try to find the essential definition that will explain the nominal definition, just as we might start from the nominal definition of what squaring is and try to identify what explains why this is squaring. The reason why he commends this strategy is, as he says, that this nominal definition is more obvious. Being more obvious it provides us with a vantage point from which we can progress to the definition which is clearer and more intelligible.2 This is the definition which states the cause (tou pragmatos legei to aition).

If this is right, the opening of DA II.2 responds to the methodological question raised at the beginning of I.1: the method of inquiry will be the one recommended in Book 2 of the APo. We will start with a nominal definition of the soul and then proceed within the framework of the demonstrative syllogism to find a definition which explains the nominal definition. This was also the kind of answer suggested by Aristotle in DA I.1 when he said that the consequences could make a great contribution to the discovery of the essence.

But what sort of definition of the soul is ‘more obvious’ to us? The parallel with the APo and the example of squaring make us expect that the starting definition of soul will (p.36) be a nominal definition of soul, stating the, or a, common meaning of the term.3 It is clear that the definition has to state a property that is necessarily true of soul, and perhaps part of its essence, as noise in the clouds was part of the essence of thunder. Otherwise it will not play the role of a conclusion of a demonstration which proceeds through the essence.4 Finally, the property should follow necessarily from the essence of soul, while not itself reveal its complete essence. Otherwise, it would not just occur in the nominal definition. It is such attributes that Aristotle already referred to in the passage from DA I.1 as attributes ‘consequent on the nature or substance’.

Robert Bolton has suggested that the nominal definition is the common account of the soul in DA II.1,5 that is, the definition of the soul as a first fulfilment of an instrumental body potentially having life. I have already suggested what I take the function of the common account to be. But independently of this suggestion, it is clear that the common account does not fit the bill of a nominal definition for the purposes of DA II.2. Here again are the key lines:

Since clarity and what is better known in account (kata logon) comes about from things that are less clear but more obvious, we should try again to approach the soul in this way. For the defining account should not only show the fact [‘the that’], in the way the majority of definitions do, but also contain and present the cause. For as it is the propositions (logoi) of the definitions are like conclusions. For example, what is squaring? The construction of an equilateral rectangle equal in area to a given oblong. Such a definition is a proposition proper to a conclusion. But a definition which says that squaring is finding a mean states the cause of the matter. (413a11–20)

First of all, Bolton takes it that the reference in line 413a16 to the definitions that are like conclusions of an argument is primarily a reference to the definitions offered by Aristotle himself in II.1. We might think that the plural here reflects the fact II.1 in fact offered not one, but three (or, on Bolton’s count, four definitions) of the soul. However, this construal of the plural hardly squares with the claim two lines before that the majority of the definitions state the conclusion. Aristotle seems to be referring not primarily to his own definitions but to those that are currently in circulation, perhaps of the sorts referred to in his survey of other theories of the soul in DA I.3–5. Second, the common account seems to be too theoretically informed to play the role of nominal definition: terms such as ‘first fulfilment of an instrumental body’ hardly reflect common usage of the term ‘soul’ (psuchê).6 We are looking instead for an account of the soul comparable to the account of thunder as noise in the cloud or squaring as finding a square of the same size as a given rectangle, an account, that is, which stays closer to common or garden notions of the term. Third, it may be argued, as does Simplicius, that already the common account in DA II.1 had a causal aspect insofar as it refers to the soul as the form and fulfilment of the body, which suggests the (p.37) soul as a formal and final cause. Correspondingly, when Aristotle in II.4 (415b11–12) articulates the ways in which the soul is a cause he says that the soul is a cause in one way as the ousia of ensouled bodies. Here he seems to presuppose that the reader would already recognize ousia as the formal cause, alongside the efficient cause and the final cause. The causal aspect is problematic if we think that the point of the nominal definition is to state the fact to be explained rather than the causal explanation; and Aristotle’s example of squaring strongly suggests that this is part of his point. On the other hand, on the reading I suggested of the aim of DA II.1 it is not surprising that the common account should have causal implications since it is an outline of any definition of the essence of the soul. Without telling us what the essence is it tells us the kind of role it should play in relation to the body. Finally, the opening of DA II.2 asks for new starting point of inquiry also in the sense of a new account that is prior to us, that is, not one already given in the previous chapter.

The definition of the soul Aristotle chooses as his starting point is one that would indeed be obvious to us if we were Greeks of his time. It states a feature that is commonly understood by the term ‘soul’, indeed a feature that may plausible be taken to be just what the Greeks at this time generally understood by ‘soul’. Continuing from the last quotation, Aristotle says,

So (oun), taking up a starting point for the investigation, we say that the ensouled is distinguished from the soulless by living. But living is said in many ways and even if just one of these is present in something, we say that it lives, such as thinking (nous), perceiving, change and rest with respect to place and, moreover, change with respect to food and also diminution and growth.

The ‘so’ (oun) suggest that we now proceed according to the previous methodological reflections: we are now doing what Aristotle has just recommended, namely, starting with what is more obvious to us. As often in Aristotle’s investigations, what is more obvious to us coincides with ‘what we say’:7 ‘We say that the ensouled is distinguished from what has no soul by being alive’. ‘Being alive’ is, then, the distinguishing mark of what is ensouled. It seems, then, that we should seek an account of the essence of soul which explains why ensouled things are alive.

The idea that what has soul lives fits well with our notion of a nominal definition, since it captures a common, layman, understanding of the meaning of the term ‘soul’. So the Greek word for ‘soul’, psuchê, may be often translated as ‘life’ and ‘ensouled’ (empsuchon) as ‘living’.8 One reason why Socrates’ first account of death as the separation of soul from body is so readily accepted by his interlocutors in Plato’s Phaedo (64c) is that Greeks would think of something’s being alive insofar as soul is present. As in Homer,9 people die when the soul leaves the body, just as the presence of soul makes them alive.

The claim that what is ensouled is distinguished by being alive is a commonly held one. However, it is not immediately clear that the claim is of the right form to serve as (p.38) the conclusion of a demonstrative syllogism. There are two problems. First, the subject of the conclusion is not soul but what has soul or what is ensouled. We might expect the subject of the conclusion of a demonstrative syllogism through the essence of soul to be the soul itself rather than what is ensouled. So, for example, the soul is F, F is alive, therefore the soul is alive, where F would express the essential attribute of soul in virtue of which it is alive. The following is not a valid syllogism: the soul is F, F is alive, therefore ensouled beings are alive. It is not immediately clear how the definition of soul will provide the relevant middle term for a syllogism whose subject is what is ensouled. Contrast the kind of conclusion being offered by Xenocrates in APo II.4:

Those people who try to prove through conversion what the soul is (or what a man is, or anything else that exists) postulate what was set at the beginning. E.g. if someone were to claim that soul is what is itself explanatory of its own being alive, and that this is a number that moves itself: it is necessary to postulate that soul is just what is a number which moves itself, in the sense of being the same thing as it. (91a35–b2)

Here the syllogism serves to prove that the soul is what explains itself being alive, while Aristotle’s own analysis of the soul requires that it explains the body’s being alive. However, in this respect the case of the soul is not significantly different from Aristotle’s favoured examples in APo. II. Consider, again, the example of thunder: thunder is, according to the nominal definition, noise in the clouds. In this case the phenomenon to be explained is one that occurs in a certain subject. The definition of the essence of thunder similarly states that is the extinguishing of fire in the clouds. We might say, therefore, that the essence of soul relates to life in a body rather the way the essence of thunder relates to noise in the clouds. So the explanatory syllogism in the case of thunder would go: noise belongs to fire being extinguished; fire being extinguished belongs to the clouds, therefore noise belongs to the clouds (APo 93b8–12). Similarly in the case of the soul: life belongs to X,10 X belongs to what is ensouled, therefore life belongs to what is ensouled, where ‘X’ stands for an essential feature of soul. Here the essence of the soul generates a conclusion about what the soul belongs to, not about the soul itself. Similarly, in the example of thunder the conclusion is that noise belongs to the clouds. This is clearly because we take soul essentially to belong to something, just as we take thunder as a phenomenon that essentially belongs to clouds. We do not expect to find detached souls anymore than thunder apart from clouds. So the demonstration through an essential feature of soul or thunder can rightly generate conclusions about what is true of ensouled beings or clouds.

Platonic interlude: there is a parallel for Aristotle’s procedure in Plato’s Laws X (895c). Here the Athenian Stranger is seeking a definition of the essence of soul. He takes as his starting point two observations. The first is that whenever something (p.39) moves itself we say that it is alive, and the second that whenever we observe that there is soul in something we must agree that it is alive. The Stranger then proceeds to make a distinction between the thing, its name, and the account of its essence (ousia). He offers as the account of the essence of soul: ‘the motion that is able to move itself’. The overall argument seems to be that this definition of soul recommends itself because it explains that by virtue of which we say that ensouled beings live: we say that ensouled beings are alive because their soul moves them. In other words, the Athenian Stranger has offered us the tools to construct the following syllogism:

What has soul has self-motion. (895c)

What has self-motion is alive.

Therefore what has soul is alive. (895c)

If we accept this reading of the Laws passage, then it offers a clear parallel with Aristotle’s procedure in DA II.2. The explanandum is the same ‘what has soul is alive’, but the middle term of course differs, for, as DA I.3 shows, Aristotle rejects the claim that the soul moves itself. End of Platonic interlude.

Now for the second problem. We might say that what people generally think of the soul, and what Aristotle is really referring to when he says that ensouled things are alive, is that soul is what causes living beings to be alive. So the notion of soul is itself causal: soul is that which causes living beings to be alive. So it seems that the desired contrast between the conclusion which states a fact and the essential definition that states the cause is missing from Aristotle’s starting point here.

I would reply by conceding some of the objection, namely, that the claim that what is ensouled is distinguished by life is already meant to point to the soul as cause of life, while insisting that this claim can still play the role of the conclusion of an explanatory syllogism, in which the full essence is not displayed. For, we can distinguish between propositions that are causal in the sense that they are about one thing causing another and propositions that are causal in the sense that they successfully display what it is about the one thing that makes it the cause of the other. Compare Aristotle’s distinction between experience and knowledge in Metaph. I.1. The layman knows that hemlock kills but only the expert knows what it is about hemlock that kills. So it is part of the layman’s conception of hemlock that it has a certain effect, but he doesn’t know why it has this effect. Similarly, it would be part of the common conception of soul that it causes life, but it is still left to be explained why soul causes life. Therefore, it does not matter greatly for my argument whether we take Aristotle’s starting point at 413a21 to be that ensouled beings are alive or that they are alive because they are ensouled. For the essential feature because of which living beings are alive is still to be identified.

The causal account of the soul, then, will be that which fully explains the fact that ensouled beings are alive. But Aristotle goes on to point out (413a22–5) that this task breaks down into several explanatory tasks because being alive is not a simple thing: there are different ways in which one can be alive and so different ways in which the (p.40) soul can cause an ensouled being to be alive. Aristotle thus proceeds (413a25–b2) to explain the soul as the cause of each kind of being alive. First of all using nutrition is a way of being alive, characteristic of plants. This is explained by the capacity and principle (arkhê) ‘through which [plants] grow and diminish in opposite directions’ (413a26–8). Since nutrition is the minimal life form that all living beings must participate in, Aristotle also says that ‘living belongs to all living beings because of this principle’ (413b1–2). Next is the life activity of perception, characteristic of animals. The cause of this is the sense-faculties (413b2).

What needs to be shown is the way in which the various kinds of soul are causes and principles of the various kinds of life forms. And this is what the rest of DA shows. At this stage, Aristotle is content with summing up the causal role of the soul in relation to the various activities at 413b11: ‘For now, let us just say this much, that the soul is the principle (arkhê) of all these activities and is determined by the following capacities: the nutritive capacity, the perceptual capacity, the intellectual capacity and movement.’11 The account of each of these capacities, as the causes of the different life activities, will be the definition of the essence of soul.

Aristotle returns towards the end of DA II.2 to the question of the relationship between the common account in DA II.1 and the definition of the soul as a cause of various life functions. He reiterates that the soul is that by which we live and so it is the cause in the sense of the form rather than the matter. He then refers back (kathaper eipomen, 414a15) to the distinction between three manners of substance, and identifies the form as one of them. He then (414a18–19) aligns the soul as form with the fulfilment of the body. In reverting in this way to the notion of the soul as a fulfilment of the body, the passage has sometimes been taken to show the priority of the DA II.1 account over the causal definitions. However, on my reading Aristotle is returning to show the compatibility of the causal definitions with the outline account. Thereby he is using the outline account as what it is supposed to do: as a check that the essential definition of the soul is of the right general kind.

The causal account of the soul is supposed to provide middle terms for demonstrations of the various life activities of living beings.12 It gives the essence of soul as the (p.41) starting point for explaining the other attributes that are necessarily true of living beings insofar as they are alive. Now if this is true we expect the causal account of the soul also to show the soul as the cause in some of the ways that are available in scientific demonstration. Aristotle explains in APo II.11 that each of the four causes, formal, final, efficient, and material cause, can play the role of an explanatory middle term in a demonstrative syllogism. His examples of each are the following:

1) Material cause:

The angle in a semicircle is half of two right angles.

The half of two right angles is a right angle.

The angle in a semicircle is a right angle.

Here the middle term ‘half of two right angles’ expresses what a right angle is composed of, that is, its matter. So the deduction happens through the material cause as the explanatory middle term.

2) Efficient cause:

The Athenians were the aggressors.

Aggressors are made war on.

Therefore the Athenians were made war on.13

Here being the aggressor is the efficient cause and works as middle term explaining why the Athenians were attacked in the Persian War.

3) Final cause:

Shelters for belongings are roofed.

Houses are shelters for belongings.

Therefore, houses are roofed.14

Note also here that the final cause of a house, sheltering belongings, seems to be the same as its formal cause: if you were to define what a house is you would say ‘a shelter for yourself and your belongings’. The sameness of final and formal cause is a non-accidental feature of natural causation, which we shall find reproduced in the case of the soul in Chapter 5.

Aristotle offers no example in II.11 of the formal cause as the explanatory middle term, but there have been plenty in the preceding discussion. To take the familiar example we had before:

(p.42) Thunder is the extinguishing of fire.

Extinguishing of fire occurs in the cloud.

Thunder occurs in the cloud.

Here ‘extinguishing of fire’ offers the ‘what it is’, the formal cause of thunder, and occurs in the explanatory role as middle term in the demonstration.

Now Aristotle said in Phys. II that nature was an explanatory principle of nature and worked as a cause in those four familiar ways. Since knowledge of nature is meant to be an instance of demonstrative knowledge, we would expect nature in those various senses to work as an explanatory middle term in demonstrations. And Aristotle confirms the expectation in APo II.11. Having offered the three kinds of causes as middle terms, he says (94b27–37) that in many cases the same phenomenon can be explained with reference to several causes, like the final and efficient cause. So thunder can be explained both because it is necessary for it to sizzle and make a noise when fire is extinguished (efficient cause) and because, according to Pythagoreans, it frightens the denizens of Tartarus (final cause). We know of course already what the formal causal explanation of thunder is, the extinction of fire in the clouds. Aristotle underlines (94b35–6) that it is especially in the case of natural processes and products that we get such explanatory diversity. The reason is obviously that nature itself works as a cause in these various ways.

Now my argument so far has been that Aristotle seeks to define the soul as the cause of the various life phenomena, and that his way of seeking such a definition is by trying to identify the account of the soul that will serve as an explanatory middle term in demonstrations of the necessary attributes of living beings qua living. Since, as Aristotle told us at the beginning of DA I.1, the soul is also a principle of animals, and I have suggested, it constitutes the nature of living beings, we would expect that soul too could possibly be understood as the cause in more than one way. We can exclude already from the common account of soul, in DA II.1, that the soul works as such as the material cause since soul is said to be form rather than matter. However, as the discussion of the nutritive soul in DA II.4 will confirm,15 the soul works as a cause in the three other ways, as formal, final, and efficient cause. We can exemplify these roles by some examples taken from the soul’s role in nutrition and generation. Many more could be found according to which aspect of nutritive behaviour we are interested in, and in which kind of living being we are considering.

Final cause:

People are immortality-lovers.

Immortality-lovers have intercourse.

Therefore people have intercourse.

(p.43) Understand here immortality-lovers as meaning lovers of the perpetuation of the eternal species form, which is the soul. Soul in this way works as final cause, like an object of desire.

Efficient cause:

What is heated up internally is concocted.

Ingested food is heated up internally.

Therefore ingested food is concocted.

Here, as Aristotle explains in DA II.4, it is the soul that regulates the internal heat, as its contributory cause, in concocting the food. The reference to soul is necessary in fully specifying the efficient cause since it is the soul that ensures the proper proportion of the heating, not too much, not too little.

Formal cause:

Large animals produce only one offspring at a time.

The elephant is a large animal.

Therefore the elephant only produces one offspring at a time.

The example is taken from GA IV.4 771a17–22 but serves our purpose since it relates the reproductive behaviour of the elephant directly to its size, a formal feature. Size is part of the nature of the elephant which is explained by the form of elephant, either directly as a part of the definition of elephant or as a further necessary consequence of its essential attributes. It is the form of elephant that ensures that it is big, and so it is the form of the elephant which explains why it only has a single calf at a time.16 Since the form of the elephant is its soul, we can say that the example shows how the soul works as a formal cause in explaining the reproductive behaviour of animals.

Now there is a debate about the ways in which, if any, Aristotle meant for his model of demonstrative understanding to be applied to the biological works. Views vary accordingly about the relationship between Aristotle’s theory of science and his scientific practice, particularly, but not exclusively, in the biological works. There are in the biological works, as Jonathan Barnes has underlined, no overtly stated demonstrative syllogisms. However, this observation leaves a range of interpretations open, including the option that Aristotle meant for the demonstrative model to be applicable to his biological accounts, and that such applicability would show the correctness of those accounts. Robert Bolton, Allan Gotthelf, and James Lennox, amongst others, building on the seminal work of David Balme, have stressed the extent to which Aristotle’s biological accounts are amenable to the demonstrative model. Others, notably Geoffrey Lloyd, have criticized this interpretation. Through a survey of Aristotle’s uses of (p.44) the term apodeixis, ‘demonstration’, Lloyd highlights the way in which apodeixis is often used for arguments that have a less strict status than required of a demonstrative syllogism in the earlier chapters of APo, I.1–4. This may in itself come as no surprise, in that many similar words used by Aristotle sometimes have a stricter and sometimes a looser sense: syllogismos, ousia, or logos, to take a few. Nor need the variety of uses of apodeixis pose a difficulty to the opposing camp, in that their claim is not that apodeixis always carries with it the notion of the demonstrative model, but rather that Aristotle’s explanatory practice and his presentation of it in the biological works are consistent with the notion of apodeixis articulated by the demonstrative model. Nonetheless, Lloyd’s point is an important reminder that we cannot assume that apodeixis in the biological works always refers to demonstration strictly understood.

Even more important, perhaps, are his claims that the other uses of apodeixis point to Aristotle himself having changed his standards of scientific argument in the biological works. He cites five passages, of which Lloyd considers the following lines from PA I ‘the most enlightening’:

But the manner of the demonstration and the necessity is other in the case of physics and the theoretical sciences. This has been discussed elsewhere. But the archê in some cases is what is, in others what will be. For since health or man is such, necessarily [so-and-so] is or comes to be: it is not that because this [so-and-so] is or comes to be, necessarily it [the product] is or will be. Nor is it possible to join the necessity of such a demonstration (tês toiautês apodeixeôs) to the eternal, so as to say, since this is, that is. I have discussed these matters in another place…(PA 639b30–640a8, transl. from Lloyd (1996))

The passage is, as Lloyd says, ‘difficult and controversial’, particularly in the construal of the first sentence, which at first glance looks like it is contrasting physics and the theoretical sciences, while Aristotle says elsewhere that physics is a theoretical science (Metaph.VI.1 1025b24–7). Lloyd argues that however we try to solve this difficulty, the passage lays out two modes of apodeixis. One is ‘a mode of apodeixis where you start from what will be, the telos—health or man—and work back to what had to be, or become, to produce that telos’. Here the necessity is hypothetical: the antecedent can be demonstrated to be necessary on the assumption that the outcome ‘health’ or ‘man’ is to come about. The other mode of demonstration operates with eternal and unconditional necessities: here the antecedent can be asserted independently of the consequent. Lloyd suggests that the former mode of demonstration does not comply with the strict standards of the APo which require that the premises be necessary independently of the consequences. PA I.1 signals, then, that the teleological explanations to be employed in biology are not strictly demonstrative in the way envisaged by APo.

Lennox has replied effectively to Lloyd’s interpretation of the passage. He points to the way in which the Phys. II.9 insists that goals are first principles in natural science, a point reworked in APo I.24 85b28–86a3 in the context of an argument to the effect that since in final causal explanation we know something best when we no longer know it because of something else, this is also true in other kinds of demonstration. (p.45) And the implication of this argument that there is demonstration using final causes as principles is made explicit, as Lennox points out, in APo II.11, the chapter we were looking at just know. So if one were to agree with Lloyd that there is a difference between the two kinds of apodeixis referred to in PA I.1from the point of view of living up to the standards of APo I.1–4, then one would have to say, equally, that there is a tension between these standards and APo II.11.

Moreover, it is not clear why the use of final causes as explanatory middle terms in nature should undermine the claim that the premises necessitate the conclusion. One might think hypothetical necessitation would fail to necessitate the conclusion for at least three reasons, taken individually or in combination: (1) because other means could be found to bring about the end, or (2) because the end does not always obtain, or (3) because the end does not yet obtain when the means happens. However, none of these is decisive in the case of natural teleology. (1) Nature generally uses the same means for the same ends: our teeth, to use the example of Phys. II.9, are generally organized in the same way to facilitate our eating. While it may then be true in practical contexts that humans sometimes find different ways of achieving the same result, so that one end does not necessitate one particular kind of means, this does not seem to hold of natural causation. (2), that the end does not always obtain, seems true only for individuals but is denied by the eternity of the species: insofar as the ends we are considering are the essences and necessary features of the species it is not true to say that these ends do not always obtain. The contrast of course still obtains with beings, the stars or the planets, which are individually eternal. As for (3), Aristotle argues in Metaph. Θ.8 1049b that while the individual means precedes the individual end, for example the acorn comes before the oak, at the level of the kind, there is always a preceding member of the kind, so before this acorn there was an oak that made it, or as Aristotle is want to say, ‘man generates man’. In nature things come into being to perpetuate ends that are already realized. At the level of natural kinds, it is misleading then to say the end cannot necessitate the coming into being of the means since the end has not yet to come into being. If we take these points into account, there seems no compelling reason why the hypothetical necessitation of the means by natural ends, taken at the level of the species,17 should not be strong enough to support demonstrative argument.

These considerations might serve as a first defence of the claim in APo II.11 that final causes can play the role of an explanatory middle term in demonstrative arguments. However, it is not my intention here to argue that Aristotle has a consistent theory of demonstration through the four causes in the APo or that he implements such a theory consistently in the biological works. Such arguments are major interpretative tasks in their own right. But insofar as the tensions that Lloyd has found between the APo and the biological works would already have to be seen as located or addressed within the APo, they do not constitute reasons for thinking that Aristotle changed his mind about (p.46) the applicability of the demonstrative model between these works.18 And nor therefore do they provide grounds for doubting that Aristotle might have wished to apply the APo’s ideas about the role of the four causes in demonstration to the psychology of the DA, given the foundational role of the DA in relation to the biological works.

This is the point I ultimately want to maintain: in the DA Aristotle uses the demonstrative model of the APo as a framework for identifying the essential features of the soul, as those features that could work as middle terms in syllogisms demonstrating the necessary life behaviours of living beings. This project is foundational for the biological works in that the essential features, so identified, could form the ultimate premises for any demonstrations of why animals have the parts they do. The DA can thus be seen as the fundamental stage in Aristotle’s zoology, a point argued for by Lloyd elsewhere with supreme persuasiveness.19

I have been concerned to stress that Aristotle’s attempt to define the soul should be seen within the context of his attempt to gain a certain kind of knowledge of the soul, namely the knowledge that is expressed in demonstrative arguments. The definition of the soul will thus provide the explanatory principle for demonstrations of the characteristic attributions of soul, and, by extension, of whatever attributes living beings have qua ensouled. I have suggested that Aristotle approaches the identification of the definition of soul from the fact of its connection with life. We are firmly within the explanatory framework of the APo. In at least two ways, this framework continues to determine Aristotle’s account of the soul in the early chapters of Book II of the DA: in the account of the manner and the ways in which the soul works as a cause for living beings and in his concern with the question of whether and if so how the soul, despite being said in many ways, may present a unitary object of definition. These are the issues we shall be looking at in the next two chapters.


(1) Charles (2000).

(2) For the distinction, cf. APo I.2 71b34–72a6.

(3) I am here broadly in agreement with Bolton (1978).

(4) For more detail, see Bolton (1978) nn. 21, 22.

(5) Bolton (1978).

(6) See also Charles (2000) 168, n. 54.

(7) Cf. Owen (1986) 242–3.

(8) Cf. LSJ, s.v.

(9) E.g. Iliad IVX. 518. For further evidence and discussion, see Bremmer (1987) Chapter 3.

(10) This claim can be made without implying that the soul itself is alive irrespective of or prior to a body, just as the claim about noise can be made without implying that there is not something in which the fire is being extinguished when noise is produced. Rather the point of the syllogism is to articulate the way in which soul is responsible for the living of the body in which it is.

(11) νῦν δ̓ ἐπὶ τοσοῦτον εἰρήσθω μόνον, ὅτι ἐστὶν ἡ Ψυχὴ τῶν εἰρημένων τούτων ἀρχὴ καὶ τούτοις ὥρισται, θρεπτικῶ̣, αἰσθητικῶ̣, διανοητικῶ̣, κινήσει‎.

(12) While Kent Sprague (1996) is right in seeking to use the notion of the soul as a cause of life to construct a demonstrative syllogism, it is unclear to me why she seeks to find the preferred definition in the conclusion rather than in the premises. The root of the problems she has in reconstructing the demonstrative syllogisms for the soul may be the claim that because of the example of squaring ‘we have to assume that “an actuality or formulable essence of something that possesses a potentiality of being besouled” (414a27) qualifies as a definition that is not simply the conclusion of a syllogism but that somehow exhibits the ground’ (105). But it is quite clear that the explanatory middle term in the example of squaring should be ‘finding the mean proportional’ and that it is this term that will correspond to the essence of soul. On the reading I am suggesting ‘an actuality or formulable essence of something that possesses a potentiality of being besouled’ is not Aristotle’s preferred definition of the essence of soul, nor is it supposed to figure in the demonstrative syllogism, rather it works, in accordance with the outline account of the soul in DA II.1, as a description of the causal role that the essential definitions will play.

(13) I have changed Aristotle’s language slightly for ease of presentation.

(14) I avoid the problematic example of digestion, in favour of the simpler example Aristotle offers, as developed by Barnes (1993) 231. Leunissen (2007) 167 seeks to avoid the problem by arguing that Aristotle does use final causes as explanatory middle terms but only as predicates in the conclusion. There is on this account then no demonstration through the final cause, but only of the final cause. This seems to throw the baby out with the bathwater: the interpretation not only undermines the primacy of final causation in nature as an explanatory principle, it also makes it very hard to see how the formal cause works as an explanatory middle term in those cases, such as a house, where it is the same as the final cause. Leunissen (2007) 170–1 writes of the subsumption of the final cause under the formal cause, but it is more characteristic of natural contexts, as we shall see, to say that the formal cause is explicated in terms of the final cause.

(15) See Chapter 6.

(16) Aristotle explains further that ‘their size is the very reason why they do not produce many offspring, because in animals of this sort the nourishment gets used up to supply the growth of the body, whereas in the case of smaller animals, nature takes away from their size and adds the surplus on to the seminal residue’ (771a28–31, transl. Peck in Loeb). The fact that the amount of matter available for semen is determined by the available nourishment given the size of the animal again confirms the formal character of the explanation.

(17) Since demonstration expresses knowledge, and there is no knowledge of the individual as such, it should not surprise us that the demonstration has to be pitched in this way at the level of the species.

(18) There is a tendency in Lloyd’s comments to harden the criteria for demonstration in the APo so that tensions arise even within that work. So, for example, he uses mathematical demonstrations in the APo, which enjoy the strictest form of demonstrative necessity, as a starting for comparisons with Aristotle’s arguments in the biological works; see Lloyd (1990) 32–3. But this makes it harder to see how all the four causes could as recommended by APo II.11 work as explanatory middle terms, given that there are no final or efficient causes in mathematical reasoning. Since these causes necessitate their outcomes in different ways (as we have seen), we cannot expect mathematical style necessitation to apply in these explanations. But it is the case that the APo positively recommends that all four causes can work as explanatory principles. So if the use of these explanatory principles is in tension with the mathematical model it is a tension already within the APo, not between the APo as such and the biological works.

(19) In Lloyd (1992).