"Aristotle's Philosophical Method"

OUP UNCORRECTED PROOF – FIRST-PROOF, 01/06/12, NEWGEN Chapter 7 ARISTOTLE’S PHILOSOPHICAL METHOD C. D. C. Reeve A problem (problêma) is posed: Is pleasure choiceworthy, or not? The answerer claims that yes it is (or, alternatively, that no it isn’t). The questioner must refute him by asking questions—by offering him premises (protaseis) to accept or reject. The questioner succeeds if he forces the answerer to accept a proposition contrary to the one he undertook to defend (SE 2 165b3–4). The questioner fails if the answerer always accepts or rejects premises in a way consistent with that proposition. To a first approximation, dialectic is the art or craft (technê) enabling someone to play the role of questioner or answerer successfully (Top. I 1 100a18–21, VIII 14 164b2–4). Also to a first approximation, it is the distinctive method of Aristotelian philosophy. At the heart of dialectic is the dialectical deduction (dialektikos sullogismos). This is the argument lying behind the questioner’s questions, partly dictating their order and content, and partly determining the strategy of his attack. Understanding dialectic is primarily a matter of grasping the nature of dialectical deductions and the type of premises they employ. In Topics I 1, such deductions are contrasted with three other types of arguments: scientific, eristic, and paralogistic. In Sophistical Refutations I 2, they are distinguished from didactic, peirastic, and eristic arguments. Our task in sections 1–4 is to explore and co-ordinate these two sets of contrasts. When it is completed, we shall turn in sections 5–7 to a discussion of dialectical premises, endoxa (reputable beliefs), and aporiai (puzzles). Section 8 deals with the uses of dialectic in intellectual training (gumnasia), ordinary discussion (enteuxeis), and the philosophical sciences; section 9, with its use in regard to scientific starting-points 07_Shields_Ch07.indd 150 1/6/2012 6:01:07 PM OUP UNCORRECTED PROOF – FIRST-PROOF, 01/06/12, NEWGEN aristotle’s philosophical method 151 or first principles (archai). Section 10 returns to dialectic and philosophy and an important difference between them. 1. Dialectic, Eristic, and Sophistry Dialectical deductions differ from scientific ones only in their premises: the latter are deductions from starting-points and hence are demonstrations (apodeixeis); the former are deductions from endoxa (Top. I 1 100a1-b23; Met. III 1 995b23–4). In the case of eristic arguments the differences are potentially twofold: they are either genuine deductions from apparent endoxa or apparent deductions from genuine or apparent endoxa (Top. I 1 100b23–5). Paralogistic arguments differ from all these: unlike dialectical or eristic arguments, their premises are not endoxa, but ‘premises proper to a specialized science’ (Top. I 1 101a5–7); unlike scientific demonstrations, their premises are false (Top. I 1 101a14). ‘In dialectic,’ Aristotle tells us, ‘a sophist is so called on the basis of his deliberate choice (prohairesis), and a dialectician is so called not on the basis of his deliberate choice, but on the basis of the ability he has’ (Rhet. I 1 1355b20–1). If dialectic is understood in this way, it is a neutral craft and a dialectician who decides to employ eristic arguments is a sophist (Rhet. I 11355a24-b7). A contender (eristikos) also employs such arguments, but differs from a sophist in his purposes: ‘Sophistry . . . is a way of making money out of apparent wisdom. . . . Contenders and sophists use the same arguments, but not to achieve the same goal. . . . If the goal is apparent victory, the argument is eristic or contentious; if it is apparent wisdom, sophistic’ (SE 11 171b27–9). In the Topics and Sophistical Refutations, by contrast, the person who decides to use only genuine and never eristic arguments is a dialectician, since in both treatises dialectic differs from eristic precisely in employing genuine endoxa and genuine deductions rather than merely apparent ones (Top. I 1 100a29-b25, SE 2 165b3–8, 11 171b34–172a2). For clarity’s sake, let us say that plain dialectic is the neutral craft contenders, sophists, and honest dialecticians use for different purposes, imposing different restrictions on which of its resources may be legitimately employed. 2. Peirastic Deductions and Sophistical Refutations Peirastic (peirastikê) is ‘a type of dialectic which has in view not the person who knows (eidota), but the one who pretends to know but does not’ (SE 11 171b4–6). It is the type particularly useful in arguments with sophists, since they are the 07_Shields_Ch07.indd 151 1/6/2012 6:01:09 PM OUP UNCORRECTED PROOF – FIRST-PROOF, 01/06/12, NEWGEN 152 the framework of philosophy: tools and methods archetypal pretenders to knowledge and wisdom (SE 1 165a21). Though Aristotle usually uses the term peirastikê to refer to honest peirastic rather than to the plain craft (SE 2 165b4–6), he courts confusion, as we shall see, by using it to refer to the plain craft too. The best way to distinguish honest peirastic from honest dialectic pure and simple is by exploring sophistical refutations, which are the dishonest twins of honest peirastic arguments. Honest peirastic arguments expose the genuine ignorance of a sophist answerer, who has only apparent knowledge and wisdom (SE 11 171b3–6); sophistical refutations give the appearance of exposing the ignorance of someone who really does have scientific knowledge (SE 6 168b4–10). Such refutations are of two sorts. An a-type sophistical refutation is ‘an apparent deduction or refutation rather than a real one’; a b-type is ‘a real deduction that is only apparently proper to the subject in question’ (SE 8 169b20–3). A-type sophistical refutations are eristic arguments, therefore, whereas b-types are like paralogisms (SE 11 171b34–7). The paralogisms proper to a craft or science are those based on the startingpoints and theorems belonging to it (SE 11 171b38–172a1). Thus Hippocrates’ argument for squaring the circle by means of lunes is a geometrical paralogism, because it ‘proceeds from starting-points proper to geometry’ and ‘cannot be adapted to any subject except geometry’ (SE 11 172a4–5).1 Someone who uses Zeno’s argument that motion is impossible in order to refute a doctor’s claim that it is better to take a walk after dinner, however, has produced a b-type sophistical refutation, since Zeno’s arguments are not proper to geometry or medicine but ‘koinos (common)’ (SE 11 172a8–9). Such an argument is paralogistic, indeed, even when sound: ‘Bryson’s method of squaring the circle,2 even if the circle is thereby squared, is still sophistical because it is not in accord with the relevant subject matter’ (SE 11 171b16–18). The only difference between paralogisms and b-type sophistical refutations is that the former have premises proper to the answerer’s science but false, while the latter have premises not proper to it but true. Because paralogisms depend on premises proper to a science, it is the job of the scientist himself to diagnose and refute them. It is not his job to deal with b-type sophistical refutations (Phys. I 2 185a16–17, SE 9 170a36–8), however, but that of a dialectician: ‘It is dialecticians who study a refutation that depends on koina, that is to say, that do not belong to any [specialized] craft’ (SE 9 170a38–9). Dialecticians must also deal with Antiphon’s argument for squaring the circle, which is an a-type sophistical refutation, since by assuming that a circle is a polygon with a large but finite number of sides, it ‘does away with the starting-points of geometry’ (Phys. II 185a1–2)—in particular, with the principle that magnitudes are divisible without limit.3 It cannot be discussed in a way that presupposes those starting-points, therefore, and so must be discussed on the basis of koina (Top. II 101a35-b4). One view about koina is that they are axioms (axiômata)—starting-points common to all or many sciences (APo I 2 72a15–17, I 9 76b14–15). The laws of logic, such as the principle of noncontradiction, which hold at least analogically of all beings, are examples, as are other somewhat less general laws, such as the axioms 07_Shields_Ch07.indd 152 1/6/2012 6:01:09 PM OUP UNCORRECTED PROOF – FIRST-PROOF, 01/06/12, NEWGEN aristotle’s philosophical method 153 of equality, which are not universally applicable, but are also not proper to a single science or single genus of beings (APo I 10 76a38-b2). On one manuscript reading, indeed, SE 11 172a36–7 actually identifies koina with axiômata, with ‘identical (τ’ αὐτα) starting-points which hold true of everything.’ On another reading, it says only that there are ‘many of these (ταῦτα) [common] things in each area.’ Though most editors favour the first reading, the second is preferable.4 Axioms, as common to many sciences, cannot by themselves entail a proposition contrary to a conclusion proper to a specific science. Hence it is impossible to construct b-type sophistical refutations using axioms alone. Yet that is precisely what b-type sophistical refutations must use koina to do. As we saw in section 1, the only propositions that can figure as premises in dialectical arguments are endoxa. Since koina, too, can figure as such premises, they must be endoxa: ‘It is plain that it is the dialectician’s job to be able to grasp the various ways in which a real or apparent refutation—that is to say, one that is an example of dialectic or apparent dialectic or peirastic—can be achieved on the basis of koina’ (SE 9 170b8–11; compare Rhet. I 1 1354a1–3). The following two passages—the first referring to the second—settle the matter: ‘Even if one had the most rigorous sort of scientific knowledge, it would not be easy to persuade some people by arguments based on it . . . rather, it is necessary to construct our persuasions and arguments on the basis of koina, as we said in the Topics about ordinary discussions with the many’ (Rhet. I 1 1355a24–9); ‘(Plain) dialectic is useful in ordinary discussions because once we have catalogued the beliefs of the many, our approach to them will begin from their own views, not from other people’s, and we will redirect them whenever they appear to us to be wrong’ (Top. I 2 101a30–4). It follows that axioms that are endoxa will also be koina. Since the noncontradoxical5 views of philosophers are endoxa (section 5), it is a status that most if not all of them will have. Honest peirastic deductions ‘deduce from premises that are accepted by the answerer, and that must be known (eidenai) by anyone who claims to have the relevant scientific knowledge (epistêmê)’ (SE 2 165b4–6). Premises of this sort are said to be taken ‘not from the things from which one knows or even from those proper to the subject in question, but from the consequences that a man can know (eidota) without knowing the craft in question, but which if he does not know (eidota), he is necessarily ignorant of the craft’ (SE 11 172a21–34). In other words, such premises are not starting-points of the answerer’s science—not ‘things from which one knows’—or other starting-points proper to it, but consequences of them. Peirastic premises, unlike those of b-type sophistical refutations, must be proper to the answerer’s science, since they are syllogistic consequences of its starting-points. Later in the same passage these consequences are identified as koina (endoxa): Everybody, including those who do not possess a craft, makes use of dialectic as peirastic; for everyone tries to use peirastic to some extent in order to test those who claim to know things. And this is where the koina come in; for the testers 07_Shields_Ch07.indd 153 1/6/2012 6:01:09 PM OUP UNCORRECTED PROOF – FIRST-PROOF, 01/06/12, NEWGEN 154 the framework of philosophy: tools and methods know (isasin) these things for themselves just as well as those who do possess the craft—even if they seem to say quite inaccurate things. (SE 11 172a30–34). Hence the premises of honest peirastic deductions must be true endoxa proper to the answerer’s science—the one the sophist undergoing honest peirastic examination is pretending to know. A person who in other respects does have scientific knowledge may yet be the victim of a sophistical refutation, since he may find himself caught in a contradiction when interrogated by a clever sophist. The mere fact that someone can be bested in a dialectical argument is not enough to show that he lacks scientific knowledge. What is further required is: first, that this argument not be a sophistical refutation (its premises must be true and proper to the science in question); second, those premises must be such that anyone who knows the science would have to know them (otherwise, the answerer could reject them and still know the science); finally, they must be propositions it is possible to know without knowing the science (otherwise, they could not figure in arguments available to nonscientists). Thus the various features that the premises of an honest peirastic argument must have are entailed by the fact that their purpose is to enable nonscientists to unmask pretenders to scientific knowledge. In Topics VIII 5, Aristotle discusses ‘dialectical explorations that are not competitive, but are conducted for the sake of examination (peiras) and inquiry’ (159a32–33). From the account he provides of these, it is clear that they do not fit our characterization of honest peirastic. For example, the questioner is not restricted to using true premises; he can and sometimes must use false ones: Since arguments of this kind are conducted for the sake of practice and examination (peiras), it is clear that the questioner must deduce not only true conclusions but also false ones, and not always from true premises but sometimes from false ones as well. For often, when a true proposition is put forward [by the answerer], the dialectician is compelled to demolish it, and so he has to offer [the answerer] false premises. (Top. VIII 11 161a24–29) Moreover, the answerer may defend a position he himself does not hold (Top. VIII 5 159b27–35), and accept premises that are not proper to the topic of the argument (Top. VIII 6 160a1–2). Yet the very fact that Aristotle discusses how the answerer should deal with improper premises (Top. VIII 6) in connection with dialectical explorations that examine and enquire suggests that such explorations are at least closely related to b-type sophistical refutations and honest peirastic deductions. Indeed, it suggests that these dialectical explorations are simply exercises in plain peirastic. When Aristotle tells us in Sophistical Refutations I 2 that he has already discussed peirastic arguments, there is good reason to take him to be referring to the discussion of dialectical explorations that examine and investigate in Topics VIII 5–11. But to secure that reference, in the face of the manifest differences between what the two treatises say about peirastic, we must recognize that Sophistical Refutations mostly deals with honest peirastic, Topics with plain peirastic.6 07_Shields_Ch07.indd 154 1/6/2012 6:01:09 PM OUP UNCORRECTED PROOF – FIRST-PROOF, 01/06/12, NEWGEN aristotle’s philosophical method 155 3. Didactic Deductions Didactic deductions (didaskalikoi) are ‘those that deduce from the startingpoints proper to each subject matter and not from the opinions held by the answerer, since learners have to take things on trust’ (SE 2 165b1–3). This identifies them as scientific demonstrations of some sort—‘arguments based on scientific knowledge’ (Rhet. I 1 1355a26). But if they are scientific demonstrations, why are they included with honest dialectic, peirastic, and eristic arguments as one of the four types of argument used ‘in question and answer discussions’ (SE 2 165a38)? Didactic deductions are not deductions ‘from the opinions held by the answerer’ (SE 2 165b2). Yet ‘the student should always grant [only] what seems to him to be the case’ (Top. VIII 5 159a28–9), suggesting that didactic arguments must indeed be deductions from the student’s opinions. In Topics VIII, teaching sometimes takes the form of question and answer discussions. Yet teaching is also contrasted with asking questions: ‘the teacher should not ask questions but make things clear himself, whereas the dialectician should ask questions’ (SE 10 171b1–2). To grasp the coherence of Aristotle’s thought about didactic in the face of these apparent inconsistencies of doctrine, we need to appreciate the relevance to them of the distinction between an argument ‘taken by itself ’ and one ‘presented in the form of questions’ (Top. VIII 11 161a16–17). Suppose a student has acquired the starting-points of a science, and his teacher wants to test his knowledge of it. The natural thing for him to do is to examine the student by offering him propositions to accept or reject. And, of course, ‘the student should always grant [only] what seems to him to be the case’ (Top. VIII 5 159a28–9), since otherwise the teacher will not be able to discover what he really knows. Here the teacher’s didactic argument is ‘presented in the form of questions.’ But the admissions made by the student are not premises in the didactic argument (the scientific demonstration taken by itself) that underlies these questions and partly dictates their order and content. It is not a deduction ‘from the opinions held by the answerer’ (SE 2 165b2). Suppose a phrase occurring in a scientific proposition has a double meaning, but that the student ‘neither has considered nor knows nor conceives that a second meaning is possible’ (SE 10 171a32–4). Then ‘the teacher should not ask questions but make things clear himself ’ (SE 10 171b1–2). Here, unlike in the previous case, the teacher is not trying to find out what the student knows by asking him questions. He already knows that the student is ignorant and is providing him with information. So he uses a didactic argument ‘taken by itself ’ to make things clear. Once we see that teaching may involve question and answer discussion as well as straightforward demonstration, so that didactic arguments can be understood in two different ways, we can see that these arguments do have a place in question and answer discussions and that Aristotle’s account of them is consistent.7 07_Shields_Ch07.indd 155 1/6/2012 6:01:09 PM OUP UNCORRECTED PROOF – FIRST-PROOF, 01/06/12, NEWGEN 156 the framework of philosophy: tools and methods 4. The Classification of Deductions In Topics I 1, deductions are divided into four classes: (T1) scientific (T2) paralogistic (T3) honest dialectic (T4) eristic. In Sophistical Refutations I 2 they are also initially divided into four: (S1) didactic (S2) peirastic (S3) honest dialectic (S4) eristic. Then two more are added: (S5) a-type sophistical refutations (S6) b-type sophistical refutations. Though apparently discordant, the two classifications fit together to constitute a single systematic classification of dialectical deductions. Deductions are generally of two kinds: (D1) genuine (valid) (D2) apparent (invalid). The premises of each may be: (P1) true and proper starting-points of a science (P2) untrue but proper starting-points of a science (P3) true endoxa proper to a science (P4) true endoxa only apparently proper to a science (P5) endoxa (P6) apparent endoxa. (D1–2) and (P1–6) together determine the various kinds of dialectical deductions: (D1)-(P1) scientific demonstrations (T1); presupposed in didactic arguments (S1) (D1)-(P2) paralogisms (T2) (D1)-(P3) peirastic deductions (S2) (D1)-(P4) b-type sophistical refutations (S6) (D1)-(P5) honest dialectic arguments (T3), (S3) (D1)-(P6) eristic arguments or a-type sophistical refutations (T4), (S4), (S5) (D2)-(P5) eristic arguments or a-type sophistical refutations (T4), (S4), (S5). A striking feature of this classification is that it includes only one type of invalid deduction, namely, (D2)-(P5). This is so for a reason. The various kinds of formally valid and invalid deductions have already been studied in the Prior Analytics. Topics and Sophistical Refutations are primarily concerned not with them, therefore, but with sound or unsound ones—with the choice of premises rather than with the logical form of arguments (APr I 30 46a29–30). Hence the classification is both complete and systematic. 07_Shields_Ch07.indd 156 1/6/2012 6:01:09 PM OUP UNCORRECTED PROOF – FIRST-PROOF, 01/06/12, NEWGEN aristotle’s philosophical method 157 5. Dialectical Premises A dialectical premise consists in [a] making a question out of something that is endoxos to everyone or to the majority or to the wise—either to all of them, or to most, or to the most notable of them, provided it is not contradoxical; for a person would accept the opinion of the wise, provided it is not contrary to general opinion (doxa). Dialectical premises also include [b] things that are like endoxa, and [c] propositions that contradict the contraries of what seem to be endoxa, and also [d] all opinions that accord with [the starting-points of] the recognized crafts, . . . since a person would accept the opinions of those who have investigated the subjects in question—for example, on a question of medicine he will agree with the doctor, and on a question of geometry with the geometer. (Top. I 10 104a8–37) Later, in a reprise of this passage, Aristotle adds what seem to be two new cases to the account: [e] Furthermore, statements that seem to hold in all or in most cases, should be taken as starting-points, that is to say, as accepted theses; for such statements are accepted by those who do not notice that there is a case in which they do not hold. [f] We ought also to select [premises] from written accounts and draw up lists of them on each type of subject, putting them under separate headings— for example, ‘Dealing with good’, ‘Dealing with life’. And the one dealing with good, should deal with every kind of good, beginning with the essence. (Top. I 14 105b10–15) The fact that (b) describes propositions that are ‘like endoxa,’ that (c) speaks of the contraries of what ‘seem to be endoxa,’ and that (e) includes as endoxa statements that merely seem to be true to those ‘who do not notice that there is a case in which they do not hold’ strongly suggest that these clauses refer to apparent endoxa. Aristotle’s illustrative examples bear this out: (i) ‘If it is an endoxon that the science of contraries is the same, it might appear to be an endoxon that the perception of contraries is also the same’ (Top. I 11 104a15–17); (ii) ‘Propositions contradicting the contraries of endoxa will appear to be endoxa’ (Top. I 10 104a20–3); (iii) ‘If it is an endoxon that there is a single craft of grammar, it might also seem to be an endoxon that there is a single craft of flute-playing’ (Top. I 10 104a17–20). (i) and (ii) explicitly refer to apparent endoxa, while (iii) makes sense only if it too has them in view, since if a proposition is a genuine endoxon, its contrary cannot be (Top. VIII 5 159b4–6). Since both endoxa and apparent endoxa can serve as premises in plain dialectical deductions, we cannot identify genuine endoxa with such premises, or infer that everything said about the latter applies willy-nilly to them. The propositions referred to in (d) are in accord with the starting-points of the recognized crafts, so they must be genuine. But because they only would be accepted by anyone, they do not have to be already accepted so to count. Since written accounts are likely to have wise people or practitioners of the recognized crafts as authors, (f) is probably a new source of something already listed rather than a 07_Shields_Ch07.indd 157 1/6/2012 6:01:09 PM OUP UNCORRECTED PROOF – FIRST-PROOF, 01/06/12, NEWGEN 158 the framework of philosophy: tools and methods wholly new addition to the list. Aristotle himself suggests as much when he writes that we should note in the margins of the lists we distil from these writings the identity of the thinkers, such as Empedocles, who hold them, since ‘anyone might assent to the saying of some endoxos (reputable) thinker’ (Top. I 14 105b17–18). Because medicine is itself an acknowledged craft or recognized area of expertise, the opinions of a doctor known to have studied medicine carry weight with everyone, whether or not the doctor himself has already acquired a good reputation. Hence if a person can show that he has been trained as a doctor, that is enough, everything else being equal, to guarantee that the answerer would accept his opinion on medical matters. Of course, someone can be wise without being a practitioner of a recognized craft, but his epistemic authority cannot then flow from his training. Nor is it enough that he be wise. If his opinions are to have any standing, the answerer must recognize him as a wise person. In other words, like Solon or Thales, he must be notable for his wisdom or have a reputation as a wise man. Hence the reference to notability and reputation in the relevant clause of the definition of endoxa (Top. I 1 100b23). (a) corresponds closely to the official definition of genuine endoxa as ‘things that are held by everyone, by the majority, or by the wise—either by all of them, or by most, or by the most notable and most endoxos (reputable)’ (Top. I 1 100b21–3; repeated 101a11–13). But it also adds something new, namely, that views held by all, most, or the most reputable wise people have to meet a negative condition if they are to count as endoxa—they cannot be contradoxical or ‘contrary to general opinion’ (Top. I 10 104a11–12). Some of the endoxa characterized in (a) are accepted by all or most answerers, because they are accepted by someone whose epistemic authority stems from his reputation for wisdom. Those characterized in (d) are accepted because they are accepted by someone whose epistemic authority stems not from his reputation but from his having been trained in an acknowledged area of expertise.8 Some of the endoxa characterized in (a) and all of those characterized in (d) are thus indirect: they are (or would be) accepted by all or most answerers, because they are accepted by someone whose authority they recognize. The other endoxa characterized in (a) are direct: they are accepted on other grounds. 6. Endoxa and Phainomena From our discussion in section 5, we see that genuine endoxa fall into three classes: (1) propositions that all or most ordinary people would accept; (2) noncontradoxical propositions—propositions not contrary to what is already in (1)—that all, or most, or the most notable of the wise accept; (3) propositions in accord with—that 07_Shields_Ch07.indd 158 1/6/2012 6:01:09 PM OUP UNCORRECTED PROOF – FIRST-PROOF, 01/06/12, NEWGEN aristotle’s philosophical method 159 follow from—the starting-points of the recognized crafts, since everyone, ordinary people included, would accept them. It seems, then, that (1) is acting as a sort of gatekeeper class. If p is in (2), it cannot be an endoxon unless it can be consistently added to (1). If p is a proposition in (3), it could, apparently, conflict with those in (1) while retaining its status as an endoxon, but only by joining (1) and depriving any conflicting propositions of membership in it. The fact that all or most people believe something, Aristotle claims, leads us ‘to trust it as something based on experience’ (Div. Somn. 1 462b14–16). For ‘human beings are naturally adequate as regards the truth and for the most part happen upon it’ (Rhet. I 1 1355a15–17), so that each person ‘has something of his own to contribute’ to it (EE I 6 1216b30–1). Thus experience—whether in the form of perception or correct habituation (Top. I 11 105a3–7, EN I 4 1095b4–8, EE I 3 1214b28–1215a3)—must surely be what provides the evidence for direct endoxa in class (1). Direct endoxa are thus beliefs that seem true to us on the basis of experience. Presumably, that is why Aristotle occasionally refers to them as phainomena—as things that seem to be so (Top. I 10 104a12 with 14 105a37-b1, EE VII 2 1235b13–18 with EN VII 1 1145b2–7). Phainomena include, in the first instance, basic perceptual observations: ‘This [that the earth is spherical] is also shown by the sensory phainomena. For how else would lunar eclipses exhibit segments shaped as we see them to be?’ (DC II 14 297b23–5; also 297a2–6). But though phainomena are for this reason typically contrasted with things that are supported by proof or evidence (EE I 6 1216b26–8), there seems to be no a priori limit on the degree of conceptualization or theoryladenness manifest in them. They need not be, and in Aristotle rarely are, devoid of interpretative content. It is a phainomenon, for example, that the incontinent person ‘knows that his actions are base, but does them because of his feelings, while the continent one knows that his appetites are base, but because of reason does not follow them’ (EN VII 1 1145b12–14). Since all the crafts and sciences—indeed, all types of knowledge, however humble or exalted—rest ultimately on experience (APr I 30 46a17–18, Gen. et Corr. I 2 316a5–6), what is true of direct endoxa also seems true of indirect ones. They are propositions that seem true on the basis of experience not to the untutored eyes of people in general, but to the relatively more trained ones of craftsmen and scientists, or the relatively more reflective ones of reputable philosophers. It follows, once we make proper allowance for the division of epistemic labor, that the entire class of endoxa—direct and indirect—is epistemically homogeneous: it consists of propositions that seem true on the basis of experience. It is important to be clear, however, that Aristotle does not presuppose that endoxa are all guaranteed to be true. To be sure, an endoxon has epistemic credentials that are from the point of view of dialectic nonpareil. But that is because dialectic deals with things only ‘in relation to opinion’ not, as philosophy does, ‘in relation to truth’ (Top. I 14 105b30–1). If a proposition is an endoxon, if it would be accepted by all or most people, it is everything an honest dialectician could ask for in a premise. But that does not mean that it will retain its credibility when the philosopher has done his aporematic or aporia-related work. 07_Shields_Ch07.indd 159 1/6/2012 6:01:09 PM OUP UNCORRECTED PROOF – FIRST-PROOF, 01/06/12, NEWGEN 160 the framework of philosophy: tools and methods 7. Problems, Theses, and Aporiai A dialectical problem is a subject of inquiry . . . about which [a] people hold no opinion either way, or [b] on which the many hold an opinion contrary to that of the wise, or [c] the wise contrary to that of the many, or [d] about which the members of either of these classes disagree among themselves . . . Problems also occur [e] where deductions conflict, since there is an aporia about whether the thing holds or not, because there are strong arguments on both sides. They occur, too, [f] where we have no argument because they are so vast, and we find it difficult to give an explanation— for example, is the universe eternal or not? For one may also inquire into problems of that sort. (Top. I 11 104b1–17) If there is disagreement over some proposition, p, whether (b) between the many and the wise or (c, d) within either party, p—or more accurately the corresponding question, p?—is a problem. However, not all problems result from conflicts in opinion, or from the existence of contradoxical opinions, some exist (a) because we have no opinions about them, or (f) no arguments for or against them. If p is contradoxical, but is held by even one notable philosopher, or if there is an argument for not-p, p (or p?) is a dialectical problem of a distinctive sort: A thesis is a contradoxical belief of some notable philosopher. . . . For it would be silly to pay any attention when an ordinary person expresses views that are contrary to general opinion. Or it may be a view contrary to general opinion that is supported by an argument. . . . For even if this view is unacceptable to someone, it might well be accepted [by the answerer] because it is supported by argument. A thesis is also a problem; but not every problem is a thesis, since some problems are such that we hold no opinion about them either way. (Top. I 11 104b19–28) Whenever there is some reason, however slight, in favour of a contradoxical proposition, a problem exists. But this means that the endoxa to which such a proposition are contrary become problematic—especially as dialectical premises. The class of endoxa, as we might put it, has a built in tendency towards consistency—a tendency that dialectical practice itself helps further. An aporia, (e) suggests, is a problem of a second particular sort. There is an aporia about whether p just in case there are strong arguments for p and strong arguments against it: The sophistical argument [against incontinence] is an aporia. For because they want to refute people in contradoxical ways, so that they will be clever in ordinary discussions, the deduction they construct gives rise to an aporia; for thought is tied up in a knot, since it does not want to stand still because it dislikes the conclusion, but it cannot move forward because it cannot undo the argument. (EN VII 2 1146a21–7) Philosophy, in its aporematic capacity, is particularly concerned with problems of this sort: ‘If we want to move forward [in philosophy], our first task is to explore 07_Shields_Ch07.indd 160 1/6/2012 6:01:10 PM OUP UNCORRECTED PROOF – FIRST-PROOF, 01/06/12, NEWGEN aristotle’s philosophical method 161 the aporiai well; for we will be in a position to do so later only if we free ourselves of earlier aporiai by undoing them; but we cannot undo them if we do not know that we are tied up’ (Met. III 1 995a27–30). 8. Uses of Dialectic Dialectic has four apparently distinct uses, three of which are the focus of the present section: (a) intellectual training, (b) ordinary discussions, and (c) in relation to the philosophical sciences (Top. I 2 101a26–7). Dialectic’s usefulness for (a) training is ‘immediately evident,’ because ‘if we have a line of inquiry, we can more easily take on a question proposed to us’ (Top. I 2 101a28–30). Since all other uses provide intellectual training too, just as all sports provide physical training, this use is presumably the broadest one. If we are dialektikos—if we are dialectically proficient (Top. VIII 14 164b1–4)—we will be better able to deal with any question put to us by any sort of questioner. Contrariwise, dealing with all sorts of questioners will tend to make or keep us more dialectically proficient. Dialectic is useful in (b) ‘ordinary discussions,’ because, as we saw, ‘once we have catalogued the beliefs of the many, our approach to them will begin from their own views, not from other people’s, and we will redirect them whenever they appear to us to be wrong’ (Top. I 2 101a31–4). Here, it is dialectic’s systematic collecting and categorizing of endoxa (Top. I 14 105b12–18) that proves particularly helpful. For by knowing what people will accept as premises, we will be better able to argue effectively and persuasively against them when they seem to be mistaken— even if their own lack of dialectical training means that the argument is sometimes ‘bound to degenerate’ (Top. VIII 14 164b9–10). Aristotle sometimes applies the term ‘philosophy’ to any of the sciences that aim, in particular, at theoretical truth: ‘It is also right that philosophy should be called scientific knowledge of the truth. For the end of theoretical knowledge is truth, while that of practical knowledge is action’ (Met. II 1 993b19–20). In this sense, any non-practical science will count as philosophy. At the same time, Aristotle occasionally recognizes some non-theoretical philosophies, such as ‘the philosophy of human affairs’ (EN X 9 1180b15) or ‘political philosophy,’ classifying some of his own writings as ‘those philosophical works of ours dealing with ethical issues’ (Pol. III 12 1282b19–23). Finally, to make matters yet more complex, ‘philosophy’ also has a narrower, more specialized sense, in which it applies exclusively to sciences that provide theoretical knowledge of scientific starting-points (Met. XI 1 1059a18). It is in this sense of the term that there are ‘three theoretical philosophies, mathematical, natural, and theological’ (Met. VI (Epsilon) 1 1026a18–19).9 It is hard to know which sense of ‘philosophical sciences’ is pertinent in (c), so fortunately not much hangs on settling the matter. For what makes dialectic 07_Shields_Ch07.indd 161 1/6/2012 6:01:10 PM OUP UNCORRECTED PROOF – FIRST-PROOF, 01/06/12, NEWGEN 162 the framework of philosophy: tools and methods useful to these sciences, however we identify them, is that its ‘ability to go through the aporiai on both sides of a subject makes it easier to see what is true and false’ (Top. I 1 101a24–6). What this means is explained more fully as follows: Where knowledge (gnôsin) and philosophical wisdom are concerned, the ability to discern and hold in one view the consequences of either hypothesis is no insignificant tool, since then it only remains to make a correct choice of one of them. But a task of this sort requires euphuia. And true euphuia consists in just this—the ability to choose the true and avoid the false. For people with euphuia are the very ones who can do this well, since they judge correctly what is best by a correct love or hatred for what is set before them. (Top. VIII 14 163b9–16) Suppose that the problem a philosopher faces is, as before, to determine whether or not pleasure is always choiceworthy. If he is a competent dialectician, he will be able to follow out the consequences of supposing that it is, as well as those of supposing that it is not. He will be able to see what aporiai these consequences in turn face, and he will be able to go through these and determine which can be solved and which cannot.10 For this is just what a dialectician has to be able to do in order successfully to play the role of questioner or answerer in a dialectical argument about the choiceworthiness of pleasure. But this ability alone will not tell the philosopher where the truth lies. For that he also needs euphuia (explained in section 10). In the end, the philosopher will have concluded, we may suppose, that some sorts of pleasure are sometimes choiceworthy, while others are never choiceworthy. But in the process of reaching that conclusion some of the endoxa on both sides will almost certainly have been modified or clarified, partly accepted and partly rejected (Top. VIII 14 164b6–7). Others will have been decisively rejected as false. But these the philosopher will need to explain away: ‘We must not only state the true view, however, but also give the explanation for the false one, since that promotes confidence. For when we have a clear and good account of why a false view appears true, that makes us more confident of the true view’ (EN VII 14 1154a24–5). In other words, some beliefs that seemed to be genuine endoxa will have been revealed to be merely apparent. But if ‘most of them and the most compelling’ are still in place, that will be ‘an adequate proof’ (EN VII 1 1145b5–7) of the philosopher’s conclusion. It might seem that philosophy, at least in this aporematic role, has now simply collapsed into honest dialectic, but this is not so. In an honest dialectical argument, the answerer may refuse to accept a proposition that a philosopher would accept: The premises of the philosopher’s deductions or those of the man who is investigating by himself, though true and familiar, may be refused by the answerer because they lie too near to the original proposition, and so he sees what will happen if he grants them. But the philosopher is unconcerned about this. Indeed, he will presumably be eager that his axioms should be as familiar and as near to the question at hand as possible, since it is from premises of this sort that scientific deductions proceed. (Top. VIII 1 155b10–16; also APr I 30 46a3–10) Since the truth may well hinge on propositions whose status is just like the premises referred to here, there is no guarantee that honest dialectic and aporematic philosophy will reach the same conclusion on a given problem. 07_Shields_Ch07.indd 162 1/6/2012 6:01:10 PM OUP UNCORRECTED PROOF – FIRST-PROOF, 01/06/12, NEWGEN aristotle’s philosophical method 163 Perhaps enough has been said about this particular philosophical use of dialectic to show that it is relatively uncontroversial from the methodological and epistemological points of view. Dialectical ability helps an aporematic philosopher reach the truth in a way that is readily intelligible, but does not guarantee that he will reach it. For that he needs euphuia as well. The philosopher employs endoxa as premises of his arguments, but he does not employ all and only those available to a dialectician. And he does not simply accept them. They are presumptively true, but this presumption can be cancelled. 9. Dialectic and Starting-Points In addition to its uses in training, ordinary discussions, and the philosophical sciences, dialectic is also [d] useful with regard to the starting-points in each science. For [e] it is impossible to discuss them at all from the starting-points proper to the science proposed for discussion, since the starting-points are primary among all [the truths contained in the science]; instead they must be discussed through the endoxa about them. This is distinctive of dialectic, or more appropriate to it than to anything else; for [f] since it examines (exetastikê), it provides a way towards the starting-points of all lines of inquiry (Top. I 2 101a36-b4) According to (e), a certain kind of discussion of starting-points is impossible. Whether it is a dialectical discussion, in which starting-points appear as the contents of dialectical problems, or a philosophical investigation into starting-points, the premises involved cannot be the starting-points themselves, since they are the very things at issue. Instead, they must be endoxa. But, as we saw in the previous section, the class of endoxa the aporematic philosopher considers is typically broader than the class available to the honest dialectician, who is limited to employing endoxa that an answerer, eager not to be refuted, can reasonably be expected to accept. By the same token, when (f) tells us that dialectic provides a way towards starting-points because it examines (exetastikê), it could be referring to dialectical examination of some sort or to philosophical examination. The verb exetazein is used to refer to both sorts of activities. In the opening sentence of the Rhetoric, for example, it refers to dialectical questioning or examining in general: ‘everyone attempts either to examine propositions or maintain them’ (I 1 1354a4–5). At EN I 4 1095a28 and EE I 3 1215a6, it refers to an aporematic philosopher’s examination of various views, popular as well as expert, on the nature of happiness. Suppose that the discussion envisaged in (e) is dialectical. In that case, there are a set number of forms it can take. If p is a starting-point of geometry, the problem under discussion will be: p? If the answerer claims that p (as he may if he is a 07_Shields_Ch07.indd 163 1/6/2012 6:01:10 PM OUP UNCORRECTED PROOF – FIRST-PROOF, 01/06/12, NEWGEN 164 the framework of philosophy: tools and methods geometrician), questioner’s argument must be either an a-type or b-type sophistical refutation. If it is a b-type, the answerer’s responses (provided he is honest) must be based on an honest peirastic argument. If it is an a-type, his answers must be based on an honest dialectical argument of some other sort. If the answerer claims that not-p (as he may if he is a sophist pretender to scientific knowledge of geometry), his underlying argument must be either an a-type or a b-type sophistical refutation (or what would be such a refutation if it were being used to refute rather than to defend), while questioner’s argument (provided he is honest) must be either an honest peirastic argument or an honest dialectical argument of some other sort. In a dialectical discussion of starting-points, therefore, various types of honest dialectical argument will be involved, depending on what position the answerer takes and what sort of argument he employs in support of his position. Hence, if the way towards starting-points (f) envisages, is one that begins in such discussions, there is no reason to think that it has to be a peirastic one.11 It is useful to focus on honest peirastic arguments, nonetheless, in order to see the epistemic limitations of honest dialectic generally. Honest peirastic arguments have premises that are endoxa of a very special kind, namely, known (eidenai) truths—though not truths scientifically known (epistasthai) to the participants in these arguments (section 2). So even if these endoxa get refined through philosophical examination, they cannot be rejected or explained away. Thus honest peirastic arguments offer an epistemically better way towards starting-points than any other kind of dialectical argument. If what they offer has limitations, shifting our allegiance to some other type will simply make things worse. The epistemic weakness of honest peirastic arguments emerges most clearly if we first presuppose that the science involved in them is in fact possessed by someone other than questioner or answerer. The situation we have to imagine is something like this. The science of geometry exists in finished form as a structure of demonstrations from starting-points. q is a conclusion of one of these demonstrations that is known—although not scientifically known—to both questioner and answerer. Indeed, if the answerer did not know q, his pretense to be a geometrician would be immediately revealed as just that, since q must be known to anyone who claims to know geometry. q can then function as a premise in an honest peirastic argument: it can be used to deduce the negation of the false geometrical claim (not-p) made by the sophist answerer. Since this deduction must be sound, it establishes that p is true. Since p is a starting-point of geometry, it establishes that some starting-point of geometry is true. Since its premises are known, it leads the sophist answerer, at least, to know p. Yet, because p is a starting-point of geometry, the operating presupposition is that it is already scientifically known. Consequently, our peirastic argument does nothing to increase anyone’s store of scientific knowledge. For one cannot get scientific knowledge from premises that are not themselves known scientifically (APo I 3 72b18–23). Thus the peirastic way towards scientific starting-points is unimpressive. All it does is lead pretenders to scientific knowledge to a less profound kind of knowledge of starting-points than genuine scientists already possess. 07_Shields_Ch07.indd 164 1/6/2012 6:01:10 PM OUP UNCORRECTED PROOF – FIRST-PROOF, 01/06/12, NEWGEN aristotle’s philosophical method 165 If we now drop the presupposition that scientific knowledge of geometry is possessed by anyone, a different defect in peirastic arguments is revealed. If we do not have scientific knowledge of p as a starting-point of geometry, the peirastic deduction of p from q, will not even lead us to know that it is a starting-point, since this involves knowing its place in the demonstrative structure of completed geometry. Given this second failing of peirastic, it is hard to see it as giving us any kind of knowledge of starting-points as such. We may conclude that if the way referred to in (f) is one that begins in dialectical discussions—if the examination it refers to is peirastic examination or some other sort of honest dialectical examination—it is not a way any scientist should bother to take. Aristotle himself acknowledges as much in the following text: What causes our inability to take a comprehensive view of the agreed-upon facts is lack of experience. That is why those who dwell in more intimate association with the facts of nature are better able to lay down starting-points which can bring together a good many of these, whereas those whom many arguments have made unobservant of the facts come too readily to their conclusions after looking at only a few facts. One can see, too, from this the great difference that exists between those whose researches are based on the facts of nature and those who inquire [merely] dialectically (logikôs). (Gen. et Corr. I 2 316a6–11)12 Experience based on intimate association with the natural facts is the scientific way to starting-points, not dialectical argument. We turn now to the other alternative, where (f) is referring not to dialectical, but to philosophical examination. Experience has provided starting-points to the scientist and he has developed a finished science—a structure of demonstrations—from them. The philosopher is aware of this science and its status as such, and so accepts that its starting-points must—as inductively justified and explanatorily adequate—be true. Yet he also sees that the way towards those startingpoints is blocked by aporiai, since arguments based on endoxa entail that they cannot be true. His goal is to solve these aporiai, by undoing the arguments that seem to support them—something he can only do if he is aware of the aporiai themselves: Those who wish to be free of aporiai must first go through the aporiai well; for the subsequent aporia-free condition is reached by untying the knots produced by the aporiai raised in advance, and it is not possible for someone who is unaware of a knot to untie it. An aporia in thought, however, reveals a knot in its subject matter.13 For thought caught in an aporia is like people who are tied up, since in either case it is impossible to make progress. That is why one must have studied all the difficulties in advance, both for these reasons and because those who inquire without first going through the aporiai are like people who don’t know where they have to go, and, in addition, don’t even know whether they have found what they were inquiring about, since the end is not clear to them. But to someone who has first gone through the puzzles it is clear. Besides, one is necessarily in a better position to discern things when one has heard all the competing arguments, like opposing parties in a courtroom. (Met. III 1 995a27-b4) 07_Shields_Ch07.indd 165 1/6/2012 6:01:10 PM OUP UNCORRECTED PROOF – FIRST-PROOF, 01/06/12, NEWGEN 166 the framework of philosophy: tools and methods If he is successful in cataloguing and solving these aporiai, his way toward the starting-points will be cleared. And it is only when it is cleared that the startingpoints themselves are grasped in the way requisite for scientific knowledge that is genuinely unconditional: If we are to have scientific knowledge through demonstration, . . . we must know the starting-points better and be better convinced of them than of what is being proved, but we must also not find anything more convincing or better known among things opposed to the starting-points, from which a contrary mistaken conclusion may be deduced, since someone who has unconditional scientific knowledge must be incapable of being convinced [out of it]. (APo I 2 72a37-b4) Aporematic philosophy thus completes science by defending scientific startingpoints in a way that science itself cannot. That is why theoretical wisdom (sophia), as the most rigorous (akribês) form of scientific knowledge, must be ‘understanding plus scientific knowledge; scientific knowledge, having a head as it were’ (EN VI 7 1141a16–20). In defending some starting-points against dialectical objection, moreover, we provide a sort of demonstration of them, namely, a ‘demonstration by refutation’ (Met. IV 4 1006a11–12). Included among these are very secure or fundamental starting-points such as the principle of non-contradiction, which we must know in order to know anything. But it may also hold more generally: ‘a disputant’s refutation of what is opposed to his accounts is a demonstration of them’ (EE I 3 1215a6–7). Even when philosophy doesn’t offer us this sort of demonstration of starting-points, however, what it does offer is no puzzling knots—no impediments to clear and strict understanding (EN VII 2 1146a24–27). 10. Philosophy and Dialectic ‘Dialecticians practice dialectic about all things . . . because all things are proper to philosophy. For . . . dialectic treats the same genus as philosophy, but philosophy differs from dialectic in the type of power it has. . . . Dialectic tests in the area where philosophy achieves knowledge (esti de hê dialektikê peirastikê peri hôn hê philosophia gnôristikê)’14 (Met. IV 2 1004b19–26). Because it can draw out the consequences of each of the hypotheses (p, not-p) in a problem and go through the aporiai they face, dialectic can test those hypotheses. But it cannot achieve knowledge, because it lacks a type of power that philosophy possesses. Our task now is to explain what this power is. When dialectic has done its testing of p and of not-p, as we saw, it ‘only remains to make a correct choice of one of them’ (Top. VIII 14 163b9–12). Since euphuia is what enables people to ‘discern correctly what is best by a correct love or hatred 07_Shields_Ch07.indd 166 1/6/2012 6:01:10 PM OUP UNCORRECTED PROOF – FIRST-PROOF, 01/06/12, NEWGEN aristotle’s philosophical method 167 of what is set before them’ (Top. VIII 14 163b15–16), it seems to be the power we are seeking. The reference to ‘what is best’ suggests too that the euphuia in question may be the sort referred to in the following passage: A person doesn’t aim at the end [the good] through his own choice; rather, he must by nature have a sort of natural eye to make him discern well and choose what is really good. And the person who by nature has this eye in good condition is euphuês. For it is the greatest and noblest thing . . . and when it is naturally good and noble, it is true and complete euphuia. (EN III 5 1114b5–12) And that, in fact, is what the distinction between philosophy and sophistry, which uses all of plain dialectic’s resources, might lead us to expect, since ‘philosophy . . . differs from sophistic in its deliberate choice about how to live’ (Met. IV (Gamma) 2 1004b23–5). A deliberate choice of how to live is au fond a choice of an ultimate end or target for one’s life: ‘everyone who can live in accord with his own deliberate choice should adopt some target for the noble life, whether honour, reputation, wealth, or education, which he will look to in all his actions’ (EE I 2 1214b6–9). And what ‘teaches correct belief’ about this end or target, thereby insuring that the deliberate choice of it is itself correct, is ‘natural or habituated virtue of character’ (EN VII 8 1151a18–19). It is this, we may infer, in which euphuia consists. Hence if we possess it, when we hear from political science that the starting-point it posits as the correct target for a human life is ‘activity of the soul in accord with virtue, and if there are more virtues than one, in accord with the best and most complete’ (EN I 7 1098a16–18), we will accept it as true, and so strive to clear away the aporiai that block our road to it. If we do not possess such euphuia, we will reject this startingpoint and strive to sustain the aporiai that block our path to it, so that in our choice between p and not-p, we will go for the wrong one: ‘the truth in practical matters must be discerned from the things we do and from our life, since these are what have the controlling vote. Hence when we examine everything that has been previously said, it must be by bringing it to bear on the things we do and on our life, and if it is in harmony with what we do, we should accept it, but if it conflicts, we should suppose it mere words.’ (EN X 8 1179a17–22) In the Rhetoric, we learn of an apparently different sort of euphuia, which seems from the company it keeps to be an exclusively intellectual trait: ‘euphuia, good memory, readiness to learn, quick-wittedness . . . are all productive of good things’ (I 6 1362b24–5). When it comes to solving dialectical problems bearing on ‘truth and knowledge,’ we might conclude, such apparently intellectual euphuia is all a philosopher needs, even if, when it comes to those bearing on ‘pursuit and avoidance’ (Top. I 11 104b1–2; compare EN VI 2 1139a21–2), he also needs its apparently more ethical namesake. Whatever we decide about this, our account of intellectual euphuia can nonetheless take the account of ethical euphuia as a useful guide. Aristotle sometimes refers to what he calls ‘a well-educated person (pepaideumenos)’—someone who studies a subject, not to acquire scientific knowledge of it, but to become a discerning judge: 07_Shields_Ch07.indd 167 1/6/2012 6:01:10 PM OUP UNCORRECTED PROOF – FIRST-PROOF, 01/06/12, NEWGEN 168 the framework of philosophy: tools and methods Regarding every branch of theoretical knowledge and every line of inquiry, the more humble and more estimable alike, there appear to be two ways for the state to be, one which may be well described as scientific knowledge of the subject matter, the other a certain sort of educatedness. For it is characteristic of a person well educated in that way to be able accurately to discern what is well said and what is not. We think of someone who is well educated about the whole of things as a person of that sort, and we think that being well educated is being capable of doing such discerning. Except that, in the one case, we consider a single individual to be capable of being discerning in practically all subjects, in the other, in one of a delimited nature—for there might be another person disposed in the same way as the person we have been discussing, but about a part. So it is clear in the case of inquiry into nature, too, that there should be certain defining-marks by referring to which one can appraise the manner of its demonstrations, apart from the question of what the truth is, whether thus or otherwise. (PA I 1 639a1–15) A person well educated in medicine, for example, is capable of discerning whether someone has treated a disease correctly (Pol. III 11 1282a3–7), and the ‘unconditionally well-educated person,’ who is well educated in every subject or area, ‘seeks rigor in each area to the extent that the nature of its subject matter allows’ (EN I 3 1094b23–1095a2). Whether identical to intellectual euphuia, or a state developed from it by intellectual training in the way that habituated virtue is developed from natural virtue by adequate upbringing, it is surely this sort of educatedness the aporematic philosopher needs to perform the task Aristotle assigns to intellectual euphuia. For if he is well-educated he will be discerning in the realm of knowledge, able to distinguish genuine sciences from specious or sophistic look-alikes, and so be able to determine which starting-points he should be trying to find an aporiafree way toward. Aporematic philosophy is not the only sort of philosophy Aristotle recognizes, of course. As we saw in section 8, he also recognizes a number of philosophies or philosophical sciences, some theoretical (mathematical, natural, theological), and some practical (ethics, politics). The way to the starting-points of these, as to those of all sciences, is aporematic. But the philosophies themselves—at any rate, insofar as they are or are like genuine Aristotelian sciences—are presumably structures of demonstrations from starting-points. But that means that their methodology, when it isn’t dialectical, is simply that of such sciences. Dialectic, in other words, is not just the method of aporematic philosophy, but has a claim to being regarded as the distinctive method of Aristotelian philosophy generally. Notes 1. Hippocrates’ argument is described in homas Heath, A History of Greek Mathematics Vol. I (Oxford: Clarendon Press, 1921), pp. 183–201. 2. It is unclear just what Bryson’s method is. See Heath, A History of Greek Mathematics, pp. 223–25. 3. See Heath, A History of Greek Mathematics, pp. 221–22, citing Simplicius. 07_Shields_Ch07.indd 168 1/6/2012 6:01:10 PM OUP UNCORRECTED PROOF – FIRST-PROOF, 01/06/12, NEWGEN aristotle’s philosophical method 169 4. As Robert Bolton, ‘he Epistemological Basis of Aristotelian Dialectic,’ in Biologie, Logique et Métaphysique Chez Aristote, Daniel Devereux and Pierre Pellegrin, eds. (Paris: Éditions du Centre National de la Recherche Scientiique, 1990), pp. 215–17, convincingly argues. 5. I have translated paradoxos using the neologism ‘contradoxical’ to make clear that what is paradoxos in the relevant sense is not what we mean by ‘paradoxical’. 6. Compare Daniel Devereux, ‘Comments on Robert Bolton’s ‘he Epistemological Basis of Aristotelian Dialectic’,’ in Biologie, Logique et Métaphysique Chez Aristote, p. 272 n. 18. 7. Compare Jonathan Barnes, ‘Aristotle’s heory of Demonstration,’ in Articles on Aristotle. Vol. 1, Jonathan Barnes, Malcolm Schoield, and Richard Sorabji, eds. (London: Duckworth, 1975), pp. 80–1, and Devereux, ‘Comments on Robert Bolton’s ‘he Epistemological Basis of Aristotelian Dialectic’,’ pp. 272–73 n. 19. 8. Robert Bolton, ‘Deinition and Scientiic Method in Aristotle’s Posterior Analytics and Generation of Animals,’ in Philosophical Issues in Aristotle’s Biology, Allan Gotthelf and James Lennox, eds. (Cambridge: Cambridge University Press, 1987), pp. 122–23, conlates (a) and (d) when he claims that if an ‘expert biologist with new empirical data were not yet so lucky as to stand among the most acclaimed biologists neither he nor anyone else would be entitled to use his new results in dialectical argument no matter how empirically well-grounded they might be.’ 9. I discuss these sciences in Substantial Knowledge: Aristotle’s Metaphysics (Indianapolis: Hackett, 2000), pp. 258–60. 10. In Soph. fr. 1 (Ross), Aristotle says that Zeno invented dialectic. Zeno, too, saw the importance of examining ‘the consequences that follow from the hypothesis, not only if each thing is hypothesized to be, but also if that same thing is hypothesized not to be’ (Plato, Prm. 135d-136a). 11. Contrast Bolton, ‘he Epistemological Basis of Aristotelian Dialectic.’ 12. Also APr I 30 46a17–22, DC II 12 291b31–292a3, III 7 306a14–17, DA I 1 402b21–403a2, GA II 8 747b27–748a14, III 10 760b27–33. 13. In many texts, as here, Aristotle characterizes aporiai as knots aporematic philosophy enables us to untie (Phys. VIII 3 253a31–3, 8 263a15–18, Met. VII 6 1032a6–11, EN VII 2 1146a24–7). In others, he characterizes such philosophy as enabling us to make things—including starting-points—clear (APr I 30 46a17–30, DA II 2 413a11–13). 14. Robert Bolton, ‘Aristotle’s Conception of Metaphysics As a Science,’ in Unity, Identity and Explanation in Aristotle’s Metaphysics, T. Scaltsas, D. Charles, and M. L. Gill, eds. (Oxford: Clarendon Press, 1994), pp. 327–28, argues that the inal clause should instead be translated: ‘When it comes to those matters which (irst) philosophy deals with, dialectic should use its special peirastic form or capacity.’ Translated in this way, he continues, it ‘does not in the least require that when it deals with philosophical subjects dialectic merely probes or tests or criticizes but does not establish or lead one to know anything.’ True. But it collapses the distinction Aristotle is trying to draw between philosophy and dialectic, and conlicts with Top. VIII 14 163b9–16, which tells us unequivocally that dialectic alone cannot reach the truth. Bibliography Barnes, J. (1975) ‘Aristotle’s heory of Demonstration’, in J. Barnes, M. Schoield, and R. Sorabji, eds., (London: Duckworth), 65–87. 07_Shields_Ch07.indd 169 1/6/2012 6:01:11 PM OUP UNCORRECTED PROOF – FIRST-PROOF, 01/06/12, NEWGEN 170 the framework of philosophy: tools and methods Barnes, J., Schoield, M., and Sorabji, R., eds., (1975) Articles on Aristotle. Vol. 1. (London: Duckworth). Bolton, R. (1987) ‘Deinition and Scientiic Method in Aristotle’s Posterior Analytics and Generation of Animals’. in A. Gotthelf and J. Lennox, eds., Philosophical Issues in Aristotle’s Biology. (Cambridge: Cambridge Univ. Press), 120–66. ———. (1990) ‘he Epistemological Basis of Aristotelian Dialectic’. in D. Devereux and P. Pellegrin, eds., Biologie, Logique et Métaphysique Chez Aristote. (Paris: Éditions du Centre National de la Recherche Scientiique), 185–236. ———. (1994) ‘Aristotle’s Conception of Metaphysics As a Science’, in T. Scaltsas, D. Charles, and M.L. Gill, eds.,Unity, Identity and Explanation in Aristotle’s Metaphysics. (Oxford: Clarendon Press), 321–54. ———. (2009) ‘Two Standards for Inquiry in Aristotle’s De Caelo’. in A.C. Bowen and C. Wildberg, eds., New Perspectives on Aristotle’s De Caelo (Leiden: Brill), 51–82. ———. (2010) Science, Dialectique et Ethique chez Aristote: Essais D’Épistémologie Aristotélicienne (Louvain-La-Neuve: Éditions Peeters). Bowen, A.C. and Wildberg, C. (2009) New Perspectives on Aristotle’s De Caelo (Leiden: Brill). Brunschwig, J. (1967) Aristote: Topiques I-IV. (Paris: Les Belles Lettres). ———. (1984–5) ‘Aristotle on Arguments Without Winners or Losers’, Wissenschatskolleg Jahrbuch, 31–40. ———. (1990) ‘Remarques sur la Communication de Robert Bolton’ in D. Devereux ,and P. Pellegrin, eds., Biologie, Logique et Métaphysique Chez Aristote, 237–62. ———. (2007) Aristote: Topiques I-IV. (Paris: Les Belles Lettres). Canto-Sperber, M. and P. Pellegrin, eds. (2002) Le Style de La Pensée: Receueil de Texts en Homage à Jacques Brunschwig. (Paris: Les Belles Lettres). Devereux, D. (1990) ‘Comments on Robert Bolton’s “he Epistemological Basis of Aristotelian Dialectic”,’. in D. Devereux and P. Pellegrin, eds., pp. 263–86. Deveraux, D. and Pellegrin, P., eds. (1990) Biologie, Logique et Métaphysique Chez Aristote. (Paris: Éditions du Centre National de la Recherche Scientiique). Dorion, L.A. (1995) Les Réfutations Sophistiques d’Aristote. (Paris: Vrin). ———. (2002) ‘Aristote et L’Invention de la Dialectique’, in M. Canto-Sperber and P. Pellegrin, eds., Le Style de La Pensée: Receueil de Texts en Homage à Jacques Brunschwig. (Paris: Les Belles Lettres), 182–220. Gotthelf, A. and J. Lennox, eds. (1987) Philosophical Issues in Aristotle’s Biology (Cambridge: Cambridge Univ. Press). Heath, T. (1921) A History of Greek Mathematics. Vol. I. (Oxford: Clarendon Press). Irwin, T.H. (1988) Aristotle’s First Principles (Oxford: Clarendon Press). Nussbaum, M.C. (1986) he Fragility of Goodness (Cambridge: Cambridge Univ. Press). Owen, G.E.L. (1986) Logic, Science, and Dialectic: Collected Papers in Greek Philosophy. (Ithaca, N.Y.: Cornell Univ. Press). ———. ed. (1968) Aristotle on Dialectic: he Topics. (Oxford: Clarendon Press). Reeve, C.D C. ———. (2000) Substantial Knowledge: Aristotle’s Metaphysics. (Indianapolis, Ind.: Hackett). ———. (2012) Action, Contemplation, and Happiness: An Essay on Aristotle (Cambridge, Mass.: Harvard Univ. Press). Scaltsas, T., D. Charles, and M.L. Gill, eds. (1994) Unity, Identity and Explanation in Aristotle’s Metaphysics (Oxford: Clarendon Press). Sim, M., ed. (1999) From Puzzles to Principles? (Lanham: Lexington Books). Smith, R. (1997) Aristotle Topics Books I and VII (Oxford: Clarendon Press). 07_Shields_Ch07.indd 170 1/6/2012 6:01:11 PM