De Sophisticis Elenchis

Introduction

I. The Place of the Topica in the Organon

Both the Topica and the De Sophisticis Elenchis have always been regarded as genuine works of Aristotle. The two treatises are closely connected; the De Sophisticis Elenchis is an appendix to the Topica and its final section forms an epilogue to both treatises; indeed Aristotle himself seems sometimes to regard the two as forming a single work, since he twice quotes the De Sophisticis Elenchis under the title of the Topica.

It is generally admitted that what we call logic and Aristotle himself calls analytic was an early preoccupation of the philosopher and a direct outcome of discussions on scientific method held in the Platonic Academy. Plato himself, however, never attempted a formal treatment of the subject and the theories put forward, for example, in the Theaetetus, Sophist, Parmenides and Politicus were never developed into a regular system. But while Aristotle's systematic treatment of the process of inference and, above all, his discovery of the syllogism owe little to Plato, it has been generally recognized that the Platonic dialogues contain some of the germs from which the Aristotelian system was afterwards developed; for example, in the Theaetetus the doctrine of the categories is already implicit in the recognition of the abstract notions of substance, quality, quantity, relation, activity and passivity.

Of the logical treatises of Aristotle, which since about A.D. 200 have passed under the title of the Organon or 'instrument' of science, the most important are (1) the Prior Analytics, in which he sets forth the doctrine of the syllogism in its formal aspect without reference to the subject-matter with which it deals, (2) the Posterior Analytics, in which he discusses the characteristics which reasoning must necessarily possess in order to be truly scientific, (3) the Topica, in which he treats of the modes of reasoning, which, while syllogistically correct, fall short of the conditions of scientific accuracy. The Categories and the De Interpretatione are subsidiary treatises dealing, in the main, with the term and the proposition.

A great deal of time and ingenuity has been expended, particularly by German scholars, in an attempt to fix the exact order in which the various treatises which constitute the Organon were composed. The problem is complicated by the fact the treatises, in the form in which they have come down to us, seem to consist of rough notes, which were evidently subjected to a certain amount of revision due to the modification and development of his original doctrines. This process has naturally given rise to minor inconsistencies such as would naturally occur if corrections were made or additions inserted which were not completely adapted to the context in which they were placed.

It has been generally recognized that the whole of the Topica does not belong to the same date. H. Maier^a holds that the oldest portion consists of Books II–VII. 2 and that it was written under the direct influence of the Academy and belongs to the same period as the Aristotelian Dialogues, which have survived only in fragments; in particular, he points out that the term συλλογισμός is not used in the technical sense which it afterwards acquired (or, if it is used in that sense, e.g. in 130 a 7, it is a late insertion), whereas in the second half of Book VII the term is used in its well-known Aristotelian sense, and that, consequently, Books II–VII. 2 were composed before the philosopher made his greatest contribution to logic. He holds that Books I and VIII belong to the same period as Book VII. 4–5, and form an introduction and conclusion to the treatise written after the discovery of the syllogism and that the De Sophisticis Elenchis was a subsequent addition to the Topica. On the other hand, F. Solmsen^a and P. Gohlke^a hold that Books I–VII form the earlier portion of the work and that Book VIII and the De Sophisticis Elenchis were added subsequently.

As regards the relation of the Topica to the rest of the Organon, Maier considers the Topica as a whole to be earlier than the Analytics; Solmsen suggests that the order was Topica I–VII, (2) Posterior Analytics I, (3) Topica VIII and De Sophisticis Elenchis, (4) Posterior Analytics II, (5) Prior Analytics; Gohlke holds that the traditional order of the two Analytics is correct, and that the Topica and De Sophisticis Elenchis presuppose the Analytics.

In short, there is general agreement that the bulk of the Topica embodies Aristotle's earliest contribution to the systematic study of logic and that it was written in part before his discovery of the syllogism.

II. The Content of the Topica

The purpose of the Topica is, in the words of its author (100 a 18 ff.), 'to discover a method by which we shall be able to reason from generally accepted opinions about any problem set before us and shall ourselves, when sustaining an argument, avoid saying anything self-contradictory'; that is to say, it aims at enabling the two participants, the 'questioner' and the 'answerer,' to sustain their parts in a dialectical discussion. The subject, then, of the treatise may be described as the dialectical syllogism based on premises which are merely probable as contrasted with the demonstrative, or scientific, syllogism, which is the subject of the Posterior Analytics and is based on premises which are true and immediate. The probable premises which make up the dialectical syllogism are described (100 b 21 f.) as 'those which commend themselves to all or to the majority or to the wise.' The uses of dialectic are, we are told, three in number, (1) for mental training, (2) for general conversation, and (3) for application to the sciences, because (a) if we can argue a question pro and con, we shall be in a better position to recognize truth and falsehood, and (b) since the first principles of the sciences cannot be scientifically demonstrated, the approach to them must be through the study of the opinions generally held about them.

After the general introduction in Book I, Aristotle, in Books II–VII. 3, gives a collection of the τόποι which give their name to the treatise. The term τόποι is somewhat difficult to define. They may be described as 'commonplaces' or argument or as general principles of probability which stand in the same relation to the dialectical syllogism as axioms stand to the demonstrative syllogism; in other words, they are 'the pigeon-holes from which dialectical reasoning is to draw its arguments.'^a

Books II and III deal with the problems of accident; Books IV and V with those of genus and property; Books VI and VII. 1–3 with those of definition. Books VII. 4–5 and Book VIII, after giving some additional notes, conclude the treatise by describing the practice of dialectical reasoning.

III. The De Sophisticis Elenchis

Just as Aristotle treats of the demonstrative and the dialectical syllogism in the Posterior Analytics and the Topica, respectively, so in this treatise, which forms a kind of appendix to the Topica, he deals with the sophistical syllogism. A knowledge of this is part of the necessary equipment of the arguer, not in order that he may himself make use of it but that he may avoid it, and that the unwary may not be ensnared in the toils of sophistical argument; in fact, Aristotle is carrying on the Socratic and early-Platonic tradition by attacking the Sophists, who taught the use of logical fallacy in order to make the worst cause appear the better.

The term ἔλεγχος is strictly applied to the confutation of an actual adversary, but it is also used more widely of the confutation of an imaginary opponent. The treatise is, in fact, a study of fallacies in general, which are classified under various headings and fall into two main classes, those which depend on the language employed and those which do not. Some of these fallacies would hardly deceive the most simple minds; others, which Aristotle seems to have been the first person to expose and define, are capable not only of deceiving the innocent but also of escaping the notice of arguers who are employing them.

After two introductory chapters the work naturally falls into two parts, chapters 3–15, the refutation of fallacies, and chapters 16–33, the solution of fallacies, while chapter 34 forms an epilogue to the work.

IV. The Manuscripts

The chief manuscripts for the Topica and De Sophisticis Elenchis are:

A	Urbinas 35	saec. ix–x ineunt.
B	Marcianus 201	an. 955
C	Coislinianus 170	saec. xi
D	Coislinianus 170	saec. xi
u	Basileensis F. 11.21	saec. xiv
C	Vaticanus 1024	'satis vetustus'
P	Vaticanus 207	'non recens'
f	Marcianus App. IV. 5	saec. xiv
q	Ambrosianus M. 71	saec. xv
N	Laurentianus 72. 18	saec. xv
i	Laurentianus 72. 15	saec. xiv
T	Laurentianus 72. 12	saec. xiii
O	Marcianus 204	saec. xiv

Of these A and B are in a class by themselves. Bekker preferred A, Waitz B; the Teubner Editors give a slight preference to B, the readings of which are sometimes supported by papyrus fragments. C sometimes preserves the true reading.

V. Select Bibliography

Editions

Translations

In French:

In German:

Articles and Dissertations

In translating the Topica and De Sophisticis Elenchis I have used the text of Bekker in the Berlin Edition, and when I translate any other reading this is noted at the foot of the page. I have constantly referred to the Teubner text of Strache-Wallies, which does not, however, seem to me to mark any considerable advance on that of Bekker. I have found Waitz's edition of the Organon of great use, and the Latin version of Pacius is often helpful. I have frequently consulted the Oxford translation by W. A. Pickard-Cambridge. For the De Sophisticis Elenchis the notes and paraphrase in Poste's edition are often enlightening, though I cannot always agree with his interpretation.

My aim in translating has been to represent Aristotle's meaning as closely and faithfully as I can in simple English without resorting to paraphrase or trying to express it in modern terminology.

I have to thank my friend and former colleague Professor W. S. Macguinness, of King's College, London, for reading through my version and giving me the benefit of his fine scholarship and accuracy. He has suggested several improvements in the text which I have been glad to adopt.

Aristotle on Sophistical Refutations

I. Let us now treat of sophistical refutations, that is, arguments which appear to be refutations but are really fallacies and not refutations, beginning, as is natural, with those which come first.

That some reasonings are really reasonings, but that others seem to be, but are not really, reasonings, is obvious. For, as this happens in other spheres from a similarity between the true and the false, so it happens also in arguments. For some people possess good physical condition, while others have merely the appearance of it, by blowing themselves out and dressing themselves up like the tribal choruses^a; again, some people are beautiful because of their beauty, while others have the appearance of beauty because they trick themselves out. So too with inanimate things; for some of these are really silver and some gold, while others are not but only appear to our senses to be so; for example, objects made of litharge^b or tin appear to be silver, and yellow-coloured objects appear to be gold. In the same way also reasoning and refutation are sometimes real and sometimes not, but appear to be real owing to men's inexperience; for the inexperienced are like those who view things from a distance. Reasoning is based on certain statements made in such a way as necessarily to cause the assertion of things other than those statements and as a result of those statements; refutation, on the other hand, is reasoning accompanied by a contradiction of the conclusion. Some refutations do not affect their object but only appear to do so; this may be due to several causes, of which the most fertile and widespread division is the argument which depends on names. For, since it is impossible to argue by introducing the actual things under discussion, but we use names as symbols in the place of the things, we think that what happens in the case of the names happens also in the case of the things, just as people who are counting think in the case of their counters. But the cases are not really similar; for names and a quantity of terms are finite, whereas things are infinite in number; and so the same expression and the single name must necessarily signify a number of things. As, therefore, in the above illustration, those who are not clever at managing the counters are deceived by the experts, in the same way in arguments also those who are unacquainted with the power of names are the victims of false reasoning, both when they are themselves arguing and when they are listening to others. For this reason, therefore, and for others which will be mentioned hereafter, there exist both reasoning and refutation which appear to be genuine but are not really so. But since in the eyes of some people it is more profitable to seem to be wise than to be wise without seeming to be so (for the sophistic art consists in apparent and not real wisdom, and the sophist is one who makes money from apparent and not real wisdom) it is clear that for these people it is essential to seem to perform the function of a wise man rather than actually to perform it without seeming to do so. To take a single point of comparison, it is the task of the man who has knowledge of a particular subject himself to refrain from fallacious arguments about the subjects of his knowledge and to be able to expose him who uses them. Of these functions the first consists in being able to give a reason, the second in being able to exact one. It is essential, therefore, for those who wish to play the sophist to seek out the kind of argument which we have mentioned; for it is well worth his while, since the possession of such a faculty will cause him to appear to be wise, and this is the real purpose which sophists have in view.

It is clear, then, that a class of arguments of this kind exists, and that those whom we call sophists aim at this kind of faculty. Let us next discuss what are the various kinds of sophistical arguments and what are the various component parts of this faculty, and into what different divisions the treatment of the subject falls, and all the other elements which contribute to this art.

II. Of arguments used in discussion there are four kinds, Didactic, Dialectical, Examination-arguments and Contentious arguments. Didactic arguments are those which reason from the principles appropriate to each branch of learning and not from the opinions of the answerer (for he who is learning must take things on trust). Dialectical arguments are those which, starting from generally accepted opinions, reason to establish a contradiction. Examination-arguments are those which are based on opinions held by the answerer and necessarily known to one who claims knowledge of the subject involved (in what manner, have been described elsewhere^a. Contentious arguments are those which reason or seem to reason from opinions which appear to be, but are not really, generally accepted. Demonstrative arguments have been treated in the Analytics, and dialectical arguments and examinations have been dealt with elsewhere.^b Let us now deal with competitive and contentious arguments.

III. We must first of all comprehend the various objects at which those aim who compete and contend in argument. They number five; refutation, fallacy, paradox, solecism, and, fifthly, the reduction of one's opponent to a state of babbling, that is, making him to say the same thing over and over again; or, if not the reality, at any rate the appearance of each of these things. Their first choice is a plain refutation, their second to show that their opponent is lying, their third to lead him on a paradox, their fourth is to make him commit a solecism (that is, to make the answerer, as a result of the argument, speak ungrammatically), and, lastly, to make him say the same thing over and over again.

IV. There are two modes of refutations; one has to do with the language used, the other is unconnected with the language. The methods of producing a false illusion in connexion with language are six in number: equivocation, ambiguity, combination, division, accent and form of expression. The truth of this can be verified by induction and by syllogistic proof based on this (though some other assumption is also possible), that this is the number of ways in which we can fail to indicate the same thing by the same terms or expressions. Arguments such as the following are based on equivocation: 'Those who know, learn; for it is those who know the use of letters that learn^a what is dictated to them.' Here 'learn' is equivocal, meaning 'understand by using knowledge' and 'acquire knowledge.' Or again, 'Evils are good, for what must exist is good, and evil must exist.' Here 'must exist' is used in two senses; it means 'what is necessary,' which is often true of evils (for some evil is necessary), and we also say that good things 'must exist.'^b Or again, 'the same man is seated and standing and is a sick man and restored to health; for it is the man who stood up that is standing, and it is he who was recovering his health that is restored to health, but it was the man who was seated that stood up and the man who was sick that was recovering.' For that 'the sick man' does such and such a thing or has such and such a thing done to him, has not one meaning only but at one time means 'the man who is now sick,' and at another time 'the man who was formerly sick.' But it was the sick man who began to recover his health when he was actually sick, but he is in good health when he is not sick and is not the sick man now but the man who was formerly sick. The following examples are connected with ambiguity: 'To wish me the enemy to capture,' and 'when a man knows something, surely there is knowledge of this'; for it is possible by this expres­sion to signify both the knower and the thing known as knowing.^c And 'what a man sees, surely that (he)^a sees: a man a pillar sees, therefore the pillar sees.' Again, 'Surely you insist on being what you insist on being. You insist on a stone being: there­fore, you insist on being a stone.' Again 'Surely speaking is possible of the silent.' 'Speaking of the silent' can also be taken in two ways, either that the speaker is silent or the things spoken of are silent. There are three modes connected with equivocation and ambiguity: (1) when the expres­sion or name properly signifies more than one thing, such as ἀετός^b and κύων,^c (2) when we customarily use a word in more than one sense, (3) when a word has more than one meaning in combination with another word, though by itself it has only one mean­ing, for example, 'knowing letters'; for it may so happen that taken separately 'knowing' and 'letters' have only one meaning, but taken together they have more than one meaning, namely, either that the letters themselves have knowledge or that someone else has knowledge of the letters.

Ambiguity and equivocation then take these forms. The following examples are connected with the combination of words, for instance, 'A man can walk when sitting and write when not writing.' The significance is not the same if one utters the words separately^d as it is if one combines them, namely, 'a man can walk-while-sitting,'^e and, similarly, in the other example, if one combines the words and says 'a man can write-when-not-writing,' for it means that he can write and not write at the same time; whereas if one does not combine the words it means that, when he is not writing, he has the power to write. Again, 'He now understands letters, since he has understood what he knows'^a; and further, 'One single thing being able to carry, many things you can carry.'^b

The following propositions are connected with divi­sion: '5 is 2 and 3,' '5 is odd and even,' 'the greater is equal to the less,' for it is so much and something more.^c For the same sentence divided would not always seem to have the same meaning as when taken as a whole, for example, 'Free I made thee a slave'^d and 'goodly Achilles left a hundred (and) fifty men.'^e

It is not easy to construct an argument relating to accent in discussions which are not written down, but it is easier in written matter and poetry. For example, some people emend Homer to meet the ob­jection of critics that his phrase 'τὸ μὲν οὗ καταπύθεται ὄμβρῳ' is a strange one.^f For they solve the difficulty by a change of accent, pronouncing the ου more sharply.^g Also in the passage about Agamemnon's dream^h they say that Zeus himself did not say, 'But we grantⁱ him to secure the fulfilment of his prayer' but bade the dream to grant it.^j Such examples, then, depend on accentuation.

Refutations which depend on the form of expres­sion occur when what is not the same is expressed in the same form; for example, when the masculine is expressed by the feminine or vice versa, or the neuter by the masculine or feminine; or again when a quality is expressed by a quantity or vice versa, or the active by a passive or a state by the active, and so forth according to the distinctions previously made.^a For it is possible for something which is not of the nature of an action to signify by the language used something which is of the nature of an action; for example, to 'flourish' is a form of expression like to 'cut' or to 'build'; yet the former denotes a quality and a certain disposition, the latter an action. So too with the other possible examples.

Refutations, then, connected with language are based on these commonplaces. Of fallacies un­connected with language there are seven kinds: (1) those connected with Accident; (2) those in which an expression is used absolutely, or not abso­lutely but qualified as to manner or place or time or relation; (3) those connected with ignorance of the nature of refutation; (4) those connected with the consequent; (5) those connected with the assump­tion of the original point to be proved; (6) those which assert that what is not a cause is a cause; (7) the making of several questions into one.

V. Fallacies connected with Accident occur when it is claimed that some attribute belongs similarly to the thing and to its accident; for since the same thing has many accidents, it does not necessarily follow that all the same attributes belong to all the predicates of a thing and to that of which they are predicated. For example, 'If Coriscus is different from "man" he is different from himself, for he is a man'; or 'if he is different from Socrates, and Socrates is a man,' they say that it has been admitted that Coriscus is different from a man, because it is an accident that the person from which he said that Coriscus is different is a man.

Fallacies connected with the use of some particular expression absolutely or in a certain respect and not in its proper sense, occur when that which is pre­dicated in part only is taken as though it was predi­cated absolutely. For example, 'If that-which-is-not is an object of opinion, then that-which-is-not is'; for it is not the same thing 'to be something' and 'to be' absolutely. Or again, 'That-which-is is not, if it is not one of the things which are, e.g. if it is not a man.' For it is not the same thing 'not to be some­thing' and 'not to be' absolutely; but, owing to the similarity of the language, 'to be something' appears to differ only a little from 'to be,' and 'not to be something' from 'not to be.' In like manner when something is predicated in a certain respect and absolutely; for example, 'If an Indian, being lack all over, is white in respect of his teeth, then he is white and not white.' Or if both attributes belong in a certain respect, they say that the contrary attributes belong simultaneously. In some cases this sort of fallacy can be easily perceived by anyone; if, for example, after securing an admission that the Ethiopian is black, one were to ask whether he is white in respect of his teeth, and then, if he be white in this respect, were to think that he had finished the interrogation and had proved dialectically that he was both black and not black. In some cases, on the other hand, the fallacy escapes detection, namely, where, when an attribute is ascribed in some respect only, an absolute attribution would also seem to follow, and where it is not easy to see which of the attributes can be properly assigned. An instance of this occurs when both the opposite attributes belong similarly; for then it is generally held that it must be conceded that either both or neither can be predicated absolutely; for example, if something is half white and half black, is it white or black?

Other fallacies arise because no definition has been given of what a syllogism is and what a refutation, and there is some defect in their definition. For a refutation is a contradiction of one and the same predicate, not of a name but of a thing, and not of a synonymous name but of an identical name, based on the given premisses and following necessarily from them (the original point at issue not being included) in the same respect, relation, manner and time. A false statement about something also occurs in the same manner. Some people, however, appear to refute, omitting some of the above-named points, showing, for example, that the same thing is double and not double, because two is the double of one but not the double of three. Or, they show that if the same thing is double and not double of the same thing, yet it is not double in the same respect; for it is double in length but not double in breadth. Or, if it is double and not double of the same thing and in the same respect and manner, yet it is not so at the same time; and so there is only an apparent refutation. One might, indeed, force this fallacy also into the category of those connected with language.

Fallacies connected with the assumption of the original point to be proved arise in the same manner and in the same number of ways as it is possible to beg the original point; they have an appearance of achieving a refutation because men fail to perceive at the same time what is the same and what is different.

The refutation connected with the consequent is due to the idea that consequence is convertible. For whenever, if A is, B necessarily is, men also fancy that, if B is, A necessarily is. It is from this source that deceptions connected with opinion based on sense-perception arise. Tor men often take gall for honey because a yellow colour accompanies honey; and since it happens that the earth becomes drenched when it has rained, if it is drenched, we think that it has rained, though this is not necessarily true. In rhetorical arguments proofs from signs are founded on consequences; for, when men wish to prove that a man is an adulterer, they seize upon the consequence of that character, namely, that the man dresses him­self elaborately or is seen wandering abroad at night—facts that are true of many people, while the accusation is not true. So, too, in dialectical reason­ings; for example, the argument of Melissus that the universe is infinite assumes that the universe has not come into being (for nothing could come into being from what does not exist) and that everything which has come into being has come from a beginning; if, therefore, the universe has not come into being, it has no beginning and therefore is infinite. But this does not necessarily follow; for even if what has come into being always has a beginning, anything that has a beginning need not have come to be, any more than it follows that a man who is hot must be in a fever because a man who is in a fever is hot.

The refutation connected with taking as a cause what is not a cause, occurs when that which is not a cause is foisted into the argument as though the refutation were due to it. Such a case occurs in reasonings leading up to an impossibility; for in these one is bound to destroy one of the premisses. If, therefore, what is not a cause is enumerated among the questions which are necessary for the production of the resultant impossibility, the refutation will often seem to come about as the result of it; for example, in the argument that 'soul' and 'life' are not identical. For if coming-into-being is contrary to perishing, then a particular kind of coming-into-being will be contrary to a particular kind of perishing; now death is a particular kind of perishing and con­trary to life; life, therefore, is a coming-into-being and to live is to come-into-being. But this is im­possible; and so the soul and life are not identical. But this conclusion is not the result of reasoning; for the impossibility occurs even if one does not assert that life is identical with the soul but merely says that life is contrary to death, which is a perishing, and that coming-into-being is contrary to perishing. Such arguments are not absolutely inconclusive but only inconclusive as regards the point at issue, and the questioners themselves are often equally uncon­scious of such a state of affairs.

Such, then, are the arguments connected with the consequent and the falsely imputed cause. Those which are connected with the union of two questions in one occur, when it is not noticed that they are more than one and one answer is given as though there was only one question. Sometimes it is easy to see that there is more than one question and that an answer should not be given, for example, when it is asked 'Is the earth sea, or is the sky?' Sometimes, however, it is less easy, and thinking that there is only one question, people either give in by not answering the question or suffer an apparent refutation. For example, 'Is A and is B a man?' 'If so, if a man strikes A and B, he will strike a man, not men?' Or again, 'Where part is good and part evil, is the whole good or evil?' Either answer might possibly seem to involve an apparent refuta­tion or false statement; for to say that something is good when it is not good or not good when it is good is a false statement. Sometimes, however, if certain premisses are added, there might be a genuine refutation. For example, if one agrees that a single thing and a number of things are alike called 'white' or 'naked' or 'blind.' For if 'blind' is used of something which does not possess sight though it is its nature to possess it, it will also describe a number of things which do not possess sight though it is their nature to possess it. When, therefore, one thing has sight while another has not, they will either both be able to see or both be blind; which is impossible.

VI. We must either divide apparent reasonings and refutations in the manner just described or else refer them all to a false conception of refutation, making this our basis; for it is possible to resolve all the kinds of fallacy which we have mentioned into violations of the definition of refutation. Firstly, we must see if they are inconclusive; for the conclusion ought to follow from the premisses laid down, so that we state it of necessity and do not merely appear to do so. Next, we ought to see if they accord with the remaining parts of the definition. For of the fallacies connected with language, some are due to a double meaning, for example equivocation and ambiguous phraseology and similarity of formation (for it is customary to indicate everything as a particular substance), whereas composition, division and accen­tuation are due to the phrase not being the same or the name different. For the name also, like the thing signified, ought to be the same, if refutation or reasoning is to result. For example, if the subject is a mantle, you should come to a conclusion about a mantle, not about a cloak; for the latter con­clusion is also a true one, but the reasoning is not complete, and a further question must be asked to prove that words mean the same, thing, if the answerer asks how you have refuted him.

Fallacies connected with Accident become obvious when 'proof' has been defined. For the same definition ought to be true also of refutation, except that 'the contradictory' is added; for refutation is a proof of the contradictory. If, therefore, there is no proof of the accident, no refutation takes place. For if, when A and B are, C is, and C is white, it does not necessarily follow that it is white because of the syllogism. And again, if the triangle has its angles equal to two right angles, and it happens to be a figure, element or principle, it does not necessarily follow that because it is a figure, element or principle it has this character; for the demonstration is con­cerned with it not qua figure or qua element but qua triangle. And so likewise with the other instances. Thus, if refutation is a kind of proof, an argument depending on an accident could not be a refutation. Yet it is along these lines that specialists and men of science in general are refuted by the unscientific; for they argue with the men of science with reasonings based on accident, and the latter, being incapable of making distinctions, either give in when questioned, or think that they have done so when they have not.

Fallacies which depend on whether a statement is made in a limited sense or absolutely occur be­cause the affirmation and denial are not of the same thing. For 'not partly white' is the negation of 'partly white,' and 'not absolutely white' of 'abso­lutely white.' If, then, one takes the admission that something is partially white to mean that it is abso­lutely white, he does not cause a refutation but only seems to do so owing to ignorance of what a refuta­tion is.

The clearest fallacies of all are those already mentioned^a as connected with the definition of refutation (hence also their name)^b; for the semblance of a refutation is due to the defect in the definition, and, if we distinguish fallacies in this way, we must put down 'defect of definition' as common to all these cases.

Fallacies due to assuming the original point and stating as a cause what is not a cause are clearly exposed by means of the definition. For the con­clusion ought to follow because this and that is so, which is not the case when the alleged cause is not the cause; and, again, the conclusion should follow without the original point being included, which is not true of arguments based on the begging of the original point.

Fallacies connected with the consequent form part of those due to accident; for the consequent is an accident but differs from the accident because the accident can be secured in the case of a single thing by itself, for example, a yellow thing and honey are identical, and so is a white thing and a swan, whereas the consequent always exists in more than one thing; for we claim that things which are the same as one and the same thing are the same as one another; and this is how refutation proceeds when the con­sequent is involved. It is not, however, always true, for example, in the case of accidental whiteness; for both 'snow' and 'swan' are the same in respect of whiteness.^a Or again, as in the argument of Melissus,^b someone takes 'to have come into being' and 'to have a beginning' as the same thing, and 'to become equal' as the same thing as 'to take on the same magnitude.' For because what has come into being has a beginning, he claims also that what has a beginning has come into being, on the ground that 'having come into being' and 'being finite' are both the same thing, because both have a begin­ning. Similarly, too, in the case of things which become equal, he assumes that, if things which take on one and the same magnitude become equal, then also things which become equal take on the same magnitude. In doing so he is assuming the conse­quent. Since, then, the refutation where accident is concerned depends on ignorance of the nature of refutation. so also, it is clear, does the refutation where the consequent is concerned. But we must examine this question from other points of view also.^c

Fallacies connected with the union of several questions in one are due to our failure to differentiate or distinguish the definition of the term 'proposi­tion.' For a proposition is a single predication about a single subject. For the same definition applies to 'one single thing' and to 'the thing' simply; the definition, for example, of 'man' and of 'one single man' is the same, and so, too, with the other instances. If, therefore, a 'single proposition' is one which claims a single predicate for a single sub­ject, a 'proposition,' simply, will also be a question of this kind. And since reasoning is based on pro­positions, and refutation is a process of reasoning, refutation will also be based on propositions. If, therefore, a proposition is a single predication about a single thing, clearly this fallacy also depends on ignorance of the nature of refutation; for what is not a proposition appears to be one. If, therefore, a man has given an answer as though to a single question, there will be a refutation, but if he has not given it but only appears to have done so, there will be only an apparent refutation. Thus all the kinds of fallacy fall under the heading of ignorance of the nature of refutation—those connected with language because the contradiction, which is a particular char­acteristic of refutation, is only apparent, and the rest because of the definition of reasoning.

VII. In fallacies connected with verbal equivoca­tion and ambiguous phrases the deception arises from the inability to distinguish the various meanings of a term (for there are some which it is not easy to distin­guish, for example, the meanings of 'unity,' 'being' and 'identity'). In fallacies connected with combina­tion and disjunction the deception is due to the supposi­tion that it makes no difference whether the term is combined or disjoined, as indeed is generally the case. So, too, in those connected with accentuation; for it does not seem ever, or seems very seldom, to alter the significance of the word whether it is pronounced with a lower or a higher pitch. In fallacies connected with the form of expression the deception is due to similarity of language; for it is difficult to distinguish what sort of things belong to the same and what to different categories; for he who can do this very nearly approaches a vision of the truth. What in particular seduces us into giving our assent to the fallacy is the fact that we suppose that every predi­cate of something is an individual thing and it pre­sents itself to our ears as a single thing; for it is to the one and to substance that 'individuality' and 'being' are generally held most truly to be attached. On this account also this kind of fallacy must be classed among those connected with language; firstly, because the deception occurs more commonly when we are inquiring with others than by ourselves (for an inquiry with someone else is carried on by means of words, whereas in our own minds it is carried on quite as much by means of the thing itself); secondly, because, even in solitary inquiry, a man is apt to be deceived when he carries on his inquiry by means of words; and, thirdly, the deception arises from the similarity, and the similarity arises from the language. In fallacies connected with accident the deception is due to inability to distinguish the identical and the different, the one and the many, and what kinds of predicates have all the same accidents as their sub­ject. So, too, in fallacies connected with the con­sequent; for the consequent is a branch of the accident. Furthermore, in many cases it appears to be true and is treated as axiomatic that, if A is inseparable from B, then also B is inseparable from A. In fallacies connected with the defect in the definition of refutation and with the distinction between a qualified and an absolute statement the deception is due to the minuteness of the difference; for we regard the qualification of a particular case or respect or manner or time as having no extra signifi­cance and concede the universality of the proposition. So, too, when people assume the original point and when the wrong cause is assigned and when several questions are united in one; for in all these cases the deception is due to the minuteness of the differ­ence; for we fail accurately to carry out the definition of 'proposition' and 'reasoning' from the above-mentioned cause.

VIII. Since we know the various sources from which apparent reasonings arise, we also know those from which sophistical reasonings and refutations would arise. By sophistical refutation and reasoning I mean not only the seeming but unreal reasoning or re­futation but also one which, though real, only seems to be, but is not really, germane to the subject in hand. These are those which fail to refute and show up ignorance within the sphere of the subject in hand, and this is the function of examination. Now this is a department of dialectic, but it may reach a false con­clusion owing to the ignorance of the person under examination. But sophistical refutations, even if they rove the contradictory of his view, do not make clear whether he is ignorant; for men try to entrap even the man of scientific knowledge by these arguments.

That we know them by the same method is clear; for the same reasons which make the hearers think that a conclusion has been reached as a result of questions, would make the answerer think so too, so that there will be false proofs as a result of all or some of these causes; for what a man thinks he has conceded without being questioned, he would grant if he were to be questioned. But of course it some­times happens that, as soon as we ask the requisite question, we make the falsehood obvious, as happens in verbal fallacies and those due to solecism. If, therefore, false proofs of the contradictory depend on the apparent refutation, it is clear that proofs of false conclusions must be also due to the same number of causes as the apparent refutation. Now the apparent refutation depends on the elements which compose a genuine one; for, if any one of these is lacking, there would only be an apparent refutation, for example, that which is due to the conclusion not resulting from the argument (the reduction to an impossibility), and that which unites two questions in one and is due to a fault in the pro­position, and that which is due to the substitution of an accident for the essence of a thing, and—a subdivision of the last mentioned—that which is due to the consequent; moreover, there is the case where the result follows in word only and not in reality, and also where, instead of the contradiction being uni­versal and in the same respect, relation and manner, there is a restriction in extent or in connexion with another of these qualifications; and then again there is the case of the assumption of the original point due to a disregard of the principle of not reckoning it in. Thus we should know the various conditions under which false proofs occur, for there are no further conditions under which they could occur, but they will always result from the above causes.

A sophistical refutation is not an absolute refuta­tion but is relative to some person, and so likewise is a sophistical proof. For unless the refutation which depends on equivocation assumes that the equivocal term has only a single meaning, and unless that which depends on similarity of termination assumes that there is only substance, and so on, neither refutation nor proof will be possible, either absolutely or relatively, to the answerer; whereas, if they do make these assumptions, they will be possible relatively to the answerer, but not absolutely; for they have not secured a statement which has a single meaning but only one which appears to be such, and only from a particular person.

IX. Without a knowledge of everything which exists we ought not to try and grasp the various ways in which the refutation of those who are refuted is brought about. This, however, is not the function of any art; for the sciences are possibly infinite, and so clearly demonstrations are also infinite. Now there are true as well as false refutations; for wherever demonstration is possible, it is possible also to refute him who maintains the contradictory of the truth; for example, if a man maintains that the diagonal of a square is commensurate with its sides, one should refute him by proving that it is incommensurate. So we shall need to have scientific knowledge of everything; for some refutations will depend on the principles of geometry and their conclusions, others on those of medicine, and others on those of the other sciences. Moreover, false refutations also are among things which are infinite; for every art has a false proof peculiar to it, geometry a geometrical proof and medicine a medical proof. By 'peculiar to an art' I mean 'in accordance with the principles of that art.' It is clear, then, that we need not grasp the commonplaces of all refutations but only those which concern dialectic; for these are common to every art and faculty. And it is the function of the scientific man to examine the refutation which is peculiar to each science and see whether it is apparent only and not real, or, if it is real, why it is so; whereas it is the function of dialecticians to examine a refutation which depends on common principles which do not fall under any one art. For if we know the sources of generally accepted proofs about any particular subject, we know also the sources of the refutations; for a refutation is a proof of a contradictory, and so one or two proofs of a contradictory make up a refutation. We know, then, the various sources of all such proofs, and, knowing these, we also know their solutions; for the objections to these are the solutions. We also know the various sources of apparent refutations—apparent, that is, not to everyone but only to a certain kind of mind; for it would be an endless task to examine the various ways in which they are apparent to the man in the street. It is, therefore, clear that it is the function of the dialectician to be able to grasp the various ways in which, on the basis of common principles, a real or apparent refutation, that is, dialectical or apparently dialectical or part of an examination, is brought about.

X. No real distinction, such as some people propose, exists between arguments used against the word and those used against the thought; for it is absurd to suppose that some arguments are used against the word and others against the thought, and not the same in both cases. For what is failure to use the argument against the thought except what happens when a man does not apply the term in the meaning about which the man questioned thought that he was being questioned when he made the concession? And this is equivalent to using it against the word; whereas to use it against the thought is to apply it to the sense about which the man was thinking when he made the concession. If, then, when the word has more than one meaning, both the questioner and the man questioned were too think that it had only one meaning—as, for example, 'unity' and 'being' have several meanings but both the answerer answers and the questioner puts his question on the supposition that there is only one meaning and that the argument is that all things are one—will the argument have been directed against the word and not rather against the thought of the man questioned? If, on the contrary, one of them thinks that the word has several meanings, obviously the argument is not directed against the thought. For application to the word and application to the thought belong primarily to arguments which signify several things ambiguously, but, secondarily, to any argument whatsoever; for the application to the thought does not depend on the argument but on a certain attitude of mind in the answerer towards what has been conceded. Next, it is possible for all arguments to be applied to the word; for in the case under discussion 'to be applied to the word' means 'not to be applied to the thought.' For if all are not applied to the word or the thought, there will be a third class not applied to either; but they declare that the classification is exhaustive and divide them into those applied to the word and those applied to the thought, and there is no other class. But, as a matter of fact, reasonings dependent on the word are amongst those dependent on a multiplicity of meanings. For it is an absurd statement that 'dependent on the name' describes all arguments connected with language. The truth is that there are some false arguments which do not depend on a particular attitude of mind on the part of the answerer towards them but are due to the fact that the argument itself involves the kind of question which can bear more than one meaning.

It is quite absurd to discuss refutation without previously discussing proof; for refutation is a proof, and so we ought to discuss proof before discussing false refutation; for such refutation is an apparent proof of a contradiction. Therefore the cause of falsity will lie either in the proof or in the contradiction (for the contradiction must be added), but sometimes in both, if there be a merely apparent refutation. In the argument that 'the silent speaks,' the refutation lies in the contradiction, not in the proof; in the argument that 'a man can give away what he has not got,' it lies in both; in the argument that 'Homer's poetry is a figure' because it forms a 'cycle,' it lies in the proof. The argument that errs in neither respect is a true proof.

But to resume from the point whence the argument digressed,^a Are mathematical arguments always applied to the thought or not? If anyone thinks that the term 'triangle' has several meanings and has granted it in a sense other than a figure which he has proved to contain two right angles, has the questioner reasoned against the answerer's thought or not?

Further, if the name has several meanings but the answerer does not think or imagine that this is so, has not the questioner reasoned against his thought? Or how else must the question be asked except by offering a distinction? In which case one will ask, 'Is it or is it not possible for a man to speak when silent, or is the answer in one sense "No," in another "Yes"?' But if the answerer were to refuse to grant the possibility in any sense and the questioner were to argue that it is possible, has he not argued against the thought of his opponent? Yet the argument is generally regarded as among those connected with the name; there is not, therefore, any class of argument which is directed against the thought. Some arguments are directed against the name, and such arguments are not all of them even apparent refutations, still less true refutations. For there are also apparent refutations which are not connected with language, for example, amongst others, those connected with accident.

But if one claims to make distinctions, saying, 'By "the silent speaking" I mean sometimes one thing and sometimes another,' this claim is, in the first place, absurd (for sometimes the question does not seem to involve any ambiguity, and it is impossible to make a distinction where no ambiguity is suspected); and, secondly, what else will didactic argument be but this? For it will make clear the position to one who neither has considered nor knows nor conceives that a second meaning is possible. For why should not the same process be used where there is no double meaning? 'Are the units in four equal to the twos? Bear in mind that the twos are contained in one sense in one way and in another sense in another way.' Again, 'Is the knowledge of contraries one or not? Notice that some contraries are knowable, others are not.' Thus the man who makes this claim seems not to know that didactic is one thing and dialectic another, and that the man who employs didactic should not ask questions but himself make things clear, while the dialectician asks questions.

XI. Further, to demand that the answerer should either affirm or deny is not the function of one who is displaying something but of one who is making an examination. For the art of examination is a kind of dialectic and has in view not the man who knows but the man who is ignorant and pretends to know. The man, then, who views general principles in the light of the particular case is a dialectician, while he who only apparently does this is a sophist. Now one form of contentious and sophistic reasoning is reasoning which is only apparent, with which dialectic deals as a method of examination, even though the conclusion be true; for it is deceptive in the matter of cause. Then there are those false reasonings which do not accord with the method of inquiry peculiar to the subject yet seem to accord with the art concerned. For false geometrical figures are not contentious (for the resultant fallacies accord with the subject-matter of the art), and the same is the case with any false figure illustrating something which is true, for example, Hippocrates' figure or the squaring of the circle by means of lunules.^a On the other hand, Bryson's method of squaring the circle, even though this be successful, is nevertheless sophistical, because it does not accord with the subject-matter concerned. And so any merely apparent reasoning on these topics is a contentious argument, and any reasoning which merely appears to accord with the subject-matter, even though it be genuine reasoning, is contentious argument; for it only apparently accords with the subject-matter and so is deceptive and unfair. For just as unfairness in an athletic contest takes a definite form and is an unfair kind of fighting, so contentious reasoning is an unfair kind of fighting in argument; for in the former case those who are bent on victory at all costs stick at nothing, so too in the latter case do contentious arguers. Those, then, who behave like this merely to win a victory, are generally regarded as contentious and quarrelsome, while those who do so to win a reputation which will help them to make money are regarded as sophistical. For, as we have said,^a the art of the sophist is a money-making art which trades on apparent wisdom, and so sophists aim at apparent proof. Quarrelsome people and sophists use the same arguments, but not for the same reasons; and the same argument will be sophistical and contentious but not from the same point of view. If the semblance of victory is the motive, it is contentious; if the semblance of wisdom, it is sophistical: for sophistry is an appearance of wisdom without the reality. The contentious arguer bears much the same relation to the dialectician as the drawer of false geometrical figures bears to the geometrician; for he reasons falsely on the same basis as the dialectician, while the drawer of false figures argues on the same basis as the true geometrician. But the latter is not a contentious reasoner, because he constructs his false figure on the principles and conclusions which come under the art of geometry, whereas the former, arguing on principles which come under dialectic, will clearly be contentious on the other subjects. For example, the squaring of the circle by means of lunules is not contentious, whereas Bryson's method is contentious. It is impossible to transfer the former outside the sphere of geometry because it is based on principles which are peculiar to geometry, whereas the latter can be used against many disputants, namely, all those who do not know what is possible and what impossible in any particular case; for it will always be applicable. And the same is true of the way in which Antiphon used to square the circle.^a Or, again, if someone were to deny that it is better to take a walk after dinner because of Zeno's argument,^b it would not be a medical argument; for it is of a general application. Accordingly, if the contentious argument stood in every respect in the same relation to the dialectical as the constructor of false figures stands to the geometrician, there would be no contentious argument on those topics. But, as it is, dialectical argument has no definite sphere, nor does it demonstrate anything in particular, nor is to of the nature of the universal. For there is no genus which includes all things, and, if there were, it would not be possible for them to come under the same principles. So no art which aims at showing the nature of anything proceeds by interrogation; for it is impossible to grant either one of two portions of the question; for a proof cannot result from both of them. Dialectic, however, does proceed by interrogation, whereas, if it aimed at showing something, it would refrain from questions, if not about everything, at any rate about primary things and particular principles; for if the opponent refused to grant these, dialectic would no longer have any basis on which to argue against the objection. Dialectic is at the same time an art of examination; for neither is the art of examination of the same nature as geometry but it is an art which a man could possess even without any scientific knowledge. For even a man without knowledge of the subject can examine another who is without knowledge, if the latter makes concessions based not on what he knows nor on the special principles of the subject but on the consequential facts, which are such that, though to know them does not prevent him from being ignorant of the art in question, yet not to know them necessarily involves ignorance of it. Clearly, therefore, the art of examination is not knowledge of any definite subject, and it therefore follows that it deals with every subject; for all the arts employ also certain common principles. Accord­ingly, everyone, including the unscientific, makes some kind of use of dialectic and the art of examina­tion; for all, up to a certain point, attempt to test those who profess knowledge. Now this is where the common principles come in; for they know these of themselves just as well as the scientists, even though their expression of them seems to be very inaccurate. Thus they all practise refutation; for they perform unmethodically the task which dialectic performs methodically, and the man who carries out an examination by means of an art of reasoning is a dialectician. Now there are many identical principles in every sphere, but these are not such as to have a particular nature and form a particular class—resembling, in this respect, negations—while others are not of this kind but limited to special spheres; it is, therefore, possible by means of these to hold ex­aminations on every subject, and that there can be an art of doing this, though not of the same kind as the demonstrative arts. For this reason the contentious arguer is not in all respects in the same position as the constructor of a false geometrical figure; for the contentious arguer will not reason falsely on principles of a definite class but will deal with every kind. These, then, are the modes of sophistical refuta­tions. It is easy to see that to investigate them and to be able to apply them is the task of the dialectician; for the method of dealing with propositions constitutes the whole of this study.

XII. We have now dealt with apparent refutations. As for showing that the answerer is stating a fallacy and leading the argument towards a paradox—for this was the second aim of the sophist—this is, in the first place, best achieved by some kind of inquiry and by questioning. For to ask a question without defining it in relation to a subject laid down is a good method of hunting out things of this sort; for people are more likely to fall into error when they speak at random, and they speak at random when they have no definite subject set before them. Also to ask a number of questions, even though the point against which one is arguing is defined, and to demand that the answerer should say what he thinks, gives ample opportunity of leading a man into a paradox or fallacy, and also, if, when asked, he says 'yes' or 'no' to any of the questions, of leading him to topics on which one has abundant material for attacking him. This unfair method, however, is much less practicable than formerly; for people demand, 'What has this to do with the original question?' An elementary rule for obtaining a fallacious or paradoxical statement is not to put any thesis directly but to pretend that one is asking from a desire to learn; for this method of inquiry gives an opening for attack.

A special method of showing up a fallacy is the sophistical method, namely, to lead one's opponent to the kind of statements against which one has plenty of arguments; it will be possible to do this in a right and in wrong way, as has already been said.^a

Again, to elicit a paradox, you should see to what school the person who is discussing with you belongs, and then question him on some pronouncement of that school which most people regard as paradoxical; for every school has some tenet of this kind. An elementary rule in this connexion is to have a ready-made collection of the theses of the different schools among your propositions. The proper solution here too is to make it clear that the paradox does not result because of the argument; now your opponent always desires that this should be so.

Furthermore, you should seek for paradoxes in men's wishes and professed opinions. For they do not wish the same things as they declare that they wish, but they give utterance to the most becoming sentiments, whereas they desire what they think is to their interest. They declare, for example, that a noble death ought to be preferred to a pleasurable life and honourable poverty to discreditable wealth; but their wishes are the opposite of their words. He, therefore, whose statements agree with his wishes must be led to express the opinions usually professed, and he whose statements agree with the latter must be led to state the opinions usually hidden; for in both cases they must necessarily fall into paradox, for they will contradict either their professed or their secret opinions.

A commonplace rule which makes men utter para­doxes in abundance is the application of the standards of nature and law, which Callicles is represented as applying in the Gorgias^a and which all the ancients regarded as valid; for according to them Nature and Law are opposites, and justice is a good thing accord­ing to law but not according to nature. Therefore, to a man who speaks in terms of nature you must reply in terms of law, and when he speaks in terms of law you must lead the argument to terms of nature; for in both cases the result will be that he utters paradoxes. In the view of the ancients what accorded with nature was the truth, while what accorded with law was the general opinion of mankind. It is, there­fore, clear that they also, like the men of to-day, tried to refute the answerer or to make him utter paradoxes.

Some questions involve a paradox whichever way they are answered; for example, 'Ought one to obey the wise or one's father?' and, 'Ought one to do what is expedient or what is just?' and 'Is it preferable to suffer or to inflict a wrong?' You ought to lead men to opinions opposed to those of the majority and of the wise—if a man speaks as trained arguers do, you should lead him to opinions opposed to the majority; if he speaks as do the majority, to opinions opposed to expert reasoners. For some say that the happy man is necessarily just, but in the view of the majority it is paradoxical that a king should not be happy. To lead a man to paradoxes of this kind is the same thing as to bring him into opposition to the standards of nature and law; for law is the opinion of the majority, but the utterances of the wise accord with the standards of nature and truth.

XIII. It is, then, by these commonplace rules that you should seek to obtain paradoxes. Next, as to making people babble, we have already said what we mean by this term.^a Arguments of the following kind all have this end in view; 'If it makes no difference whether one uses the term or the definition of it, and "double" and "double of half" are the same thing, then if "double" is "double of half," it will be "double of half of half"; and if "double of half" be substituted again for "double," there will be a triple repetition, "double of half of half of half."' Again, 'Is not" desire" "desire of pleasure?" Now "desire is an appetite for pleasure": therefore "desire is an appetite for pleasure of pleasure."'

All arguments of this kind take place (a) when relative terms are used, where not only the genera but the terms themselves are relative and are ren­dered in relation to one and the same thing (for example, appetite is appetite for something, and desire is desire of something, and double is double of something, namely, double of half), and (b) where terms are used of which, though they are not relative at all, the substance (namely, the things of which they are states or affections or the like) is indicated in their definition, since they are predicated of these things. For example, 'odd' is a 'number which has a middle unit,' and an 'odd number' exists, therefore an 'odd number' is 'number-that-has-a middle-unit number.' Again, if 'snubness' is 'concavity of the nose,' and there is a 'snub nose,' then a 'snub nose' is a con­cave-nose nose.'

Men sometimes appear to induce 'babbling' when they do not really do so, because they do not further inquire whether 'double' used by itself has a signifi­cation or no, and, if it has, whether the same or a different one, but they appear to draw the conclusion immediately. It appears, however, to have the same signification also because the word is the same.

XIV. What solecism is has already been stated.^a It is possible to commit it, and not to commit it, yet to seem to do so, as well as to commit it, yet seem not to do so. If, as Protagoras used to say, μῆνις (wrath) and πήλνξ (helmet) are masculine, according to him, he who calls wrath a 'destructress' (οὐλομένην) commits a solecism, though he does not appear to anyone else to do so,^b but he who calls it a 'destructor' (οὐλόμενην) appears to commit a solecism but does not do so. It is obvious, therefore, that one might produce this effect by art also; therefore many arguments appear to infer a solecism, when they do not really do so, as happens also with refutations.

Almost all apparent solecisms occur owing to the word 'this' or 'it' (τόδε) and when the inflection denotes neither the masculine nor the feminine but the neuter. 'He' (οὗτος) denotes a masculine, 'she' (αὕτη) a feminine, whereas 'this' or 'it' (τοῦτο), though meaning to signify a neuter, often signifies either a masculine or a feminine. For example, 'What is this (τοῦτο)?' 'It is Calliope,' or 'It is a log' or 'It is Coriscus.' The case-forms of the masculine and feminine are all different, but some of those of the neuter are different and others not. Often, therefore, when 'it' (τοῦτο) has been granted, people argue as if 'him' (τοῦτον) had been used, and they similarly use another case in place of some other. The false reasoning arises because 'it' (τοῦτο) is common to more than one case; for it signifies sometimes 'he' (οὗτος) and sometimes 'him' (τοῦτον). It ought to signify them alternately; with the indicative 'is' (ἐστι) it ought to signify the nominative 'he' (οὗτος); with the infinitive 'to be' (εἶναι) it ought to signify 'him' (τοῦτον), for example, 'It is Coriscus,' '[I believe] it to be Coriscus.' So likewise with feminine nouns and with so-called articles of use, which can have either a masculine or a feminine designation; for only those which end in -ον have the designation which belongs to an article of use, e.g., ξύλον (log), σχοινίον (rope). Those which do not take this form have a masculine or a feminine termination, and some of these we apply to articles of use; for example, ἀσκός (wine-skin) is masculine and κλίνη (bed) is feminine. Therefore, in such cases there will be the same difference when the indicative 'is' (ἔστι) is used and the infinitive 'to be'(εἶναι). Also, in a way, solecism resembles the kind of refutation which is due to the use of similar terms for dissimilar things^a; for as in the one case it happens that we commit a solecism in the category of actual things, so in the other we commit it in that of names; for 'man' and 'white' are both names and things.

Clearly, then, we must try and argue up to a solecism on the basis of the above-mentioned case-forms.

These are the branches of competitive arguments and their sub-divisions, and the above are the methods of employing them. Now it makes no small difference whether the accompaniments of the question are arranged in a certain way with a view to concealment, as in dialectics. Therefore, as a sequel to what has been said above, we must first treat of this subject.

XV. To effect a refutation one expedient is length; for it is difficult to keep many things in view simul­taneously. To produce length the above-mentioned elementary rules must be employed. One resource is speed; for when people lag behind they see less far ahead. Further, there are anger and contentious­ness; for when people are agitated they are always less capable of being on their guard. Elementary rules for rousing anger are to make it plain that one wishes to act unfairly and to behave in an altogether shameless manner. Another device is to put one's questions alternately, whether one has several argu­ments leading up to the same point or whether one has arguments proving both that this is so and that this is not so; for the result is that the answerer is on his guard at the same time against either several or contrary attacks. In a word, all the resources for concealment mentioned before^a are also useful against competitive arguments; for concealment is for the purpose of escaping detection, and escape from detection is for the purpose of deception.

When dealing with those who refuse to consent to anything which they think is in favour of your argument, you must put your question in a negative form, as though you wanted the opposite of what you really want, or, at any rate, as if you were asking your question with indifference; for people are less troublesome when it is not clear what one wants to secure. Often, when in dealing with particulars a man grants the individual case, you ought not, in the process of induction, to make the universal the subject of your question but assume that it is granted and use it accordingly; for sometimes people think that they have themselves granted it and appear to their hearers to have done so, because they recall the process of induction and think that the question would not have been asked without some object. Where there is no term to signify the universal, you should nevertheless use the resemblance of the particulars^a for your advantage; for the resemblance often passes unnoticed. Also, in order to secure your premiss, you should contrast it with its contrary in your question. For example, if you want to secure the premiss that one ought to obey one's father in all things, you should ask whether one should obey one's parents in all things or disobey them in all things. If you want to establish that the multiplica­tion of a number many times over results in a large number, you should ask whether it should be con­ceded that it is a large or that it is a small number; for, if pressed, one would rather that it should seem to be large. For the juxtaposition of contraries increases the quantity and quality of things, both relatively and absolutely, in the eyes of men.

Often the most sophistical of all frauds practised by questioners produces a striking appearance of refutation, when, though they have proved nothing, they do not put the final proposition in the form of a question but state conclusively, as though they had proved it, that 'such and such a thing, then, is not the case.'

Another sophistical trick is, when the thesis is a paradox, to demand, when the generally accepted view is originally proposed, that the answerer should reply what he thinks about it, and to put one's question in some such form as 'Is that your opinion?' For, if the question is one of the premisses of the argument, either a refutation or a paradox must result. If he grants the premiss, there will be a refutation; if he refuses to grant it and even denies that it is the generally accepted view, he utters a paradox; if he refuses to grant it but admits that it is the generally accepted view, there will be the appearance of a refutation.

Moreover, as in rhetorical arguments, so likewise also in refutations, you ought to look for contradic­tions between the answerer's views and either his own statements or the views of those whose words and actions he admits to be right, or of those who are generally held to bear a like character and to re­semble them, or of the majority, or of all mankind. Also, just as answerers, when they are being refuted, often draw a distinction, if they are on the point of being refuted, so questioners also ought sometimes, when dealing with objectors, if the objection is valid against one sense of the word but not against another, to resort to the expedient of declaring that the opponent has taken it in such and such a sense, as Cleophon does in the Mandrobulus.^a They ought also to withdraw from the argument and cut short their other attacks, while the answerer, if he perceives this move in time, should raise anticipatory objections and get his argument in first. One should also sometimes attack points other than the one mentioned, excluding it if one can make no attack on the position laid down, as Lycophron did when it was suggested that he should deliver an encomium on the lyre. To those who demand that one should take some definite point of attack (since it is generally held that one ought to assign the object of a question, whereas if certain statements are made the defence is easier), you should say that your aim is the usual result of refutation, namely, to deny what your opponent affirmed and affirm what he denied, and not to prove that the knowledge of contraries is the same or not the same. One should not ask the conclusion in the form of a proposition, and some propositions should not be asked at all but treated as admitted.

XVI. We have now dealt with the sources of ques­tions and how they ought to be asked in competitive arguments. We must next treat of answering, and how solutions are brought about, and what are their sub­jects, and for what purpose such arguments are useful. They are useful for philosophy for two reasons. In the first place, as they generally turn on language, they put us in a better position to appreciate the various meanings which a term can have and what similarities and differences attach to things and their names. Secondly, they are useful for the questions which arise in one's own mind; for he who is easily led astray by another person into false reasoning and does not notice his error, might also often fall into this error in his ow mind. A third and last reason is that they establish our reputation, by giving us the credit of having received a universal training and of having left nothing untried; for that one who is taking part in an argument should find fault with arguments with­out being able to specify where their weakness lies, rouses a suspicion that his annoyance is apparently not in the interests of truth but due to inexperience.

How answerers should meet such arguments is obvious if we have adequately described above^a the sources of false arguments and distinguished the fraudulent methods of questioning. To take an argu­ment and see and disentangle the fault in it is not the same thing as to be able to meet it promptly when one is asked a question. For we often fail to recognize something which we know when it is pre­sented in a different form. Furthermore, as in other spheres a greater degree of speed or slowness is rather a question of training, so in argument also; therefore, even though something may be clear to us, yet, if we lack practice, we often miss our opportuni­ties. The same thing happens sometimes as with geo­metrical diagrams; for there we sometimes analyse a figure but cannot reconstruct it; so too in refuta­tions we know how the argument is strung together, but we are at a loss how to take it to pieces.

XVII. In the first place, then, just as we say that we ought sometimes deliberately to argue plausibly rather than truthfully, so too we ought sometimes to solve questions plausibly rather than according to truth. For, generally speaking, when we have to fight against contentious arguers, we ought to regard them not as trying to refute us but as merely appear­ing to do so; for we deny that they are arguing a case, so that they must be corrected so as not to appear to be doing so. For if refutation is unequi­vocal contradiction based on certain premisses, there can be no necessity to make distinctions against ambiguity and equivocation; for they do not make up the proof. But the only other reason for making further distinctions is because the conclusion looks like a refutation. One must, therefore, beware not of being refuted but of appearing to be so, since the asking of ambiguities and questions involving equi­vocation and all similar fraudulent artifices mask even a genuine refutation and make it uncertain who is refuted and who is not. For when it is possible in the end, when the conclusion is reached, to say that one's opponent contradicted what he asserted only by means of an equivocation, however true it may be that he happened to be tending in the same direction, it is uncertain whether a refutation has taken place; for it is uncertain whether he is speaking the truth now. If, however, one had made a distinction and questioned the equivocal or ambiguous term, the re­futation would not have been uncertain. Also, the object of contentious arguers—though it is less their aim in these days than formerly—would have been carried out, namely, that the person questioned should answer 'Yes' or 'No'; as it is, however, because the questioners put their questions im­properly, the person questioned is obliged to add something in his answer by way of correcting the unfairness of the proposition, since, if the questioner makes adequate distinctions, the answerer must say either 'Yes' or 'No.'

If anyone is going to imagine that an argument which rests on equivocal terms is a refutation, it will be impossible for the answerer to avoid being refuted in a certain sense; for in dealing with visible things one must necessarily deny the term which he asserted and assert that which he denied. For the correction which some people suggest is useless. For they do not say that Coriscus is musical and unmusical, but that this Coriscus is musical and this Coriscus is un­musical. For it will be making use of the same expression to say that this Coriscus is unmusical (or musical) as to say that this Coriscus is so; and one is affirming and denying this at the same time. But perhaps it does not mean the same thing; for neither did the name in the former case; so what is the difference? But if he is going to assign to the one person the simple appellation 'Coriscus,' while to the other he adds 'a certain' or 'that,' it is absurd; for the addition belongs no more to the one than to the other; for it makes no difference to whichever of the two he adds it.

However, since, if one does not distinguish the meanings of a doubtful term, it is not clear whether he has been confuted or not, and since the right to draw distinctions is conceded in arguments, it is obvious that to grant the question simply, without making distinctions, is a mistake; so that, even if the man himself does not appear to be refuted, yet his argument certainly appears to be so. It frequently happens, however, that, though people see the ambiguity, they hesitate to make the distinction, because of the numerous occasions on which people propose subjects of this kind, in order to avoid seeming to be acting perversely all the time. Then, again, though people would never have thought that the argument would hinge upon this point, they are often confronted with a paradox. So, since the right to draw a distinction is conceded, we must not hesitate to use it, as was said before.

If one does not make two questions into one, the fallacy which depends on equivocation and ambiguity would not exist either, but either refutation or absence of refutation. For what is the difference between ask­ing whether Callias and Themistocles are musical and asking the same question about two people both with the same name? For if one indicates more things than one, one has asked more questions than one. If, therefore, it is not correct to demand simply to receive one answer to two questions, clearly it is not proper to give a simple answer to any equivocal question, even though the term is true of all the subjects, as some people claim that one ought. For this is just the same as asking 'Are Coriscus and Callias at home or not at home?,' whether they are both at home or not there; for in both cases the number of propositions is more than one. For if the answer is true, it does not follow that the question is a single one. For it is possible that it is true to say 'yes' or 'no' when asked a countless number of questions; but, for all that, one ought not to answer them with a single reply, for that means the ruin of discussion. This resembles the case of the same name being applied to different things. If, therefore, one must not give one answer to two questions, it is obvious that neither should one say 'yes' or 'no' where equivocal terms are used; for then the speaker has not given an answer but made a statement, but it is regarded in a way as an answer amongst those who argue, because they do not realize what is the result.

As we said, then, since there are some seeming refutations which are not really refutations, in like manner also there are some seeming solutions which are not really solutions. These we say that we ought sometimes to bring forward in preference to true refutations in competitive argument and in meeting ambiguity. In the case of statements which appear to be true one must answer with the phrase 'granted'; for then there is the least likelihood of any accessory refutation; but if one is obliged to say something paradoxical, then in particular one must add that it seems so, for then there can be no appearance either of refutation or of paradox. Since it is clear what 'begging the original question' means and since people always consider that assumptions which lie near the conclusion must be demolished and that some of them must not be conceded on the ground that the opponent is begging the question, so when someone claims something of such a nature that it must necessarily follow from the thesis and it is false or paradoxical, we must use the same plea; for the necessary consequences are generally regarded as part of the same thesis. Furthermore, when the universal which has been obtained has no name but is indicated by a comparison only, we must say that the questioner takes it not in the sense in which it was granted nor as he proposed it; for a refutation often hinges on this point too.

When we are excluded from these expedients, we must have recourse to the plea that the argument has not been properly set forth, attacking it on the basis of the classification of fallacies given above.^a

When terms are used in their proper senses, one must answer either simply or by making a distinction. It is when our statement implies our meaning without expressing it—for example, when a question is not asked clearly but in a shortened form—that refutation ensues. For instance, 'Is whatever belongs to the Athenians a property of the Athenians?' 'Yes; and this is likewise true of everything else.' 'Well, then, does man belong to the animals?' 'Yes.' 'Then man is a property of the animals. For we say that man "belongs to" the animals because he is an animal, just as we say that Lysander "belongs to" the Laco­nians because he is a Laconian.' Obviously, there­fore, when the premiss is not clear, it must not be conceded simply.

When it is generally held that, if one of two things is true, then the other is necessarily true, but, if the second is true, the first is not necessarily true, when asked which is true, we ought to concede the less inclusive; for the greater the number of premisses, the more difficult it is to draw a conclusion. If the disputant tries to establish that A has a contrary while B has not, if his contention is true, we ought to say that both have a contrary but that no name is laid down for one of the two.

Regarding some of the statements which they make, most people would declare that anyone who did not concede them was lying, while they would not say so about others, for example, about subjects on which people disagree (for instance, most people have no decided opinion whether the soul of living creatures is destructible or immortal). Therefore, when it is uncertain in which sense the suggested premiss is generally used, whether as maxims are employed (for people call both true opinions and general affirmations by the name of 'maxims') or like the statement, 'the diagonal of a square is incommensurate with its sides,' and further, where the truth is a matter of uncertainty,—in these cases one has an excellent opportunity of changing the terms without being found out. For, because it is uncertain in which sense the premiss bears its true meaning, one will not be regarded as playing the sophist, and, because of the disagreement on the subject, one will not be regarded as lying; for the change will make the argument proof against refutation.

Furthermore, whenever one foresees any question, one must be the first to make one's objection and say what one has to say, for thus one can best disconcert the questioner.

XVIII. Since a correct solution is an exposure of false reasoning, indicating the nature of the question on which the fallacy hinges, and since 'false reasoning' can mean one of two things (for it occurs either if a false conclusion has been reached or if what is not a proof appears to be such), there must be both the solution described just now,^a and also the rectification of the apparent proof by showing on which of the questions it hinges. The result is that one solves the correctly reasoned arguments by demolishing them, the apparent reasonings by making distinctions. Again, since some correctly reasoned arguments are true, while others are false, in their conclusions, it is possible to solve those which are false in their conclusion in two ways, either by demolishing one of the questions or by showing that the conclusion is not as stated. Those arguments, on the other hand, which are false in their premisses can only be solved by the demolition of one of the premisses, since the conclusion is true. Those, therefore, who wish to solve an argument should observe, firstly, whether it has been correctly reasoned or is not reasoned, and, next, whether the conclusion is true or false, in order that we may achieve a solution either by making a distinction or by demolishing a premiss and doing so in one or other of the two ways just described.^b There is a very wide difference between solving an argument when one is being questioned and when one is not; for in the latter case it is difficult to see what is coming, but when one is at leisure it is easier to see one's way.

XIX. Of the refutations which hinge upon equivocation and ambiguity some involve a question which bears more than one sense, while others have a conclusion which can bear several meanings; for example, in the argument about 'the speech of the silent,' the conclusion has a double meaning, and in the argument that 'a man who knows is not conscious of what he knows,' one of the questions involves ambiguity. Also, that which has a double meaning is sometimes true and sometimes false, the term 'double' signifying that which is partly true and partly untrue.

When the diversity of meaning occurs in the conclusion, no refutation takes place, unless the questioner secures a contradiction beforehand, as, for example, in the argument about the 'seeing of the blind'; for there never was refutation without contradiction. Where the diversity of meaning occurs in the questions, there is no need to deny the ambiguity beforehand; for the argument is not directed towards it as a conclusion but carried on by means of it. At the beginning, therefore, one ought to reply to an ambiguous term or expression in the following manner, that 'in one sense it is so and in another it is not so'; for example 'the speaking of the silent' is possible in one sense but not in another. Or again, 'what needs must is to be done sometimes and not at other times'; for the term 'what needs must' can bear several meanings. If one does not notice the ambiguity, one should make a correction at the end by adding to the questioning: 'Is the speaking of the silent possible?' 'No, but speaking of this particular man when he is silent is possible.' So likewise also where the variety of meaning is contained in the premisses: 'Are not people conscious of what they know?' 'Yes, but not those who know in this particular way'; for it is not the same thing to say that is is not possible for those who know to be conscious of what they know and that those who know in a particular way cannot be conscious of their knowledge. Generally speaking, too, even though one's opponent argues in a straightforward manner, one must contend that what he has contradicted is not the actual fact which one affirmed but merely its name, and so there is no refutation.

XX. It is evident, too, how fallacies which turn on the division and combination of words should be solved; for, if the expression signifies something different when it is divided and when it is combined, when the opponent is drawing his conclusion we must take the words in the contradictory sense. All such expressions as the following turn upon the combination or division of words; 'Was so-and-so being beaten with that with which you saw him being beaten?' and 'Did you see him being beaten with that with which he was being beaten?'^a The argument here has something of the fallacy due to ambiguous questions, but it actually turns on combination. For what turns on the division of words is not really ambiguous (for the expression when divided differently is not the same), unless indeed ὄρος and ὅρος, pronounced according to the breathing, constitute a single word with different meanings. (In written language a word is the same when it is written with the same letters and in the same manner, though people now put in additional signs,^b but the words when spoken are not the same.) Therefore an expression whose meaning turns on division is not ambiguous, and it is clear also that all refutations do not turn upon ambiguity, as some people say.

It is for the answerer to make the division; for 'I-saw-a-man-being-beaten with my eyes' is not the same thing as to say 'I saw a man being-beaten-with-my-eyes.'—Then there is Euthydemus' saying, 'Do you know now in Sicily that there are triremes in Piraeus?^a—And, again, 'Can a good man who is a cobbler be bad?' 'No.' 'But a man who is good can be a bad cobbler; therefore he will be a good-bad cobbler.'—Again, 'Things of which the knowledge is good are good objects of learning, are they not?' 'Yes.' 'But the knowledge of evil is good; therefore evil is a good object of learning.' 'But, further, evil is both evil and an object of learning, so that evil is an evil object of learning; but it has already been seen that the knowledge of evils is good.'—'Is it true to say at the present moment you are born?' 'Yes.' 'Then you are born at the present moment.' Does not a different division of the words signify something different? For it is true to say-at-the-present-moment that you are born, but not to say you are born-at-the-present-moment.—Again, 'Can you do what you can and as you can?' 'Yes.' 'And when you are not playing the harp you have the power of playing the harp; and so you could play the harp when you are not playing the harp.' In other words, he does not possess the power of playing-when-he-is-not-playing, but he possesses the power of doing it when he is not doing it.

Some people solve this in another manner also. If he has granted that a man can do what he can do, they say that it does not follow that he can play the harp when he is not playing it; for it has not been granted that he will do it in every way in which he can,—for it is not the same thing to do it in the way in which he can and in every way in which he can. But clearly this solution is not a good one; for the solution of arguments which turn on an identical principle is identical, whereas this solution will not suit every argument nor every form of question into which it can be put, but is directed against the questioner, not against the argument.

XXI. Arguments do not arise owing to accentuation either in written or in spoken language, though a few might occur such as the following: A house is 'where you lodge' (οὗ καταλύεις), isn't it? Yes. Is not 'you do not lodge' (οὐ καταλύεις) the negation of 'you lodge' (καταλύεις)? Yes. But you said that 'where you lodge' (οὗ καταλύεις) was a house; therefore a house is a negation. It is obvious how this must been even solved; for the spoken word is not the same with the acuter and with the graver accent.^a

XXII. It is plain also how we must meet arguments that turn on the identical expression of things which are not identical, seeing that we possess the various kinds of categories. Suppose that one man when questioned has granted that something which denotes a substance is not an attribute, and another man has shown that something is an attribute which is in the category of relation or quantity but generally held, because of its expression, to denote a substance, as for example in the following argument: Is it possible to be doing and to have done the same thing at the same time? No. But it is surely possible to be seeing and to have seen the same thing at the same time and under the same conditions. Or again, Is any form of passivity a form of activity? No. Then 'he is cut,' 'he is burnt,' 'he is affected by a sensible object' are similar kinds of expression and all denote some form of passivity; and, on the other hand, 'to say,' 'to run,' and 'to see' are forms of expression similar to one another; but 'to see' is surely a way of being affected by a sensible object, so that passivity and activity occur at the same time. In the former case, if someone, after granting that it is impossible to be doing and to have done the same thing at the same time, were to say that it is possible to see thing and to have seen it, he has not yet been refuted supposing that he declares that seeing is a form not of activity but of passivity. For this further question is necessary, though he is supposed by the hearer to have granted it when he granted that 'to cut' is 'to be doing something' and 'to have cut' is 'to have done something,' and so with similar forms of expres­sion. For the hearer himself adds the rest, on the supposition that the significance is similar, whereas it is not really similar but only appears so owing to the expression. The same thing occurs as in fallacies of ambiguity; for in dealing with ambiguous terms the man who is not an expert in argument thinks that his opponent has denied the fact which he asserted, not the term, whereas yet another ques­tion needs to be asked, namely, whether he is using the ambiguous term with his eye upon one meaning only; for if he grants this, a refutation will be achieved.

Similar to the above are also the following argu­ments: Has a man lost what he had and afterwards has not? For he who has lost one die^a only will no longer have ten dice. Is not what really happens that he has lost something which he had before but no longer has, but it does not follow that he has lost the whole amount or number which he no longer has? In the question, therefore, he is dealing with that which he has, in the conclusion with the total number; for the number was ten. If, therefore, he had asked in the first place whether a man who formerly possessed a number of objects which he no longer possesses, has lost the total number of them, no one would have granted this, but would have said that he had lost either the total number or one of the objects. Again, it is argued that a man could give what he had not got; for what he has not got is one die only. Is not what really happens that he has not given that which he has not got but has given it in a manner in which he has not got it, namely, as a single unit? For 'single unit' does not denote either a particular kind of thing or a quality or a quantity but a certain relation to something else, namely, dissociation from anything else. It is, therefore, as though he had asked whether a man could give what he has not got, and on receiving the answer 'No,' were to ask whether a man could give something quickly when he had not got it quickly, and, on receiving the answer 'Yes,' were to infer that a man could give what he had not got. It is obvious that he has not drawn a correct inference; for 'giving quickly' does not denote giving a particular thing but giving in a particular manner, and a man could give something in a manner in which he did not get it; for example, he could get it with pleasure and give it with pain.

Similar also are all the following arguments: 'Could a man strike with a hand that he has not got or see with an eye that he has not got?' For he has not got only one eye. Some people, therefore, solve this by saying that the man who has more than one eye (or whatever it is) has also only one. There is also^a the argument of some people that 'what a man has, he has received': A only gave one pebble, and B has, they say, only one pebble from A.^b Other people argue by directly demolishing the question raised, saying that one can have what one has not received; for example, one can receive wine that is sound but have it in a sour condition if it has gone bad in the process of transfer. But, as was said before,^c all these people direct their solutions not to the argu­ment but to the man. For if this were a real solution, it would be impossible to achieve a solution by grant­ing the opposite, as happens in all other cases; for example, if 'it is partly so and partly not so' is the solution, an admission that the expression is used without qualification makes the conclusion valid; but if no conclusion is reached, there cannot be a solution. In the above examples, even though every­thing is conceded, yet we say that no proof has been effected.

Moreover, the following also belong to this class of arguments: 'If something is written, did someone write it?' It is written that 'you are sitting'; this is a false statement, but was true at the time when it was written; therefore what was written is at the same time false and true. No, for the falsity or truth of a statement or opinion does not denote a substance but a quality; for the same account applies to an opinion as to a statement. Again, 'Is what the learner learns that which he learns?' A man learns a slow march quick; it is not then what he learns that is meant but how he learns it. Again, 'Does a man trample on that through which he walks?' But he walks through the whole day. Was not what was meant not what he walks through but when he walks? Just as when we talk of a man drinking a cup, we refer not to what he drinks but to that out of which he drinks. Again, 'Is it not either by learning or by discovery that a man knows what he knows?' But, supposing that of two things he has discovered one and learnt the other, he has not either discovered or learnt the two taken together. Is it not true to say that what he knows is each single thing, but not all the things taken together? There is also the argument that there is a 'third man' beside 'man' and 'individual men.' This is not so, for 'man' and every generic term denotes not an individual substance but a quality or relation or mode or something of the kind. So, too, with the question whether 'Coriscus' and 'the musician Coriscus' are the same thing or different. For the one term denotes an individual substance, the other a quality, so that it is impossible to isolate it; for it is not the process of isolation which produces the 'third man' but the admission that there is an individual substance. For 'man' will not be an individual substance as Callias is, nor will it make any difference if one were to say that what is isolated is not an individual substance but a quality; for there will still be a one as con­trasted with the many, for instance 'man.' It is obvious, therefore, that it must not be granted that the term predicated universally of a class is an in­dividual substance, but we must say that it denotes either a quality or a relation or a quantity or some­thing of the kind.

XXIII. To sum up, in dealing with arguments which turn on language the solution will always de­pend on the opposite of that on which the argument turns; for example, if the argument turns on com­bination, the solution will be by division, if on division, by combination. Again, if it turns on acute accentua­tion, grave accentuation will be the solution, and vice versa. If it turns on equivocation, it can be solved by the use of the opposite term; for example, if it so happens that one says something is inanimate after having denied that it is so, one must show that it is animate; and, if one has said that it is inanimate and one's opponent has argued that it is animate, one must assert that it is inanimate. Similarly, too, in the case of ambiguity; if the argument turns on similarity of language, the opposite will be the solu­tion. 'Could one give what one has not got?' Surely not what he has not got but he could give it in a way in which he has not got it, for example, a single die^a by itself. 'Does a man know the thing which he knows by learning or discovery?' Yes, but not 'the things which he knows.' Also a man tramples on the thing through which he walks, not on the time through which he walks.^b And similarly, too, with the other instances.

XXIV. To meet arguments which turn upon acci­dent one and the same solution is universally appli­cable. It is undetermined on what occasions the attribute should be applied to the subject where it belongs to the accident, and sometimes it is generally held and stated to belong and sometimes it is denied that it necessarily belongs. We must, therefore, when a conclusion has been reached, assert in every case alike that it does not necessarily belong. But we must have an example to bring forward. All such arguments as the following turn on accident: 'Do you know what I am about to ask you?' 'Do you know the man who is coming towards us?' or 'the man with his face covered?' 'Is the statue your work?' or 'Is the dog your father?'^a 'Is the result of multiplying a small number by another small number itself a small number?' It is obvious that in all these instances it does not necessarily follow that the attribute which is true of the accident is also true of the subject. For it is only to things which are in­distinguishable and one in essence that all the same attributes are generally held to belong; but in the case of the good, it is not the same thing to be good and to be about to be the subject of a question.^b Nor in the case of 'the man who is coming towards us' (or 'with his face covered'), is 'to be coming towards us' the same thing as 'to be Coriscus'; so that, if I know Coriscus but do not know the man who is coming towards me, it does not follow that I know and do not know the same man. And again, if this is 'mine' and if it is also 'a piece of work,' it is not therefore 'a piece of my work' but may be my possession or chattel or something else. The other instances can be treated in the same way.

Some people obtain a solution by demolishing the thesis of the question; for they say that it is possible to know and not to know the same thing but not in the same respect; when, therefore, they do not know the man who is coming towards them but know Coriscus, they say that they know and do not know the same thing but not in the same respect. Yet in the first place, as we have already said,^c the method of correcting arguments which turn on the same principle ought to be identical, yet this will not be so, if one takes the same axiom to apply not to 'knowledge' but to 'existence' or 'being in a certain state'; for example, 'this dog is a father, this dog is yours.'^a Though it is sometimes true and it is possible to know and not to know the same thing, yet the suggested solution is quite inapplicable in the above instance. But there is no reason why the same argument should not contain several flaws, but it is not the exposure of every fault that forms a solution; for it is possible for a man to show that a false conclusion has been reached without showing on what point it turns, as, for instance, in Zeno's argument that motion is im­possible. Even, therefore, if one were to attempt to infer the impossibility of this view, he is wrong, even though he has given countless proofs; for this procedure does not constitute a solution, for a solu­tion is, as we saw, an exposure of false reasoning, showing on what the falsity depends. If, therefore, he has not proved his case or else if he attempts to draw an inference, whether true or false, by false means, the unmasking of this procedure is a solution. But perhaps, though in some cases there is nothing to prevent this happening, yet it would not be gener­ally admitted in the instances given above; for he knows that Coriscus is Coriscus and that what is coming towards him is coming towards him. But there are cases in which it is generally held to be possible to know and not to know the same thing; for instance, one can know that someone is white but be ignorant of the fact that he is musical, thus knowing and not knowing the same thing but not in the same respect; but as to what is coming towards him and Coriscus, he knows both that it is coming towards him and that he is Coriscus.

An error similar to that made by those whom we have mentioned is committed by those who solve the argument that every number is small; for if, when no conclusion has been reached, they pass over the fact and say that a conclusion has been reached and is true because every number is both large and small, they are committing an error.

Some people, too, solve these reasonings by the principle of ambiguity, saying, for example, that 'yours' means 'your father' or 'your son' or slave.' Yet it is obvious that, if the refutation turns upon the possibility of several meanings, the term or expression ought to be used literally in several senses; but no one speaks of A as B's child in the literal sense if B is the child's master, but the com­bination is due to accident. 'Is A yours?' 'Yes.' 'Is A a child?' 'Yes.' 'Then A is your child,' for he happens to be both yours and a child; but for all that he is not 'your child.'

There is also the argument that 'something "of evils" is good; for wisdom is a knowledge "of evils."' But the statement that this is 'of so-and-so'^a is not used with several meanings but denotes possession. Granting, however, that the genitive has more than one meaning (for we say that man is 'of the animals,' though not a possession of theirs), and if the relation of so-and-so to evils is expressed by the genitive, it is therefore a so-and-so 'of evils,' but so-and-so is not one of the evils. The difference seems to be due to whether the genitive is used in a particular sense or absolutely. Yet it is perhaps possible for the saying 'Something of evils is good' to be ambiguous, though not in the example given above, but rather in the phrase 'a slave is good of the wicked.' But per­haps this example is not to the point either; for if something is 'good' and 'of so-and-so,' it is not at the same time 'so and-so's good.' Nor is the state­ment that 'man is of the animals' used with several meanings; for a phrase does not acquire several senses every time we express its meaning in an ellip­tical form; for we express, 'Give me the Iliad' by quoting the half line 'Sing, goddess, the wrath.'

XXV. Arguments which turn upon the use of an expression not in its proper sense but with validity in respect only of a particular thing or in a particular respect or place or degree or relation and not ab­solutely, must be solved by examining the conclusion in the light of its contradictory, to see if it can possibly have been affected in any of these ways. For it is impossible for contraries and opposites and an affirma­tive and a negative to belong absolutely to the same subject; on the other hand, there is no reason why each should not belong in a particular respect or relation or manner, or one in a particular respect and the other absolutely. Thus if one belongs absolutely and the other in a particular respect, no refutation has yet been reached. This point must be examined in the conclusion by comparison with its contradictory.

All the following arguments are of this kind: Is it possible for what is-not to be? But surely it is something which is not. Similarly, too, Being will not be; for it will not be any particular thing which is.—Is it possible for the same man at the same time to keep and to break his oath?—Is it possible for the same man at the same time to obey and disobey the same order? Is it not true, in the first place, that being something and Being are not the same thing? On the other hand, Not-being, even if it is something, has not absolute being as well. Secondly, if a man keeps his oath on a particular occasion or in a par­ticular respect, it does not necessarily follow that he is a keeper of oaths, but he who he has sworn that he will break his oath keeps his oath on this particular occasion only by foreswearing himself, but is not a keeper of oaths; nor is he who disobeys obedient, except to a particular order. The argument is similar which deals with the question whether the same man can say what is at the same time both true and false; but it presents apparent difficulties be­cause it is not easy to see whether the qualification 'absolutely' should be applied to 'true' or to 'false.' But there is no reason why the same man should not be absolutely a liar yet tell the truth in some respects, or that some of a man's words should be true but he himself not be truthful. Similarly, too, if there are qualifications of relation or place or time. All the following arguments turn upon a point of this kind. Is health (or wealth) a good thing? But to the fool who misuses it, it is not a good thing; it is, therefore, a good thing and not a good thing.—Is health (or political power) a good thing? But there are times when it is not better than other things; therefore the same thing is both good and not good for the same man. Or is there no reason why a thing should not be absolutely good but not good for a particular person, or good for a particular person, but not good at the present moment or here?—Is that which the wise man would not wish, an evil? But he does not wish for the rejection of the good; therefore, the good is an evil. This is not true; for it is not the same thing to say that the good is an evil and that the rejection of the good is an evil. So likewise with the argument about the thief; it does not follow, if the thief is an evil, that to acquire things is also an evil. The thief, therefore, does not wish for what is evil but for what is good; for to acquire something good is good. Also disease is an evil, but to get rid of disease is not an evil.—Is what is just preferable to what is unjust and are just circumstances prefer­able to unjust? But it is preferable to be put to death unjustly.—Is it just that each man should have his own? But judgements which a man passes in accordance with his personal opinion, even if they are false, are valid in the eyes of the law; the same thing, therefore, is just and not just.—Again, should judgement be given in favour of him who says what is just or of him who says what is unjust? But it is just for the victim of injustice to state in full the things which he has suffered, and these things were unjust. For if to suffer something unjustly is an object of choice, it does not follow that unjust cir­cumstances are preferable to just, but, absolutely, justice is preferable; but this does not prevent unjust circumstances being preferable to just in a particular case. Again, it is just that a man should have his own, and it is not just that he should have what belongs to another; but there is no reason why any judgement which is given in accordance with the judge's opinion should not be just; for, if it is just in a particular case and in particular circumstances, it is not also absolutely just. Similarly, too, there is no reason why, though things are unjust, merely saying them should not be just. For if to say things is just, it does not follow that they are just, any more than, if it is expedient to say things, it follows that those things are expedient. Similarly, too, with things that are just. So that if what is said is unjust, it does not follow that it is a case of the man who uses unjust pleas winning his cause; for he is saying things which it is just for him to say but which are, absolutely, unjust for anyone to suffer.

XXVI. Refutations which are connected with the definition of the refutation must, as suggested above,^a be met by examining the conclusion in the light of its contradictory and seeing how the same term shall be present in the same respect and in the same relation, manner and time. In putting this additional question at the beginning, you must not admit that it is impossible for the same thing to be both double and not double but must admit the possibility but not in the way that was once admitted to fulfil the conditions of a refutation. All the following argu­ments depend upon a point of this kind. 'Does he who knows that A is A, know the thing A?' And, similarly, 'Does he who does not know that A is A, not know the thing A?' But one who knows that Coriscus is Coriscus, might not know that he is musical, so that he both knows and is ignorant of the same thing.—Again, 'Is an object which is four cubits long greater than an object which is three cubits long?' But an object three cubits long might become four cubits long. Now the greater is greater than the less; therefore the object is itself greater and less than itself.

XXVII. In refutations which are connected with the begging and assuming of the original point at issue, it should not be granted to a questioner, if his procedure is obvious, even though his view is gener­ally accepted, but you should state the truth. If, on the other hand, his procedure is not detected, you should, owing to the badness of such arguments, make your ignorance recoil on the head of the ques­tioner, on the ground that he has not argued properly; for refutation must proceed without any assumption of the original point. Next, you must argue that the point was granted with the idea that he was going to use it not as a premiss but in order to argue the opposite view to it or for the purpose of refutations on side issues.

XXVIII. Again, those refutations which draw their conclusions through the consequent must be exposed in the argument itself. There are two ways in which consequences follow: Either as the universal follows from the particular, as 'animal' follows from 'man'; for it is claimed that, if A accompanies B, then B also accompanies A. Or else the process goes by opposites; for if A follows B, A's opposite will follow B's opposite. It is on this, too, that the argu­ment of Melissus depends; for he claims that, if that which has come to be has a beginning, that which has not come to be has no beginning, and so, if the heaven has not come to be, it is also eternal. But this is not true; for the sequence is the reverse.

XXIX. In refutations which are argued by means of some addition, you must examine whether the impossibility occurs none the less when the addition has been withdrawn. If so, then the answerer should make this fact clear and should state that he granted the addition not because he believed in it but for the sake of the argument, but that his opponent has made no use of it at all for his argument.

XXX. In dealing with those who make several questions into one, you should draw a distinction; immediately at the beginning. For a question is single to which there is only one answer, so that one must not affirm or deny several things of one thing nor one thing of several things, but one thing of one thing. But just as in the case of equivocal terms, a predicate is sometimes true of both meanings and sometimes of neither, and so, though the question is not simple, no detriment results if people give a simple answer, so too with these double questions. When, therefore, the several predicates are true of one subject, or one predicate of several subjects, no con­tradiction is involved in giving a simple answer, though he has made this mistake. But when the predicate is true of one subject but not of the other, or several predicates are true of several subjects, then there is a sense in which both are true of both but another sense, on the other hand, in which they are not; so one must be on one's guard against this. The follow­ing arguments illustrate this: (1) Supposing A is good and B evil, it is true to say that they are good and evil and, on the other hand, that they are neither good nor evil (for A is not evil and B is not good), so that the same thing is good and evil and neither good nor evil; (2) If everything is the same as itself and different from anything else, since things are not the same as other things but the same as themselves, and also different from themselves, the same things are both different from themselves and the same as themselves; (3) Moreover, if that which is good becomes evil and that which is evil is good, they would become two; and of two unequal things each is equal to itself, so that they are both equal and unequal to themselves.

These refutations also fall under other solutions; for the terms 'both' and 'all' have several meanings, so that to affirm or deny the same thing is verbal only, and this, as we saw, is not a refutation. But clearly, if one of the several questions is not asked but the answerer affirms or denies a single predicate of a single subject, the reduction to an impossibility will not occur.

XXXI. As regards those who lead one on to repeat the same thing several times over, it is clear that one must not allow that predications of relative terms have any signification in themselves when separated from their correlatives; for example, that 'double' apart from the expression 'double of half' is significant, just because it appears in that expression. For 'ten' appears in the expression 'ten minus one' and 'do' in the expression 'not do,' and affirmations in general in negations; but, all the same, if one were to say 'this is not white,' one is not saying that it is white. 'Double' has possibly no signification at all, just as 'the' in 'the half' too signifies nothing. If it has any signification, it is not the same as in the combined expression. Nor is 'knowledge' of a specific kind, such as 'medi­cal knowledge,' the same as 'knowledge' as a general term; for the latter has always meant 'knowledge of the knowable.' When dealing with terms which are predicated of the terms by means of which they are defined, you must say that the term defined is not the same when taken separately as it is in the combined expression. For 'concave' has the same general meaning when used of the snub-nosed and of the bandy-legged, but when it is combined in the one case with the nose and in the other with the leg, there is no reason why it should not signify different things, for in the first case it signifies 'snub,' in the other 'bandy,' and it makes no difference whether you say 'a snub nose' or 'a concave nose.' Further, the expression must not be allowed to pass without qualification; for it is a falsehood. For snubness is not a concave nose but something, namely a condition, appertaining to a nose; so there is nothing absurd in supposing that a snub nose is one which possesses nasal concavity.

XXXII. As regards solecisms, we have already stated^a the apparent cause of their occurrence; how they should be solved will be clear in the actual arguments. All the following arguments aim at producing this result: 'Is a thing truly that which you truly affirm it to be?' You affirm something to be a stone (accusative masculine)^b; therefore something (nominative neuter) is a stone (accusative masculine). Or does speaking of a stone (a masculine word) involve the use of the relative 'whom' rather rather than 'which' and the pronoun 'him' rather than 'it'? If, then, one were to ask, 'Is a stone him whom you truly state him to be?,' he would not be considered to be talking good Greek any more than if he were to ask, 'Is he whom you state her to be?' But the use of the word 'stick,' or any other neuter word, in this way, involves no difference between the nominative and accusative; therefore no solecism is committed if you say, 'Is this what you affirm it to be?' You affirm it to be a stick; therefore it is a stick. 'Stone,' however, and 'he' have the masculine gender. Now if one were to ask, 'Can "he" be a "she"?,' and then again, 'Why? Is he not Coriscus?' and then were to say, 'Then he is a she,' he has not proved the solecism even if Coriscus signifies a 'she,' though the answerer refuses to concede this; but this must be the subject of a further question. But if neither this is so nor does he concede it, then the solecism has not been proved either in fact or relatively to the person to whom the question was put. Similarly, therefore, in the first example also, 'he' must signify the stone. If, however, this is neither true nor is conceded, the conclusion must not be stated, though it is apparently true, because the case which is used of the word, which is unlike, appears to be like.—'Is it true to say that this object is what you affirm it to be?' You affirm it to be a shield (accusative), therefore it is a shield (accusative). Or is this not necessarily so, if 'this object' (nominative) signifies not shield (accusative) but shield (nominative), while 'this object' (accusative) signifies shield (accusative).—Nor again if he is what you affirm him to be, and you affirm him to be Cleona (accusative of Cleon), is he therefore Cleona? for he is not Cleona^a; for the statement was that he not him is what I affirm him to be. For the question if asked in this form^b would not be Greek either.—'Do you know this?' But this is a stone (nominative); therefore you know a stone (nominative). Has not 'this' a different force in the question 'Do you know this?' and in 'This is a stone,' in the first case standing for an accusative and in the second for a nominative?—When you exercise recognition of an object, do you not recognize it? You exercise recognition of a stone; therefore you recognize 'of a stone.' Do you not in the one case put the object in the genitive and say 'of the stone,' and in the other case in the accusative and say 'a stone'? But it was granted that, when you exercise recognition of a thing, you recognize 'it' not 'of it,' so that you recognize not 'of a stone' but 'a stone.'

That arguments of this kind, then, do not prove solecism but only appear to do so, and why they appear to do so and how you must face them, is clear from what has been said.

XXXIII. It must be noted about arguments in general that in some it is easier and in some more difficult to say why they mislead the listener, though often the latter are identical with the former. For an argument must be called identical when it depends on the same principle, but the same argument might be held by some people to depend on diction, by others on accident and by others on something else, because each, when applied in different contexts, is not equally clear. So, just as fallacies due to equivocation, which are generally regarded as the stupidest form of fallacy, some are obvious even to ordinary minds (for almost all the most laughable remarks depend upon diction). For example, 'A man was carried over the standing board of the framework of the chariot'^a; and 'Whither are you bound?' 'To the yard-arm'^b; 'Which of the two cows will calve in front?' 'Neither, but both behind.'^c 'Is the north wind^d clear?' 'No, certainly not; for he has killed the beggar and the purchaser.'^e 'Is he Evarchus?' 'Certainly not; he is Apollonides.'^f And so on with almost all the rest of the ambiguities, but some even the most expert seem to fail to discern. A proof of this is that people often dispute about the terms used, for example, whether 'Being' and 'Unity' always mean the same thing or some thing different; for some people hold that 'Being' and 'Unity' are identical in meaning, while others solve the argument of Zeno and Parmenides by saying that 'Unity' and 'Being' are used in several senses. Similarly, too, of the arguments which are dependent on accident and each of the other classes, some will be easier to detect and others more difficult, and it is not always equally easy to grasp into which class they fall and whether refutation takes place or not.

A shrewd argument is one which causes most embarrassment; for it bites deepest. Embarrassment is of two kinds. In a reasoned discussion one is in doubt which of the questions one should subvert, whereas in contentious arguments it is about the way in which one is to express the proposition. Hence it is in reasoned discussions that shrewder arguments are more stimulative of inquiry. Now a reasoned argument is shrewdest when from the most generally accepted premisses possible it subverts the most generally accepted thesis possible. For the single argument, if the contradictory is changed about, will result in all the syllogisms being alike; for from generally accepted premisses it will subvert or establish an equally generally accepted conclusion; therefore embarrassment must necessarily arise. Such, then, is the shrewdest argument which puts the conclusion on an equality with the premisses. The next shrewdest is that which argues from premisses which are all on an equality; for this will cause an equal embarrassment as to which question ought to be subverted. The difficulty lies in this, that something must be subverted but it is not clear what. The shrewdest of contentious arguments is that which, in the first place, immediately makes it uncertain whether the reasoning is conclusive or not, and also whether the solution is due to a false premiss or a distinction. Of the rest, that comes next which clearly depends on a distinction or a subversion, but it is not clear which of the premisses it is on the subversion or distinction of which the solution depends, but only whether this process depends upon the conclusion or one of the premisses.

Now sometimes an inadequately reasoned argument is stupid if the premisses assumed are too paradoxical or false; but sometimes it is not deserving of contempt. For when some question is wanting such as concerns the argument or the means of carrying it on, the reasoning which has failed to supply this and is not properly argued is stupid; but when something which is merely extraneous has been omitted, the reasoning is by no means to be lightly condemned but is respectable, though the questioner has not asked his questions well.

As it is possible to address the solution sometimes to the argument, sometimes to the questioner and his mode of questioning and sometimes to neither of these, so likewise also it is possible to address one's questions and reasonings both to the thesis and to the answerer and to the time, when the solution needs more time than the present occasion supplies.

XXXIV. The number, then, and the nature of the sources from which fallacies arise in discussion, and how we are to show up a pretender and make him utter paradoxes, and, further, in what circumstances a solecism occurs, and how to ask question, and what is the right arrangement of questions, and, moreover, what is the use of all such arguments, and also about all answering of questions in general and in particular how to solve arguments and solecisms, on all these subjects let the treatment we have given suffice. There remains to call to mind our original purpose and say a few words about it and then bring our treatise to an end.

Our purpose, then, was to discover a faculty which could reason on the problem set before us from the most generally accepted premisses that exist; for this is the function of dialectic in itself and of the art of examination. But, since there is further added to it, on account of its close affinity with the art of sophistry, that it can undertake an examination not only dialectically but also with a pretence of knowledge, we therefore proposed as the purpose of our treatise not only the above-mentioned task of being able to conduct an argument but also the discovery how, when supporting an argument, we are to defend our thesis by means of the most generally accepted premisses in a consistent manner. Of this we have given the reason; for this was why Socrates used to ask questions but never answered them, because he confessed ignorance. An indication has been given, in what has been said above, of the number of cases in which this will apply and of the various kinds of material which can be used for this and the various sources from which we may obtain an abundance of them; moreover also how questions must be asked and about the arrangement of questions in general, and about answers and solutions applicable to the reasonings employed. All the other points have also been set forth which belong to the same system of argument. In addition to these we have also explained about fallacies, as we have already remarked above. That what we purposed has been satisfactorily carried through to the end is plain; but we must not fail to observe what has happened regarding this inquiry. In all discoveries, either the results of other people's work have been taken over and after having been first elaborated have been subsequently advanced step by step by those who took them over, or else they are original inventions which usually make progress which at first is small but of much greater utility than the later development which results from them. It is perhaps a true proverb which says that the beginning of anything is the most important; hence it is also the most difficult. For, as it is very powerful in its effects, so it is very small in size and therefore very difficult to see. When, however, the first beginning has been discovered, it is easier to add to it and develop the rest. This has happened, too, with rhetorical composition, and also with practically all the other arts. Those who discovered the beginnings of rhetoric carried them forward quite a little way, whereas the famous modern professors of the art, entering into the heritage, so to speak, of a long series of predecessors who had gradually advanced it, have brought it to its present perfection—Tisias following the first inventors, Thrasymachus following Tisias, Theodorus following Thrasymachus, while numerous others have made numerous contributions; hence it is no wonder that the art possesses a certain amplitude. Of our present inquiry, however, it is not true to say that it had already been partly elaborated and partly not; nay, it did not exist at all. For the training given by the paid teachers of contentious argument resembled the system of Gorgias. For some of them gave their pupils to learn by heart speeches which were either rhetorical or consisted of questions and answers, in which both sides thought that the rival arguments were for the most part included. Hence the teaching which they gave to their pupils was rapid but unsystematic; for they conceived that they could train their pupils by imparting to them not an art but the results of an art, just as if one should claim to be about to communicate knowledge for the prevention of pain in the feet and then were not to teach the cobbler's art and the means of providing suitable foot-gear, but were to offer a selection of various kinds of shoes; for he has helped to supply his need but has not imparted an art to him. Also, material enunciated in the past, whereas regarding reasoning we had absolutely no earlier work to quote but were for a long time labouring at tentative researches. If, therefore, on consideration, it appears to you that, in view of such original conditions, our system is adequate when compared with the other methods which have been built up in the course of tradition, then the only thing which would remain for all of you, or those who follow our instruction, is that you should pardon the lack of completeness of our system and be heartily grateful for our discoveries.