1. Varieties of Aristotle's methods of ethics

Controversy continues as to what method Aristotle uses in his ethical enquiry. In the recent literature, it is widely agreed that dialectic does not play a central role in defining the highest good in ethics, i.e. the concept of happiness (εὐδαιμονία). Certainly, Aristotle applies what is called the ‘endoxic method’ (as formulated in Nicomachean ethics [Eth. Nic.] 7.1 (or Eudemian ethics [Eth. Eud.] 6.1)Footnote ¹ 1145b2–7) when he examines the problems of the weakness of will (ἀκρασία). The endoxic method sets out people's common beliefs or appearances, goes through puzzles and attempts to prove the common beliefs by resolving the puzzles.Footnote ² This endoxic method is, however, not identical with dialectic. As Brunschwig (Reference Brunschwig1967) xxii–xxiv, Devereux (Reference Devereux, Henry and Nielsen2015) 131–4 and Fink (Reference Fink2012) 2 put it, dialectic is the art of rule-governed debates between a questioner and an answerer in public. It is directed at someone else and concerned with opinion, not necessarily with the truth (Topics [Top.] 1.1 100a18–b23; 1.14 105b30–31; 8.1 155b7–16). To be sure, Aristotle considers common opinions in his ethical enquiry, but he does so only when he compares his own outlook with other people's ideas, such as in Eth. Nic. 1.8, after he has formulated his view of happiness. Thus, dialectic is not key to characterising Aristotle's method for formulating ethical concepts.Footnote ³

Recent studies draw attention to the philosophical, scientific aspects of Aristotle's method of ethics and claim that his scientific method is key to reaching the truth (Karbowski (Reference Karbowski2015, Reference Karbowski2019), Natali (Reference Natali2010), Nielsen (Reference Nielsen, Henry and Nielsen2015)). They refer to Aristotle's Posterior analytics [An. post.] as a basis for his view of science, treat his method of ethics as aimed at a demonstrative explanation, not a dialectical explanation, and explore how he uses demonstrative arguments for formulating the first principle of ethics – the concept of happiness. However, a demonstrative argument per se cannot prove its first principle; rather, it lays down the principle as an assumption from which it explores its subject. How to define the concept of happiness is a major issue of the Aristotelian methodology.

In what follows, I will examine the Eth. Eud. method first, because in the light of this version it is possible to illuminate the characteristics of the Eth. Nic. method more clearly. In particular, I will draw attention to the less-known method, expressed in Eth. Eud. 1.6 1216b26–35, where Aristotle says, ‘If we start from what is truly yet not clearly spoken (ἐκ γὰρ τῶν ἀληθῶς μὲν λεγομένων οὐ σαφῶς δέ), clarity will be won as we make progress’. This expression comes from Plato's Statesman 275a and 281d.Footnote ⁴ Aristotle then seems to be influenced by the Eleatic stranger's methodFootnote ⁵ in clarifying the initial unclear statement that happiness is the greatest and best of human goods. This influence of Plato's Statesman on Aristotle's Eudemian methodological passages seems to me to illuminate the method for defining the concept of happiness in Aristotle's two ethical treatises. Daniel Devereux (Reference Devereux, Henry and Nielsen2015) and Joseph Karbowski (Reference Karbowski2015, Reference Karbowski2019) examine the Eth. Eud. passage (1.6 1216b26–35) in detail, but do not draw attention to this influence.Footnote ⁶ Although Karbowski (Reference Karbowski2019) 51 maintains that ‘Aristotle's [negative] attitude towards dialectic constitutes a major break from his Platonic legacy’, I think that the influence of the later Platonic method on Aristotle deserves more attention in the literature on ancient philosophical methodology.

In section 2, I will explore what method Aristotle uses in Eth. Eud. to define the concept of human happiness. In particular, I will illustrate how Aristotle begins with the true yet unclear statement and clarifies it by distinguishing the target concept from other similar concepts and using examples as illustrations. In section 3, I will argue that this method is influenced by the Eleatic stranger's arguments offered in Plato's Statesman. I will also explore how the stranger and Aristotle differ in their formulations of argument. In section 4, I will explore the similarities and differences between Eth. Eud. and Eth. Nic. and show how Eth. Nic. also reflects the Eth. Eud. method in some way to define the concept of human happiness, even though Eth. Nic. does not make explicit the Eudemian method in its methodological passages. Overall, the current article seeks to illustrate the Platonic background for Aristotle's ethical enquiry. The appeal to a Platonic antecedent will illuminate the characteristics of Aristotle's method from a new angle.

2. The Eudemian ethics method

For our consideration of the method that Aristotle adopts to define the concept of human happiness in Eth. Eud., the following passage is of special importance:

There are three things to note here. First, it is not specified what the term ‘appearances’ (φαινομένοις) means in (1). The term could mean either sense-perceptions or opinions that people form from their observations.Footnote ⁸ The quoted text in (3) seems to suggest that everyone has some cognitive faculty for perceiving the truth, and that their ethical enquiry relies on this faculty. At b31, ‘from these’ (ἐξ ὧν) does not have an antecedent in a plural form but presumably picks up οἰκεῖόν τι πρὸς τὴν ἀλήθειαν (lit. ‘something akin to the truth’), signifying appearances that people have collectively. The ‘for’ (γὰρ) in (3) implies that this manner of enquiry helps to reach an agreement with others, as stated in (2), and the ‘for’ (γὰρ) in (2) suggests that this agreement is a form of seeking conviction through argument by using appearances. There is a need to examine what appearances one is going to use as witnesses and examples and how.

Second, in (2) Aristotle considers it second best that everyone will agree after a change of mind. What is unclear is how people change their mind. The term (μεταβιβάζω) has a Platonic background. In Phaedrus 262b, Socrates considers that one cannot lead others little by little through similarities away from what is the case on each occasion to its opposite without knowing what each thing is. In its wider context, Socrates criticises rhetorical art as a way of directing the soul by means of speech (ψυχαγωγία τις διὰ λόγων) (261a).Footnote ⁹ Aristotle also acknowledges the significance of understanding what the case is, as expressed in (3) and (4). His remark in (3) and (4) resists the argument that is not concerned with the truth. How then does Aristotle change the course of his argument to clarify the concept of happiness?

Third, it is also unclear what Aristotle means by ‘what is truly yet not clearly spoken’ in (4). Clarity (τὸ σαφῶς) seems to be contrasted with what is usually spoken of indiscriminately (συγκεχυμένως). The phrase οὐ σαφῶς at 1216b33 is translated into ‘not clearly’ (Inwood and Woolf (Reference Inwood and Woolf2013), Woods (Reference Woods1992)), ‘unenlightening’ (Kenny (Reference Kenny2011)) or ‘obscurely’ (Solomon (Reference Solomon and Barnes1984)). When something is obscure or vague, however, it is difficult for us to tell whether it is true or not, though Aristotle here states that the initial views are expressed truly and unclearly. Furthermore, he is going to clarify the initial views by introducing what is more known (τὰ γνωριμώτερα). This is a typical Aristotelian term by which he means either what is more known to us or what is more known by nature. The text does not specify which. Therefore, there is need to explain the meaning of clarity (τὸ σαφῶς) and the method of making clear by introducing what is more known.

In order to address these problems, it is necessary to explore how Aristotle presents the argument that follows. In Eth. Eud. 1.7, he adopts this method, referring to the methodological remark as a prologue (1217a18–21). The starting point of enquiry that is spoken of correctly yet unclearly is people's view that happiness is the greatest and best of human goods (Eth. Eud. 1.7 1217a21–22). To clarify this view, Aristotle first notes that his argument is concerned about ‘human goods’, though there might be happiness for some other being such as god (1217a22–24). Next, he draws two distinctions in Eth. Eud. 1.7. First, good things are achievable by human action or not; some things are not performed by humans because some things are not susceptible to change though they might be by nature the best of all things, such as order in the universe. Second, the things achievable by human action are the things for the sake of which we act – that is, the purposes of actions – or the things we do for the sake of these purposes. He concludes that happiness is the best of the things achievable by human action (1217a39–40).

He develops this view in Eth. Eud. 2.1 (in Eth. Eud. 1.8 he criticises Plato's form of the Good and does not advance his own view). At the very beginning, he lays down three assumptions: that good things are either within a soul or external to it and the former is better than the latter; that among things in a soul, some are states or capacities, while others are activities and processes; that virtue is the best disposition, state or capacity of any thing that has some use or function. As to this third assumption, he verifies it by induction, using the examples of a cloak, a boat and a house. Subsequently, he goes on to argue that the function of a thing is better than its state, because the function is a purpose, which is the best for the sake of which other things exist. Furthermore, he identifies the function of a soul with providing life, and that of virtue with an excellent life. From these considerations he concludes that happiness is the activity of the good soul. The reason is that happiness is the best thing, which lies in an activity or function, not a state, and within a soul, not external to it. He adds that happiness is a complete thing, and for this reason it is the activity of a complete life, not a short life, in accordance with complete virtue (Eth. Eud. 2.1 1218b31–1219a39).Footnote ¹⁰

To sum up, Aristotle draws several distinctions and leads the readers or audience in the direction of an account developed from these distinctions to define the concept of human happiness. Although the initial view was indistinct, it becomes clearer that happiness is the activity of a complete life in accordance with complete virtue. The initial view might be mistakenly interpreted as the claim that happiness is the state of a human soul in a short life, but after being elaborated by argument, it means specifically that happiness is identical with exercising complete virtue in a complete life. The initial statement is not wrong. It is true that happiness is the greatest and best of human goods (Eth. Eud. 1.7 1217a21–22). However, it is too imprecise to offer useful insights for sharing the view of happiness. After reflection, the initial view becomes more precise and helpful in defining the meaning of happiness.Footnote ¹¹

Hence, the term μεταβιβάζω does not necessarily mean to change minds radically, but rather to lead the argument in one specific direction and clarify the initial imprecise assumption. Aristotle uses this term (μεταβιβάζω) in Top. 1.2 to mean ‘changing the course of an argument’ in a conversation with ordinary people (101a33). In its wider context, he points out three manners of the use of dialectical studies: exercise, encounters and philosophical sciences. As to the second manner, Aristotle says that a dialectician will speak to the public ‘not from the beliefs of others, but from their own beliefs, changing their minds about anything they may seem to us not to have stated well’ (101a31–34, trans. Smith (Reference Smith1997)).Footnote ¹² Furthermore, in Top. 8.11, this term expresses the appropriate manner of dialectical persuasion. Aristotle distinguishes between dialectical and eristic discussion in that the former is a cooperative work whereas the latter is abusive. A dialectician can persuade his interlocutor by offering an argument stemming from what the interlocutor believes is the case. Aristotle supports this manner of persuasion in saying that ‘anyone who is to change minds well must change them dialectically, not contentiously’ (161a33–34, trans. Smith (Reference Smith1997)). Therefore, this term implies that the method aims to persuade the public with reference to what they believe to be the case.

Eth. Eud. has another argument that relies on the same method. At the end of Eth. Eud. 2.1 (1220a15–18), he again states that we must investigate ethical virtue ‘so as to try, always by working through what is expressed truly but unclearly, to arrive at what is both true and clear’ (ὥστε ἀεὶ διὰ τῶν ἀληθῶς μὲν λεγομένων οὐ σαφῶς δὲ πειρᾶσθαι λαβεῖν καὶ τὸ ἀληθῶς καὶ σαφῶς). As examples of true yet unclear views, he states that health is the best disposition of the body and that Coriscus is the darkest-skinned fellow in the marketplace. These statements do not yet express the features of each in detail but offer an initial clue to having a deeper understanding of each.Footnote ¹³

The subsequent argument is an application of this method: first, he lays down the true but unclear assumption that the best disposition is brought about by the best things, and that the best actions for each thing result from the virtues of each. These assumptions are clarified by induction from examples such as exercise and food; a good physical condition is brought about by the best exercise and food. He applies this general notion to virtue and asserts that virtue is a disposition of this kind, which is brought about by the best movements of the soul. Here the initial assumption is clarified, not by leading the argument in one specific way, but rather by way of examples. What appears to us is used to understand the initial assumption better.

In the aforementioned argument, an inductive argument from more intelligible examples is offered at the beginning of Eth. Eud. 2.1. There, the assumption that virtue is the best disposition, state or capacity of each of the things that have some use or function is verified by induction (1218b37–a5); a cloak has a virtue or excellence since it has a function or use. The same applies to a boat and a house, and hence to the soul. Here, as elsewhere, the general assumption is true but unclear at first, but it is made clearer with the help of familiar examples. This approach corresponds to the methodological passage (Eth. Eud. 1.6 1216b26–35), where Aristotle says that we should use ‘appearances’ as witnesses and examples. Therefore, ‘appearances’ are what are intelligible to people that help them to agree with more abstract, general assumptions.Footnote ¹⁴

At the end of Eth. Eud. 1.6 (1217a10–17), Aristotle again warns us that we should pay attention to appearances as well as to the points made by an argument. It is a good idea to distinguish between inferences and conclusions, because unless we regard the appearances, we have no choice but to accept what is being inferred. Here, the appearances are expected to play a vital role not only in clarifying an abstract concept of ethical values, but also in doubting what is asserted in discussion and offering an alternative perspective, even if we cannot refute the argument. Appearances are not the starting points or assumptions from which ethical beliefs are demonstrated, but are rather evidence or examples in the light of which we can examine the soundness of an assumption or piece of reasoning.Footnote ¹⁵

In (4) of the methodological passage, Aristotle maintains that we must substitute ‘what is more known (τὰ γνωριμώτερα) for what is usually spoken of indiscriminately (συγκεχυμένως)’ (Eth. Eud. 1.6 1216b33–35). Physics [Ph.] 1.1 is very helpful in clarifying these concepts. He claims that we must ‘advance from what is more obscure by nature, but clearer to us, towards what is clearer and more knowable by nature. Now what is to us plain and clear at first is rather confused masses (συγκεχυμένα), whose elements and principles become known to us later by analysis’ (184a19–23). In this context, he claims that in physics we aim to know the principles, causes or elements of natural objects. To this end, we should begin from what is known to us, i.e. a whole that is more perceptible (184a24–25), and analyse it in terms of its elements. For example, the circles we perceive are indistinct, including their variations. By way of analysis, a circle can be defined as a plane figure with all points on its circumference equidistant from a given point. This analysis of its defining elements serves to distinguish a complete circle from similar ones. Aristotle raises another example: children first call all men ‘father’ and all women ‘mother’ indiscriminately. Subsequently, they come to recognise the distinctions of father, grandfather, uncle etc. The children's initial understanding of ‘father’ is unhelpful in specifying who, but drawing distinctions serves to identify whom they call.Footnote ¹⁶

There are, however, differences between Ph. 1.1 and Eth. Eud. 1.6: in the former the confused masses are perceived objects while in the latter what is confused is the general statement about which common people also agree. Furthermore, in the former the confused masses will be defined in terms of their elements or causes, whereas in the latter the confused statement will be replaced by the more precise formulation. In the Eth. Eud. methodological passage, therefore, what is more known (τὰ γνωριμώτερα) is presumably what is more known without qualifications and not what is known to us. In the clarification of happiness, the initial uninformed view that happiness is the greatest and best of human goods is elaborated into its more precise statement.Footnote ¹⁷

Hence, the Eth. Eud. approach (i.e. beginning from a true but unclear statement) is not absolutely identical with the standard version of the Aristotelian method that we should start from what is known to us in order to recognise what is known by nature without qualifications (cf. Metaph. Z.3 1129b1–12).Footnote ¹⁸ Usually, what is known to us is a familiar example or an evident case. The initial unclear statement is, however, not what is known by our perception in this sense, but rather an abstract statement about which common people agree. This abstract statement is clarified during the process of the argument, with reference to clearer examples. Thus, the standard version of the Aristotelian doctrine is incorporated within this approach.

To sum up, in Eth. Eud., Aristotle takes two main approaches to the clarification of ethical values. First, he refers to appearances or examples to clarify the abstract concepts of ethical values. Second, he clarifies ethical concepts by drawing distinctions and leading the argument in the direction of one specific position. The initial true but unclear view is developed by these two approaches, so that it becomes a more precise account verified by more recognisable examples. Aristotle claims that one must prove (δεικνύναι at 1216b32) the initial assumption by taking these two approaches. Here ‘proof’ is not deductive demonstration from first principles. Rather, it is either an inductive inference from clear examples to the general statement or making the initial statement more precise by leading the argument in one specific direction.Footnote ¹⁹

3. The Platonic background for Aristotle's method

This section will show that the Eleatic stranger's method in Plato's Statesman offers an important background for Aristotle's Eudemian method.Footnote ²⁰ Plato's Statesman influences Aristotle enormously, as it was written in the mid-360s (Wallach (Reference Wallach2001) 335), when Aristotle began studying in the Academy. On the one hand, it is well known that Aristotle criticises Plato's Statesman in many respects. In the beginning of the Politics [Pol.], he doubts that the positions of statesman, king, household manager and master of slaves are the same (Pol. 1.1 1252a7–16). In its book 1, he explains how they differ by dividing a polis into its constituent parts and distinguishing a polis from a household.Footnote ²¹ Although he derives his six-fold scheme of constitutions from Plato's Statesman, the standards for distinguishing between correct and deviant constitutions are not identical.Footnote ²² More importantly, Aristotle criticises the Platonic method of division in Prior analytics [An. pr.] 1.31, An. post. 2.5 and On the parts of animals [Part. an.] 1.2 and 3. It is, however, well known that Aristotle uses the method of division to define a target object. Deslauriers (Reference Deslauriers1990) 208 claims that ‘Plato's aim was not primarily to classify but to define […]. Aristotle indicates the same concern when he maintains that a definition proper must express what the thing defined really is’.Footnote ²³

Our task lies in the examination of how Aristotle incorporates the stranger's method in his enquiry for defining ethical concepts. By the stranger's method, I mean the enquiry beginning from a true yet unclear statement and aiming to define a target object by distinguishing it from other similar objects and using illustrations. Aristotle uses this stranger's method not for the classification of various objects or for the deductive demonstration of a statement but for the definition of a concept. It is an age-old question whether it is more fruitful to stress that Aristotle is an anti-Platonist or an independent Platonist. It seems better to me to emphasise, rather than to understate, Aristotle's Platonic background, at least for his method of defining ethical concepts.Footnote ²⁴

Aristotle adopts the Eleatic stranger's method when he seeks to define ethical concepts such as happiness and virtue. The key phrase to analysing this reception lies in (4) of the methodological passage (Eth. Eud. 1.6 1216b26–35), where Aristotle suggests that we should begin from what is truly yet not clearly spoken. As mentioned in the abstract that begins this article, this expression comes from Plato's Statesman 275a and 281d. In 275a, the Eleatic stranger is undertaking a task in defining a statesman's rule and considers it ‘true but incomplete and unclear’ (ἀληθές, οὐ μὴν ὅλον γε οὐδὲ σαφὲς) that a statesman is one who rules a state as a whole. He recognises that he failed to explain how a statesman rulesFootnote ²⁵ and goes on to elaborate this view, first, by treating the art of a statesman in a broader category – some sort of expertise in ‘herd-keeping’, ‘looking-after’ or ‘caring for’ (275e). He prefers ‘caring for’ to ‘rearing’ as its title (276c–d). Second, he distinguishes between divine herdsman and human carer. He also divides the art of human carer into management by constraint and that by consent (276d). This division is necessary for distinguishing kingship from tyranny because their manners of ruling are totally different. Thus, he defines the art of statesman or king as ‘the herd-keeping that is voluntary and relates to two-footed living things’ (276e).

The Eleatic stranger is, however, not fully satisfied with this definition. He says, ‘our account, just like a portrait, seems adequate in its superficial outline, but not yet to have received its proper clarity’ (277b–c). This statement also implies that the earlier view serves as an outline for enquiry, though it is not yet perfect. The stranger first draws a rough sketch for the target object. Next, he seeks to give the accurate picture of it by drawing several distinctions. The initial account is true in outline but too indistinct or unilluminating to reach the target.

In the context of 281d, the stranger seeks to define the art of weaving, considering his initial statement to be ‘something true, but not clear or complete’ (τι ἀληθές, οὐ μὴν σαφές γε οὐδὲ τέλεον) that weaving is the finest and most important process in the production of woolen clothing. He goes on to elaborate this view, first by distinguishing between a contributory cause and a direct cause (281d). The expertise relating to the former does not make the thing itself but provides tools. As to wool-making, contributory causes are the skills of making spindles and shuttles. The expertise relating to a direct cause is the skill that brings the product to completion. The skills of carding and spinning are this type of expertise in wool-making and the stranger gives them the title ‘wool-working’ (282a). Next, he divides the skill of wool-working into two: the expertise of separation and that of combination. He treats the skill of carding as the former. As to the expertise of combination, he distinguishes between the skill of twisting and that of intertwining. The skill of twisting is the manufacturing of firm warp and soft woof. Finally, he defines the art of weaving as the skill of intertwining warp and woof to produce woolen clothes (283a).Footnote ²⁶

In Plato's Statesman, thus, the stranger elaborates the initial true but unclear view by drawing many distinctions and advancing the argument in the direction of an account given by the introduction of distinctions. The stranger asserts that he and young Socrates will not have a clear or complete view until they remove all of the skills relating to woolen clothing from the art of weaving (281d). The initial view is true in including the target object, but it is too broad to provide useful insights into a statesman's rule. The method of articulating a true and clear view is dividing the closely related types and offering a precise account of the target object. When the stranger eliminates all the similar but unnecessary objects from the target, he offers its clear and complete account. By ‘clarity’, thus, the stranger means ‘precision’.

The Eleatic stranger does not use his method for classifying various phenomena, but for specifying the target object. He compares this use to gold-refining (303d–e). Gold refiners need to remove earth, stones and plenty of other materials. They also separate from pure gold something akin to it – copper, silver and adamant. In the similar manner, the stranger has distinguished sophists and demagogues from a genuine statesman. He also separates military art, judicial judgement and political rhetoric from statesmanship. The method of division is primarily used for removing unnecessary elements and identifying the target object of statesmanship (262a–b, 268c, 291c, 292d). The stranger compares himself with hunters or trackers (263b, 264a, 285d).Footnote ²⁷ Aristotle receives this manner of using the method of division for identifying a target object. He is a good student of Plato, as he adopts the method of division not for classification but for the definition of a concept that is initially expressed in an imprecise statement. Given that Aristotle is greatly influenced by the later Platonic works, it is likely that his methodological ideas are also shaped by this Platonic method of division to define the concept of human happiness.

Furthermore, in the Statesman, the task of defining the art of weaving is a part of a broader project of defining statesman-rule; the art of weaving is used as an analogy for statesman-rule. Before introducing the art of weaving, the stranger explains the details of how this illustration works; in 277d, he first says, ‘It's a hard thing, my fine friend, to demonstrate any of the more important subjects without using models’ (μὴ παραδείγμασι χρώμενον). By offering the example of letter-reading, he describes the point of using models as comparing something they know to something else – that they have not yet recognised (278a–c).Footnote ²⁸ He is going to gain expert knowledge of the statesman-rule or kingship through the use of an analogy, i.e. by examining less significant cases and transferring our attention from them to the most important issue, kingship (278e). The term παράδειγμα is the same as that in the Eudemian methodological passage, where the appearances are intended to be used as illustrations. This Platonic background also confirms that Aristotle is referring to ‘appearances’ as something familiar, and through them illuminates something else that we do not yet have a clear notion of: the concept of human happiness.Footnote ²⁹

The Eleatic stranger uses the term παράδειγμα not to mean one particular example but to refer to a general concept that is formulated after examining many examples (278b, 278c). In the stranger's method, a model is needed to argue about what can be recognised not by our perception, but only through our language. In the context of 277e–278c, what can be recognised by our perception is construed as becoming a model only after it is examined in many different contexts.Footnote ³⁰

The stranger explains this characteristic of a model by using the example of reading letters in the syllables. Children first learn letters in the easiest and shortest syllables and next try to identify the same letters in other syllables. If they fail to recognise the letters, they are required to compare what they can recognise with what they cannot, until they can identify the letters in all syllables. At this point, they are creating a model by distinguishing one letter from different letters in all syllables. The stranger sums up the epistemological role of a model by saying that it issues a single judgement about the same thing in different objects (277e–278c). What is easily recognised by our perception in one context is not identical with a model. To be a model, it must be examined in other contexts so that we can judge a case in any context. Soon afterwards, the stranger clearly says that one should not begin from false beliefs, even to get to a small part of the truth (278e). He suggests that we should examine a model carefully before using it to clarify a more important concept.

Aristotle technically uses the term παράδειγμα to mean an inductive or analogical inference from evident cases. In his Rhetoric [Rh.] 1.2, he compares παράδειγμα (which is usually translated as ‘example’) in rhetoric with induction in dialectic: ‘when we base the proof of a proposition on a number of similar cases, this is induction in dialectic and example in rhetoric’ (1356b14–16). This inductive inference is contrasted with deduction in dialectic and enthymeme in rhetoric. The latter is a proof that, when some propositions are true, a further distinct proposition must also be true universally or for the most part (1356b16–18). When it comes to example in rhetoric, Aristotle stresses that rhetoricians make inferences mainly from one to another similar case, rather than from similar cases to a general proposition. For instance, he says that ‘Dionysius, in asking as he does for a bodyguard, is scheming to make himself a despot. For in the past Peisistratus kept asking for a bodyguard in order to carry out such a scheme, and did make himself a despot as soon as he got it; and so did Theagenes at Megara’ (1357b30–33; cf. An. pr. 2.24 68b41–69a13). Although this inference implies the general principle that ‘a man who asks for a bodyguard is scheming to make himself a despot’, in Aristotle's ‘example’ a rhetorician aims to reach a conclusion about a particular case.Footnote ³¹

Aristotle points out that when one statement is more familiar than the other, the former is an example (1357b29–30). In An. pr. 2.24 (69a13–16), he also stresses that an example infers from one to another similar case and that the former case is familiar. In Eth. Eud. 2.1 1218b37–1219a6, as examined in the previous section, he refers to a cloak, a boat and a house as familiar cases to explain virtue in a soul. In 1220a22–37, he considers health as a model for explaining that any disposition is brought into being and destroyed by the same things. Here too, he draws a conclusion about a further distinct case – virtue – though he formulates the general principle.

Of course, the Eleatic stranger's method is not identical with Aristotle's. First, in the stranger's arguments, the starting point for division is not what people generally agree about, but rather what he himself establishes. In Statesman 258b, he sets out the assumption that a statesman is a person who has knowledge, and goes on to divide the types of knowledge to define the characteristics of statesmanship. To be sure, he asks whether the young Socrates agrees with this assumption, but he does not treat this starting point as what seems to people at large to be true. In the context of 275b, the initial statement that a statesman is one who rules a city is not an agreed view, but rather a result of his own division. In contrast, Aristotle's starting point for ethical enquiry is people's view that happiness is the greatest and best of human goods (Eth. Eud. 1.7 1217a21–22). He begins from this common view in the hope that he can offer a more precise account of the target concept, happiness.

Second, in the stranger's framework, the relation between the two approaches – drawing distinctions and using a model – is clear; a model is needed to separate the target from other things. For example, the analogy of weaving is used to correctly distinguish the art of statesmanship from other arts. Just before introducing the art of weaving, he says, ‘since tens of thousands of people dispute the role of caring for cities with the kingly class, what we have to do is to separate all these off and leave the king on his own; and it was just for this purpose that we said we needed a model’ (279a). Creating a model serves to facilitate the process of separation. In contrast, Aristotle does not maintain that he refers to appearances to draw a wide variety of distinctions and change the course of an argument. Instead, appearances are used for defining an ethical concept, such as a function or disposition, from which he infers other ethical concepts.

To sum up, Plato's Statesman offers an important background for Aristotle's Eudemian method, though they differ in some respects. The Eleatic stranger begins arguing from a true yet unclear statement. He uses the method of division with the help of a model to clarify the initial statement. He can enhance clarity by making the initial statement more precise and explaining it from more evident cases. Aristotle uses this method to define the concept of happiness in the Eudemian ethics. Although Lane (Reference Lane1998) 94–7 draws a very sharp contrast between the stranger's method and Aristotle's rhetorical example, Aristotle adopts the former's philosophical method in the central topic of ethical enquiry.Footnote ³² Certainly, he criticises the Platonic method of division in that it is useless for demonstrating a definition and classifying various species of creatures (An. pr. 1.31, An. post. 2.5 and Part. an. 1.2–3), but he does not discourage us from using it for the definition of a target concept. In fact, he does adopt it to define the concept of happiness by distinguishing the target object from its similar objects and using familiar examples. Aristotle incorporates the following features from the stranger's method: (1) beginning from a true yet unclear statement; (2) separating the target object from its similar objects and (3) using examples to explain an abstract concept. Aristotle was Plato's student for two decades and underwent philosophical training in the Academy. Again, it is more fruitful to stress, rather than neglect, the fact that Aristotle learned from Plato's Statesman that an imprecise concept can be clarified by leading the argument in the direction of one specific position and referring to familiar examples.

4. The Nicomachean ethics method

In the light of Eth. Eud., Eth. Nic. also can be interpreted as using the Eth. Eud. method to some extent. To clarify this point, this section will examine the following three elements: first, the Eth. Nic. methodological passage; second, the Eudemian elements used in defining the concept of human happiness in Eth. Nic.; and third, methodological differences between Eth. Nic. and Eth. Eud. It is beyond the scope of this article to consider every ethical topic and conclude by determining which treatise was written earlier or later in Aristotle's career. The present article only aims to clarify the methodological relationship between the two versions.

First of all, the following well-known methodological passage is indispensable:

This passage does not express the Eleatic stranger's method, but rather follows the standard Aristotelian doctrine that the starting points of enquiry are things known – more specifically, things known to us, not without qualification. As discussed previously, what is known to us is defined as what is nearer to perception. In Eth. Nic., then, Aristotle suggests that we should start arguing from what is more accessible to our perceptive faculty to define abstract concepts of values.

Aristotle raises the same point in Eth. Nic. 2.2; he notes that we must make use of evident cases as witnesses (τοῖς φανεροῖς μαρτυρίοις χρῆσθαι (1104a13–14, cf. 1104a29–30)) for what is not evident, and clarify the concept of virtue as a mean by analogy with health and physical strength. He finds in health and strength the mean between excess and deficiency as an essential element for preserving a good condition, and finds the same characteristic in virtue as well, on account of the similarity between virtue as a disposition in a soul and health as a disposition in a body (see Eth. Eud. 2.3 for exploring the idea of virtue as a mean, and Eth. Eud. 2.1 relating to the fact that virtue is produced and destroyed by the same things, namely pleasure and pain). Therefore, in Eth. Nic. as well as in Eth. Eud., something that is empirically clear is used as evidence or illustrations.

Although these Nicomachean methodological passages do not state the use of the Eudemian method explicitly, the structure of the argument shows that it is similar to the Eudemian method. First, in the same way as Eth. Eud., the Eth. Nic. argument starts its enquiry about the highest good by seeking to clarify people's unclear beliefs. At the very beginning of Eth. Nic. 1.4, Aristotle says that as far as its name goes, most people agree that the highest good is happiness, though they do not agree about what happiness is. Moreover, soon before engaging in the function argument in Eth. Nic. 1.7, he says that it is clearly agreed that happiness is the best good, but we still need a clearer statement of what the best good is (1.7 1097b22–24). Although he has already examined this view by considering the characteristics of the highest good in terms of finality and self-sufficiency, he offers the function argument to clarify people's unclear view. If Eth. Nic. was written later than Eth. Eud., the Eth. Eud. framework is still used in Eth. Nic.Footnote ³³

Second, the terms used in the function argument show that this argument refers to what is more known to us as illustrations. Aristotle writes, ‘just as the good, i.e., doing well, for a flautist, a sculptor, and every craftsman, and, in general, for whatever has a function and action, seems to depend on its function, the same would seem to be true for a human being, if a human being has some function’ (Eth. Nic. 1.7 1097b25–28).Footnote ³⁴ He also argues, ‘just as eye, hand, foot, and, in general, every part apparently has its function, should we likewise assume that a human being has some function apart from all of these?’ (Eth. Nic. 1.7 1097b30–33).Footnote ³⁵ As to the functions of craftspersons and parts of the human body, ‘seems’ (δοκεῖ) (1097b27) and ‘apparently’ (φαίνεται) (1097b31) imply that it is known to us that the function expresses the good of a skill, and that some function exists in each part of the human body. Furthermore, the optatives ‘would seem' (δόξειεν) (1097b27) and ‘should assume’ (θείη) (1097b32) in the apodoses imply that the apodoses are not beliefs that people actually hold, but rather the consequences of inference from the protases. The apodoses are expected to work out an agreement, on the assumption that the explanation about skill or parts can be applied to the explanation about human virtue or humanity as a whole.

In Eth. Nic., therefore, Aristotle's arguments for defining the concept of happiness include the Eudemian elements, first because the argument clarifies the initial unclear view that happiness is the highest good, and second because the argument defines an abstract concept according to what seems to people to be true. The standard Aristotelian doctrine does not contradict these elements, because what seems to people to be true is also used to illustrate an abstract concept. The initial true but unclear view only offers a framework for enquiry, within which the method of the standard doctrine can be used. The concept of happiness is made more precise from the agreed, unclear statement in Eth. Nic.

In Eth. Nic. 6.1, there is an additional example of Aristotle using the same approach as in Eth. Eud. At the very start (1138b25–26), he states, ‘to say this [i.e. that the mean states are in accordance with correct reason], though true, is not clear at all (ἔστι δὲ τὸ μὲν εἰπεῖν οὕτως ἀληθὲς μέν, οὐθὲν δὲ σαφές)’. He then considers it necessary to clarify what correct reason is, and treats this problem as the task of book 6. Subsequently, he develops this argument by distinguishing various types of intellectual virtue, and at the end of book 6, he defines the concept of correct reason as practical wisdom (6.13 1144b27–28). In book 6, the initial unclear statement offers a basic framework for this enquiry, demonstrating that there is a need to clarify various types of intellectual virtue. In this approach, he leads the readers or audience in the direction of the distinctive characteristic of practical wisdom, which is separated from other intellectual virtues.

This is a piece of evidence in support of the view that the common books originally belonged to Eth. Eud.,Footnote ³⁶ because of the similarity of the method. This also implies that the Eudemian framework remains in book 1 of Eth. Nic. to formulate the highest good in ethics. In Eth. Eud. 8.3 1249b5–6, Aristotle writes that ‘this [the requirement that action should be in accordance with reasoning] is true, but not clear (τοῦτο δ' ἀληθὲς μέν, οὐ σαφὲς δέ)’. In this context (Eth. Eud. 8.3 1249a21–b25), he discusses the definition (ὅρος) of good action by analogy with health in the same manner as he does in Eth. Nic. 6.1. There is a close similarity between Eth. Eud. 8.3 and Eth. Nic. 6.1.

Although I have been stressing the similarities between Eth. Eud. and Eth. Nic., they do differ in some methodological respects. First, in Eth. Eud. Aristotle seeks to offer a scientific causal account of ethical concepts. He claims that ‘one should not, even when it comes to politics, regard as superfluous the kind of study that makes clear not only what something is but also its cause’ (Eth. Eud. 1.6 1216b37–39). According to the standard Aristotelian doctrine, we have scientific knowledge about an object when we know its cause (An. post. 1.2 71b9–12; 2.11 94a20; Ph. 1.1 184a10–16; Metaph. 1.1 981a28–b6). What Aristotle has in mind by a causal account in the Eudemian context is better construed with reference to Eth. Eud. 1.5 1216b20–21, where he says, ‘when it comes to virtue, knowing what it is is not the most valuable point, but understanding what brings it about’. Here, a causal account is meant to be an explanation of how to cultivate virtue.Footnote ³⁷ In Eth. Nic., by contrast, Aristotle does not think that offering a causal account is the primary task in ethics (cf. Eth. Nic. 1.4 1095b6–7, quoted previously). He asserts that one should not demand an explanation in the same way in all cases. He says, ‘A good proof that something is the case will suffice in some instances, as with first principles, where the fact itself is a starting-point, that is, a first principle’ (Eth. Nic. 1.7 1098a33–b3, trans. Crisp (Reference Crisp2014)).

Second, in Eth. Eud. Aristotle makes a universal statement about the concept of virtue by induction, whereas in Eth. Nic. he develops the concept of virtue by drawing an analogy between human virtue and other goods. In Eth. Eud., Aristotle refers to the term ‘induction’ (ἐπαγωγή) (2.1 1219a1–2, 2.1 1220a28–29 and 2.3 1220b30) and verifies the first premises or assumptions of his arguments inductively in the following two arguments: about virtue (2.1 1218b37–1219a5) and about the best disposition (2.1 1220a22–29) (n.b. there is no induction about the mean relative to us (2.3 1220b27–30)). What is characteristic in these Eudemian arguments is thus that they offer inductive arguments from particular examples in order to provide their comprehensive definitions, and deduce the concept of human virtue from those universal definitions.Footnote ³⁸

In Eth. Nic., by contrast, Aristotle does not refer to the term ‘induction’ in the equivalent arguments at all, but rather offers those arguments by analogy. First of all, in the Nicomachean version of the function argument, he does not make a universal, general statement about goodness but instead draws an analogy, as discussed above, between a human and a craftsman, and between a human as a whole and the parts of a body. The terms ‘just as’ (ὥσπερ) – ‘so’ (οὕτω) and ‘just as’ (καθάπερ) – ‘so’ (οὕτω) clearly express the analogical formulation of this argument. The same pattern applies to other arguments (2.1 1103b13–14, 2.2 1104a14–27). Therefore, in Eth. Nic., he does not deduce the concept of human virtue from a comprehensive definition of functional activity or disposition, but instead applies the concept of skill and physical disposition analogously to the concept of human virtue.Footnote ³⁹

This difference between Eth. Eud. and Eth. Nic. shows us that the Eth. Nic. argument is more elaborate in this respect at least.Footnote ⁴⁰ One problem with the inductive argument is the question of whether virtue and other goods, such as health and skill, can be generalised to identify the concept of good. By contrast, although an analogical argument also draws attention to similarities between two objects, it can maintain a reservation about their differences. In fact, in Eth. Nic. 2.4 (1105a26–b5), Aristotle notes an important difference between virtue and skill in the sense that while having knowledge is one important condition for developing skill, it is not important for acquiring virtue. In the latter case, he argues, it is crucial for an agent to choose virtuous actions for the sake of the actions themselves on the basis of his stable, firm character. This remark is not made in Eth. Eud. Therefore, one point that Aristotle makes by analogy in Eth. Nic. refers to what is evident as good to explain the characteristics of virtue while retaining some reservation about the differences between various types of goods. Although he does not always draw a sharp contrast between induction and analogy (Metaph. Θ.6 1148a35–b9), and analogy is one type of induction in the sense that it infers from particulars, the way the arguments are formulated differs between the two ethical treatises. In this respect, the Nicomachean arguments are more elaborate versions of his argument about the human good than the Eudemian arguments.Footnote ⁴¹

Furthermore, this fact shows us that Eth. Eud. is intended to offer a philosophically rigorous account (1.6 1216b35–40) and makes no remark about the difference in accuracy between mathematical and ethical arguments, whereas this difference is an important methodological claim in Eth. Nic. (1.3 1094b19–27, 1.7 1098a26–33). In Eth. Eud., as previously discussed, ethical notions are inferred from general principles as in a mathematical argument. In Eth. Nic., by contrast, ethical enquiry is not compared to mathematical sciences. Although Eth. Nic. also aims to clarify ethical concepts with reference to evident cases, it is less schematic than Eth. Eud. when formulating the general principles of ethics.Footnote ⁴²