Sophistical Refutations

Aristotle · a new plain-English translation from the original language

1| ON SOPHISTICAL REFUTATIONS. Let us now discuss sophistical refutations and refutations that are apparent but are in fact not refutations but fallacies, beginning, as is natural, from the first things. That some deductions are genuine while others seem to be but are not, is clear. For just as this happens elsewhere through some resemblance, so too it holds with arguments. In physical condition, some people really have it well, others merely appear to, by puffing themselves up and dressing themselves up in a showy way; and some are beautiful because of real beauty, others merely appear to be, by adorning themselves. It is the same with lifeless things too: of these, some are really silver and some really gold, while others are not but appear so to perception — for instance things made of litharge and tin appear silver, and things dipped in gall-nut appear gold. In the same way a deduction and a refutation are in one case genuine, in another not genuine but only apparent, on account of inexperience; for the inexperienced are like people viewing things from a distance. For a deduction proceeds from certain things laid down, such that it states something other than what was laid down, of necessity, through what was laid down; a refutation is a deduction accompanied by a contradiction of the conclusion. Some people do not achieve this but seem to, for many reasons, of which the most natural and most widespread source is the one that works through names.

2| Since it is not possible to bring the very things themselves into a discussion, but we use names as symbols in place of things, we suppose that what follows for the names also follows for the things, just as those who calculate suppose the same about their counters. But the case is not similar. For names, and the multitude of arguments, are finite, while things are infinite in number. It is therefore necessary that the same expression, and the single name, signify several things. Just as in the case of counters, then, those who are not skilled at handling them are tricked by the experts, in the same way, in arguments too, those inexperienced in the force of names are led into fallacy, both when arguing themselves and when listening to others. For this reason, then, and for the reasons to be stated, there both is a deduction and a refutation that is apparent but is not real. Since for some people it matters more to seem wise than to be wise without seeming so (for sophistry is apparent wisdom, not real wisdom, and the sophist is one who makes money from apparent, not real, wisdom), it is clear that for such people it is necessary to seem to perform the work of the wise man rather than to perform it and not seem to.

3| To put it, so to speak, as one thing set against one thing: it is the task of the man who knows, with respect to each subject, both not to be in error himself about the things he knows, and to be able to expose the one who is in error. Of these, the one consists in the ability to give an account, the other in the ability to take one. Those, then, who wish to practice sophistry must necessarily seek out this class of arguments just described; for it serves the purpose, since this capacity will produce the appearance of wisdom, which is the very thing such people happen to have chosen as their aim. That there is, then, some such class of arguments, and that those we call sophists aim at this capacity, is clear. Let us now say how many species there are of sophistical arguments, out of how many elements this capacity is composed, and how many parts the subject actually has, and about the other things that contribute to this art. Now of the arguments used in discussion there are four classes: didactic, dialectical, examination-arguments, and contentious. Didactic arguments are those that deduce from the principles proper to each subject and not from the opinions of the answerer (for the learner must trust the teacher); dialectical arguments are those that deduce a contradiction from reputable opinions; examination-arguments are those that deduce from what seems true to the answerer and what is necessarily known to one who claims to possess the knowledge in question (the manner of this has been distinguished elsewhere); contentious arguments are those that deduce, or appear to deduce, from what appear to be, but are not, reputable opinions. Demonstrative arguments have been discussed in the Analytics; dialectical and examination-arguments in other works. Let us now speak of the competitive and contentious kind.

4| First we must grasp how many things those who compete and quarrel in arguments aim at. These are five in number: refutation, falsehood, paradox, solecism, and fifth, making the person one is arguing against talk nonsense — that is, forcing him repeatedly to say the same thing over and over; or rather, making each of these appear to happen rather than actually happen. For their chief preference is to appear to refute; second, to show someone saying something false; third, to lead into paradox; fourth, to make him commit a solecism — that is, to make the answerer, through the argument, speak in a barbarous way in his diction; and last, to make him say the same thing repeatedly.

5| There are two ways of producing refutation: some depend on the diction, others lie outside the diction. Those depending on diction that produce the illusion are six in number: homonymy, amphiboly, combination, division, accent, and form of expression. Proof of this comes both through induction and through deduction, if one takes some other premise, and from the fact that the same names and expressions and from the fact that in just so many ways we could fail to signify the same thing by the same names and expressions. The following are examples of arguments that depend on homonymy: that those who know, learn — for grammarians learn the things that are dictated to them. For "to learn" is homonymous, meaning both "to come to understand by using knowledge" and "to acquire knowledge." And again, that bad things are good: for the things that are needful are good, and bad things are needful. For "needful" is twofold: what is necessary, which often applies even to bad things (for some evil is necessary), and we also say that good things are needful. Again, that the same person is both sitting and standing, and both sick and healthy. For whoever got up, stands; and whoever was getting well, is well; but it was the one sitting who got up, and the one sick who was getting well. For to say that the sick person does or undergoes anything whatever does not signify one thing, but sometimes means the one who is now sick or sitting, sometimes the one who was sick before.

6| Except that it was the one who was sick, even while still sick, who was getting well; but it is not the sick person who is now well, but the one who was sick before. Examples of arguments that depend on amphiboly are: "my wish is that I capture the enemy"; and "is what someone knows, known by him?" for by this expression it is possible to signify both the knower and the thing known, as being known. And, "does one see what one sees?" one sees the pillar, so the pillar sees. And, "is what you say to be, what you say?" you say a stone to be; therefore you say you are a stone. And, "is it possible to speak while silent?" for "to speak while silent" is twofold: it can mean that the speaker is silent, or that the things spoken of are silent. There are three ways in which arguments depend on homonymy and amphiboly: one, when either the expression or the name properly signifies several things, for instance "aetos" (eagle, or a kind of pediment) and "kyon" (dog, or the dog-star); another, when we are accustomed to speak that way; and a third, when the combined expression signifies several things, though each part taken separately signifies only one thing simply, for instance "he knows letters." For each of the two words, if it so happens, signifies some one thing, "knows" and "letters"; but the two together signify several things, either that the letters themselves have knowledge, or that someone else has knowledge of the letters.

7| Ambiguity and equivocation, then, occur in these ways. Fallacies from composition are of the following kind: for example, that a sitting person can walk, and that a person not writing can write. For these do not mean the same thing depending on whether one states them dividing the words or combining them: taken divided, it is possible for the sitting person to walk and for the person not writing to write; and if one likewise combines this, that the person not writing writes, it signifies that he has the capacity, while not writing, to write. But if one does not combine it, it signifies that he has, at a time when he is not writing, the capacity to write. And a person is now learning letters, if indeed he was learning what he now knows. Again: the person who has the capacity to carry only one thing at a time has the capacity to carry many things. Fallacies from division are of this kind: that five is two and three, and both odd and even, and that the greater is equal; for it is that much and something more besides. For the same expression, divided and combined, does not always seem to signify the same thing, as for example 'I made you, being a slave, free' and 'Fifty men, a hundred, noble Achilles left behind.' As for fallacies from accent, in spoken arguments without writing it is not easy to construct one, but it is easier in written texts and poems, as for instance some people emend the text of Homer against those who criticize him for having spoken absurdly

8| in the line 'Its part rots in the rain.' They resolve this by the accent, pronouncing the word for 'not' with an acute accent. And likewise, in the matter of Agamemnon's dream, they argue that it was not Zeus himself who said 'We grant him to win the glory he prays for,' but that Zeus was instructing the dream to grant it. Such cases, then, are fallacies from accent. Those from the form of the expression occur whenever what is not the same is expressed in the same way — for example, when the masculine is expressed as feminine, or the feminine as masculine, or something intermediate as one of the two; or again, when a quality is expressed as a quantity or a quantity as a quality, or what acts as what is acted on, or a state as an action, and the rest as distinguished earlier. For it is possible, by the form of expression, to signify as belonging to the class of actions what does not belong to that class. For example, 'to be healthy' is expressed, in the form of the expression, in a way similar to 'to cut' or 'to build'; and yet the one signifies a quality and a certain state, while the other signifies doing something. And the same way with the rest. The refutations that depend on language, then, arise from these topics. Of the fallacies outside language there are seven kinds: one from accident; a second from something being said without qualification, or not without qualification but in a certain respect, or place, or time, or in relation to something; a third from ignorance of what a refutation is; a fourth from the consequent; a fifth from assuming the original point; a sixth from setting down what is not the cause as the cause; and a seventh from making several questions into one.

9| Fallacies from accident, then, occur whenever anything whatever is claimed to belong alike to the thing itself and to what is accidental to it. For since many things are accidental to one and the same thing, it is not necessary that all these belong to everything that is predicated of it, and to that of which it is predicated. For example: if Coriscus is other than man, he is other than himself, since he is a man. Or: if he is other than Socrates, and Socrates is a man, they claim he has admitted that he is other than man — because it is accidental to the one he said Coriscus was other than, that this one happens to be a man.

10| Fallacies from something being said without qualification, or in a certain respect and not in the strict sense, occur when what is said in a particular respect is taken as said without qualification. For example: if what is not is an object of opinion, they conclude that what is not, is; yet being something and being unqualifiedly are two different things. Or again: that what is, is not, if some one of the things that are, is not — for example, if it is not a man. Not being something and not being at all are likewise distinct, though the wording makes them seem alike, since being something is barely different from being, and not being something barely different from not being. And likewise for the fallacy from something being said in a certain respect and without qualification. For example: if the Indian, being wholly black, is white as to his teeth, he is therefore both white and not white. Or: if both belong in a certain respect, they conclude that the opposites belong at the same time without qualification. In some cases this is easy for anyone to see through — for example, if someone, having granted that the Ethiopian is black, should then ask whether he is white; and then, if he is white in this respect, concludes that he is black and not black, and thinks he has argued syllogistically by completing the question. But in other cases it often goes unnoticed — in all the cases where, when something is said in a certain respect, the unqualified statement also seems to follow, and where it is not easy to see which of the two ought properly to be assigned. This happens in cases where the opposites belong alike; for it seems that one must grant that either both, or neither, are to be predicated without qualification — for example, if half is white and half is black, is it white or black?

11| Fallacies from not having defined what a syllogism is or what a refutation is arise from a deficiency in the account. For a refutation is a contradiction of one and the same thing — not of a name but of a fact, and of a name that is not merely synonymous but the very same name — following of necessity from what has been granted, without the original point being counted among the premises, in the same respect, in relation to the same thing, in the same way, and at the same time. And falsifying an opinion about something works in the same manner. Some people, then, by leaving out one of these conditions, appear to refute — for example, arguing that the same thing is double and not double: for two is double of one, but not double of three. Or, that the same thing is double and not double of the same thing, but not in the same respect: for in length it is double, but in breadth it is not double. Or, that it is double of the same thing, in the same respect, and in the same way, but not at the same time; and this is precisely why it is only an apparent refutation. One might, however, drag this case in among the fallacies dependent on language as well. Fallacies from assuming the original point arise in as many ways as it is possible to beg the question, and they appear to refute because one is unable to see clearly what is the same and what is different.

12| The refutation from the consequent occurs because of supposing that a relation of following can be converted. For whenever, given this, that necessarily is, people also suppose that, given that, this necessarily is. This is also the source of the deceptions about opinion that arise from perception. For people have often taken bile to be honey, because a yellow color follows upon honey; and since it happens that when it has rained the earth becomes soaked, if it is soaked we suppose that it has rained. But this is not necessary. And in rhetoric, demonstrations from a sign are drawn from the consequent. For wanting to show that someone is an adulterer, they seize on the consequent — that he grooms himself elaborately, or that he is seen wandering about at night. But many people to whom these things belong do not have the thing predicated of them actually belong to them. And likewise in syllogistic arguments — for example, the argument of Melissus that the universe is infinite, which takes it that the universe is ungenerated (for nothing could come to be out of what is not), and that whatever has come to be has come to be from a starting point. If, then, it has not come to be, the universe has no starting point, and so it is infinite. But this does not necessarily follow; for it is not the case that, if everything that has come to be has a starting point, then also whatever has a starting point has come to be — just as it is not the case that, if the feverish person is hot, the hot person must be feverish.

13| The fallacy of treating what is not the cause as the cause occurs when the non-cause is added in as though the refutation came about because of it. This happens in arguments that reduce to an impossibility, since in these it is necessary to do away with one of the things laid down. If, then, the non-cause is counted among the questions necessary for producing the impossible conclusion, it will often seem that the refutation comes about because of it — for example, the argument that soul and life are not the same thing: for if coming-to-be is contrary to perishing, then some particular coming-to-be will be contrary to some particular perishing; but death is a kind of perishing and is contrary to life; therefore life is a coming-to-be, and living is coming-to-be; but this is impossible; therefore soul and life are not the same. Now this has not in fact been proved by syllogism: for the impossibility follows even if one does not say life is the same as soul, but only that life, being contrary to death, which is a perishing, is contrary, and perishing is contrary to coming-to-be. Such arguments, then, are not without a syllogism absolutely, but they are without a syllogism relative to the point at issue. And this often escapes the notice of the questioners themselves no less than of others.

14| Such, then, are the arguments that depend on the consequent and on treating a non-cause as a cause. Those that depend on making two questions into one occur when it goes unnoticed that there are several questions and a single answer is given as though there were only one. In some cases it is easy to see that there are several and that an answer ought not to be given, for instance: is the earth sea, or is the sky? But in others it is less easy, and people either concede as though there were one question, by not answering what is asked, or appear to be refuted — for example: is this one and this one a man? So that if someone strikes this one and this one, he will have struck a man, not men. Or again: of things some of which are good and some not good, are all of them good, or not good? For whichever he says, he might seem to produce a refutation, or an apparent falsehood: for to say that something not good is good, or that something good is not good, is false. But sometimes, if certain further things are granted, even a genuine refutation would come about — for instance, if someone were to grant that 'white' and 'naked' and 'blind' are said in the same way of one thing and of many things. For if what is blind is what naturally has sight but does not have it, then things will be blind that do not have sight though naturally fitted to have it. So whenever one has it and another does not, both will be either seeing or blind — which is impossible.

15| Either, then, apparent syllogisms and refutations must be divided in this way, or else all of them must be traced back to ignorance of what refutation is, taking this as the starting point: for it is possible to resolve all the ways described into the definition of refutation. First, one should ask whether they lack a syllogism: for the conclusion must follow from the things laid down, so that it is stated of necessity and not merely appears to be. Next, one should also go by the parts of the definition. Of the fallacies that depend on language, some depend on the double meaning — such as ambiguity of a word, ambiguity of a phrase, and likeness of form (for it is customary for all things to signify a 'this something'), while composition, division, and accent differ in that the phrase or the name is not the same. But it was also necessary — just as the thing itself must be the same — that this too be the same, if there is to be a refutation or a syllogism; for example, if it is a cloak, one should conclude 'cloak,' not 'garment.' For that too is true, but it has not been proved by syllogism; a further question is still needed — namely, that it signifies the same thing — addressed to someone asking why.

16| Fallacies that depend on accident become clear once the syllogism has been defined. For the same definition must hold of refutation too, except that the contradiction must be added; for a refutation is a syllogism of a contradiction. If, then, there is no syllogism of the accidental attribute, no refutation occurs. For it is not the case that, if these things being so, this must be so, and this is white, it is therefore necessary that it be white because of the syllogism. Nor, if the triangle has angles equal to two right angles, and it is accidental to it to be a figure, or a first thing, or a principle, does it follow that it is because it is a figure or a principle or a first thing — for the demonstration holds not insofar as it is a figure nor insofar as it is a first thing, but insofar as it is a triangle. Similarly in the other cases too. So that if refutation is a kind of syllogism, refutation by accident could not be one. But it is because of this that even craftsmen, and in general those who possess knowledge, are refuted by those without knowledge: for the latter construct their syllogisms against the experts on the basis of accident, and the experts, being unable to make the distinction, either grant what is asked, or, without having granted it, suppose they have granted it. Fallacies that depend on the qualified and the unqualified occur because the affirmation and the negation are not of the same thing. For the denial of what is white in a certain respect is what is not white in that respect, whereas the denial of what is white without qualification is what is not white without qualification. So if, when someone has granted that a thing is white in a certain respect, one takes this as though it had been said without qualification, one does not produce a refutation, but it appears to, on account of ignorance of what refutation is.

17| Most evident of all are those previously described as depending on the definition of refutation, which is why they have been given that name: for the appearance of them arises from the deficiency of the argument, and, dividing them this way, one should set down the deficiency of the argument as common to all of them. And those that depend on begging the original point, and on positing the non-cause as cause, are also clear through the definition. For the conclusion must be true because these things are the causes of its coming about, which was not the case in the arguments involving a non-cause; and again, the point at issue must not be counted among the premises, which is not so with those that beg the original point.

18| Fallacies that depend on the consequent are a part of those that depend on accident: for what follows is an accident, but it differs from the accident in that the accident can be taken with respect to one thing only — for instance, that the tawny and honey are the same, and the white and a swan — whereas what follows is always taken with respect to several things; for we hold that things that are the same as one and the same thing are also the same as one another. This is why a refutation depending on the consequent comes about. But it is not always true — for instance, if a thing is white by accident: for snow and a swan are both the same as white. Or again, as in the argument of Melissus, it is taken that having come to be and having a beginning are the same, or that becoming equal and taking on the same magnitude are the same. For because what has come to be has a beginning, one also claims that what has a beginning has come to be, on the assumption that both — what has come to be and what is bounded — are the same in virtue of having a beginning. Likewise also in the case of things becoming equal: if things that take on the same and a single magnitude become equal, then things that become equal also take on a single magnitude. Thus the consequent is taken as though it were reversible. Since, then, refutation by accident lies in ignorance of what refutation is, it is clear that refutation by the consequent does too. This point should also be examined in another way.

19| Fallacies that depend on making several questions into one occur because we fail to articulate, or fail to divide, the account given in the premise. For a premise is one thing said of one thing. For the same term applies to one thing only and to that thing taken simply — for instance, to man, and to one man only; and likewise in the other cases. If, then, a premise is one thing that claims one thing of one thing, then a question of this kind will also be a premise without qualification. And since a syllogism is composed of premises, and a refutation is a syllogism, a refutation too will be composed of premises. If, then, a premise is one thing said of one thing, it is clear that this fallacy too lies in ignorance of what a refutation is: for what is not really a premise appears to be a premise. If, then, an answer has been given as though to a single question, there will be a refutation; but if it has not really been given, though it appears to have been, there will be an apparent refutation. So all the topics fall under ignorance of what refutation is — those depending on language, because there is an apparent contradiction, which was the distinctive mark of refutation, and the others, because they depart from the definition of syllogism.

20| The deception in the fallacies arising from ambiguity of name and of phrase comes from being unable to distinguish what is said in several ways (some things are not easy to divide, for instance the one, being, and the same); in those arising from composition and division, it comes from thinking that the phrase, whether composed or divided, makes no difference — as happens in most cases. Similarly with those arising from accent: for the sentence, whether relaxed or intensified, does not seem to signify anything different — in no case, or in not many. Those arising from figure of speech depend on the similarity of the expression. For it is hard to distinguish which things are said in the same way and which in a different way; indeed, whoever is able to do this is close to contemplating the truth, and above all knows how to give assent correctly, because we suppose that everything predicated of something is a "this" and understand it as one; for being a "this" and being seem most of all to accompany unity and substance. Hence this mode too must be classed among those arising from language — first, because deception occurs more when we examine together with others than by ourselves (for examination with another proceeds through words, but examination by oneself proceeds no less through the thing itself), and then because deception happens even when alone, whenever one conducts the examination upon the phrase;

21| further, the deception comes from the resemblance, and the resemblance comes from the language. Those arising from accident occur through inability to distinguish the same and the different, and one and many, and to distinguish which of the predicates all belong to equally with the thing itself. Similarly too with those arising from the consequent; for the consequent is a part of the accident. Further, in many cases it also appears, and is claimed, to hold that if this is not separated from that, then that is not separated from this either. Those arising from deficiency of statement, and those arising from the qualified taken as unqualified — the deception in general lies in the small omission; for we concede the universal as though the "something," or "in a certain respect," or "in a certain way," or "now," added nothing extra to the meaning. Similarly too with those who assume the point at issue, and those who give a non-cause as cause, and all who make several questions into one; for in all of these the deception is due to the small omission; for we do not scrutinize precisely the term of either the premise or the syllogism, for the reason stated.

22| Since we have the sources from which apparent syllogisms arise, we also have the sources from which sophistical syllogisms and refutations would arise. And I call a refutation and syllogism sophistical not only the one that appears to be a syllogism or refutation but is not, but also the one that is a genuine syllogism but only appears to be proper to the matter in question. These are the ones that do not refute in accordance with the matter and merely show people to be ignorant — which was the business of the art of testing. Now the art of testing is a part of dialectic; and this art can reason falsely because of the ignorance of the one who grants the argument. Sophistical refutations, even when they do reason to the contradiction, do not make it clear whether the person is ignorant; for they trip up even the one who knows, by means of these very arguments.

23| That we grasp them by the same method is clear: for the very things that make it appear to the hearers that a conclusion has been reasoned from what was asked would also seem so to the answerer, so that there will be false syllogisms through these — either all of them or some; for what a person thinks he has granted without being asked, he would also grant if asked. Except that in some cases it happens that, in asking the further question, one at the same time reveals both the deficiency and the falsity — as in the cases arising from language and from solecism. If, then, the paralogisms of the contradiction arise from what makes for an apparent refutation, it is clear that false syllogisms too would arise from just as many sources as the apparent refutation does. And the apparent refutation arises from the parts of the genuine one; for when each of these fails, a merely apparent refutation would appear — for instance, the one arising from what does not follow because of the argument, the one leading to the impossible, and the one that makes two questions into one, arising from the premise; and instead of what holds in a thing's own right, the one arising from accident, and a part of this, the one arising from the consequent; further, the fallacy of its happening not with respect to the thing but with respect to the argument; then, instead of the universal, taking the contradiction according to the same thing and in relation to the same thing and in the same way, arising from the qualified taken as unqualified, or from each of these individually; further, arising from assuming the point at issue when it is not being counted as part of the premises. So we would have the sources from which the paralogisms arise; for they could not arise from more sources than these, but all of them will arise from the ones stated.

24| The sophistical refutation is not a refutation without qualification, but a refutation relative to someone; and likewise the syllogism. For if the one arising from the ambiguous term does not take it as signifying one thing, and the one arising from similarity of figure does not take "only this," and the others likewise, they will be neither refutations nor syllogisms, either without qualification or relative to the person questioned; but if they do take it so, they will be refutations relative to the person questioned, but not without qualification; for they have not taken something that signifies one thing, but only something apparent, and that from this particular person.

25| One should not try to grasp the number of ways in which those who are refuted are refuted without having knowledge of all things that are. But this belongs to no single art; for the sciences are perhaps unlimited in number, so that clearly the demonstrations are too. But there are also true refutations; for whatever can be demonstrated can also be used to refute the person who posits the contradiction of the truth — for instance, if someone posited that the diagonal is commensurable, one could refute him by the demonstration that it is incommensurable. So on this reasoning one would need to have knowledge of everything; for some refutations will arise from the principles of geometry and their conclusions, others from those of medicine, others from those of the other sciences. But indeed the false refutations too are likewise unlimited in number; for in accordance with each art there is a false syllogism — for instance, in accordance with geometry the geometrical one, and in accordance with medicine the medical one. And by "in accordance with the art" I mean in accordance with that art's own principles. It is clear, then, that one need not grasp the topics of all refutations, but only those arising in connection with dialectic; for these are common to every art and capacity. And it belongs to the expert in each science to examine the refutation that arises in accordance with that science — whether it merely appears not to be a genuine refutation or actually is one, and if it is, why it is;

26| but the refutation that arises from what is common and falls under no art belongs to the dialectician. For if we have the sources of the reputable syllogisms about anything whatsoever, we have the sources of refutations too; for a refutation is a syllogism of the contradiction, so that either one or two syllogisms of the contradiction constitute a refutation. We have, then, the sources from which all such refutations come. And if we have this, we also have the solutions; for the objections to these are the solutions. And we have, from the sources from which they arise, the apparent ones too — apparent, however, not to just anyone, but to people of a certain sort; for it is indefinite if one examines the sources from which they appear convincing to random people. So it is clear that it belongs to the dialectician to be able to grasp the sources from which, by means of what is common, there arises either a genuine refutation or an apparent refutation, and either a dialectical or an apparent-dialectical or a testing argument.

27| There is no such distinction among arguments as some people state, namely that some arguments are directed against the name while others are directed against the thought. For it is absurd to suppose that there are arguments against the name and different ones, other than these same arguments, against the thought. For what is it not to be directed against the thought, except when someone does not use the name in the sense in which he thinks he is asking about what the person questioned granted? And this very thing is also being directed against the name. Being directed against the thought is when one uses it in the sense in which the person granted it, having thought it through. If, then, when a name signifies several things, both the questioner and the person questioned suppose it signifies one thing — as perhaps "being" or "the one" signifies many things, but both the answerer and the questioner, Zeno, supposing it to be one, asked his question, and the argument is that all things are one — this argument will have been conducted either against the name or against the thought of the person questioned, depending on how it has proceeded. But if someone supposes it signifies many things, clearly he is not arguing against the thought. For in the first place, "against the name" and "against the thought" hold only of arguments concerning things that signify several things; but then again "against the name" holds concerning anything whatsoever;

28| for being directed against the thought does not lie in the argument itself, but in how the person answering stands in relation to what has been granted. Next, it is possible for all arguments to be against the name. For here being against the name is the same as not being against the thought. For if not all are, there will be some arguments that are neither against the name nor against the thought; but they say all are, and they divide arguments into being either against the name or against the thought, all of them, no others. But in fact, among the syllogisms that turn on multiple meaning, some are the ones that turn on the name. For it has been stated absurdly to call all arguments that turn on expression "against the name"; still, there are some false reasonings that arise not because the answerer stands in some particular relation to them, but because the question itself, put in a certain way, contains a term that signifies several things. And in general it is absurd to discuss refutation without first discussing syllogism; for refutation is a syllogism, so one must discuss syllogism before discussing false refutation, since this kind of refutation is an apparent syllogism of a contradiction. Hence the cause will lie either in the syllogism or in the contradiction (for the contradiction must be added), and sometimes in both, if it is to be an apparent refutation. Now the fallacy of the silent one speaking lies in the contradiction, not in the syllogism; the fallacy of giving what one does not have lies in both; and the fallacy that Homer's poem is a figure by means of the circle lies in the syllogism. But a syllogism that is true in neither respect is a genuine syllogism.

29| But indeed, to return to the point from which the argument came: are the arguments used in the mathematical sciences directed against the thought or not? And if someone thinks "triangle" signifies many things, and granted it not as meaning this figure regarding which it was concluded that it has two right angles, has this person conducted the discussion against that person's thought or not? Further, if the name signifies many things, but the answerer does not have this in mind and does not suppose it, how has this person not conducted the discussion against the thought? Or how else must one ask a question, except by giving the distinction, if someone should ask whether it is possible for the silent one to speak or not, or whether it is possible in one sense and not in another? If, then, someone grants it in no sense at all, and the other then argues, has he not conducted the discussion against the thought? And yet the argument seems to belong among those that turn on the name. Therefore there is no genus of arguments that is "against the thought." Rather, some arguments are against the name; and not even all of these are such — not only are the refutations not all such, but not even all the apparent refutations are. For there are also apparent refutations that do not turn on expression, such as those that turn on accident, and others.

30| But if someone claims the right to make the distinction — saying, "by 'the silent one speaking' I mean it in this sense in one case, in that sense in another" — in the first place, this very claim is absurd (for sometimes the question does not seem to be ambiguous, and it is impossible to distinguish what one does not think is ambiguous); and next, what else will this be but teaching? For it will make plain how the matter stands to someone who has neither examined it, nor knows it, nor supposes that it is said in another sense. Since indeed, what prevents this same thing from happening even in cases that are not double in meaning? Are the units in four equal to the twos? And there are twos present in one way and in another way. And is there one science of contraries or not? Some contraries are knowable and others unknowable. So the person who makes this claim seems not to know that teaching is different from engaging in discussion, and that the teacher must not ask questions but must himself make things clear, while the discussant must ask questions.

31| Further, to claim the right to assert or deny is not the mark of one who is demonstrating, but of one who is making trial. For the art of testing is a kind of dialectic, and it examines not the person who knows but the person who is ignorant and pretends to know. Now the one who examines the common principles with a view to the subject matter is a dialectician, while the one who does this only in appearance is a sophist. And there is one kind of contentious and sophistical syllogism, which is an apparent syllogism concerning those matters about which dialectic is a means of testing, even if the conclusion is true; for it is deceptive as regards the reason why. And there are also false reasonings which, though not falling under the method proper to each subject, seem to be so according to the art. For false diagrams are not contentious (since the false reasonings occur within what falls under the art), nor is it contentious if there is some false diagram concerning something true, such as that of Hippocrates, or the squaring by means of the lunes. But as Bryson squared the circle — even if the circle is squared — still, because it is not done in accordance with the subject matter, it is for this reason sophistical. So both the apparent syllogism about these matters is a contentious argument, and the apparent syllogism concerning the subject matter, even if it is a genuine syllogism, is a contentious argument; for it is apparent with respect to the subject matter, so that it is deceptive and unjust. For just as wrongdoing in a contest has a certain form and is a kind of unjust fighting, so too in argumentative dispute the contentious is a kind of unjust fighting;

32| for there too, those who choose to win at all costs resort to everything, and here too do the contentious. Those, then, who are such for the sake of victory itself seem to be contentious people and lovers of strife, while those who are such for the sake of reputation leading to money-making are sophistical; for the sophistical art is, as we said, a kind of money-making from apparent wisdom, and this is why they aim at apparent demonstration. And the same arguments belong to both the lovers of strife and the sophists, but not for the sake of the same things. And the same argument will be both sophistical and contentious, but not in the same respect: insofar as it aims at apparent victory, it is contentious; insofar as it aims at apparent wisdom, it is sophistical. For the sophistical art is indeed a kind of apparent wisdom, but is not wisdom. The contentious person stands in somewhat the same relation to the dialectician as the false-diagram-maker stands to the geometer; for the false-diagram-maker deceives the geometer by means of the same principles as dialectic uses, and out of those same principles the contentious person misleads. But the false-diagram-maker is not contentious, because he draws his false diagram from the principles and conclusions that fall under the art; whereas the one who works under dialectic will clearly be contentious with regard to other matters. For instance, the squaring by means of the lunes is not contentious, but that of Bryson is contentious;

33| One sort cannot be transferred except to geometry alone, because it proceeds from principles peculiar to that science; the other can be transferred to many people—all those who do not know what is possible and what is impossible in each case, for it will fit them too. Or take the way Antiphon squared the circle. Or if someone refuses to admit that it is better to walk after dinner, on the strength of Zeno's argument—that is not a point of medicine, since it is common property. Now if the disputant stood toward the dialectician in exactly the way the man who draws false figures stands toward the geometer, he would not be disputatious about those other matters. But as things are, the dialectician is not concerned with some determinate genus, nor is he a demonstrator of anything, nor is he like the universal science. For not all things fall under one single genus, nor, even if they did, could the things that are fall under the same principles. So no craft that demonstrates some determinate nature proceeds by questioning: for it is not possible to grant whichever of the two parts one likes, since a syllogism does not come from both together. But dialectic does proceed by questioning. And if it demonstrated anything, even if not everything, then at least it would not put the first things and its own proper principles in question. For if people did not grant these, it would no longer have anything left from which to go on arguing against the objection.

34| The same holds for the art of testing. For the art of testing is not like geometry either, but is something one could possess even without knowing the subject. For it is possible even for someone who does not know the matter to take a test of someone else who does not know it, provided the person being tested grants his answers not from what he knows, nor from what is peculiar to the subject, but from what follows—such things as, if he knows them, nothing prevents him from not knowing the craft, while if he does not know them, he is bound to be ignorant of it. So it is clear that the art of testing is knowledge of nothing determinate. That is why it concerns everything: for all the crafts make use of certain common things as well. That is why everyone, even laymen, in some fashion make use of dialectic and the art of testing: for everyone, up to a point, tries to cross-examine those who profess a subject. Now these are the common things: for people know them no less well than the experts, even if they seem to be speaking quite outside the subject. So everyone practices refutation: for they share unartfully in that in which dialectic shares artfully, and the man who tests by the craft of the syllogism is a dialectician. Since there are many such common things, and they hold of everything, yet are not such as to constitute some single nature and genus, but are rather like negations, while other things are not of this sort but peculiar to a subject, it is possible, on the strength of these, to take a test about everything, and for there to be some craft of doing so—yet not a craft of the sort that demonstrates.

35| That is why the eristic disputant is not in every respect in the same position as the man who draws false figures: for he will not be a false reasoner working from some determinate genus of principles, but the eristic disputant will operate about every genus. These, then, are the ways of sophistical refutations. That it belongs to the dialectician to study these matters and to be able to produce them is not hard to see, for the method concerning premises covers this whole subject of study. We have now spoken about apparent refutations. As for showing something false and leading the argument into a paradoxical conclusion—this was the second aim of the sophistical purpose—this comes about above all through a kind of questioning, that is, through interrogation. For to question someone whose position is fixed with respect to nothing definite is a way of hunting for these results: for people who speak at random are more likely to go wrong, and they speak at random when they have nothing set before them. Asking many questions—even when what one is arguing about is determinate—and requiring the other person to say what seems to him to be so, provides a certain resourcefulness for leading him into a paradox or a falsehood: for whether the person questioned affirms or denies any of these points, one can lead him toward whatever gives one resources for an attack. It is possible now to do less mischief by these means than formerly, for people demand to know what this has to do with the original point. An element in hitting on some falsehood or paradox is not to ask any thesis straight out, but to claim that one is asking merely because one wants to learn: for inquiry of this kind creates room for an attack.

36| For showing something false, the special sophistical topic is leading the other person into just those things for which one has arguments ready to hand; and this can be done well or badly, as was said before. Again, for getting someone to say paradoxical things, one should consider from what genus the person one is arguing with comes, and then ask him the sort of question that people of that kind regard as paradoxical: for each group has something of this sort. An element of this technique is to have grasped, in the premises, the positions characteristic of each group. The proper solution for these cases too is to point out that the paradoxical result does not come about because of the argument—and that is always what the contestant is aiming to show. Further, one can work from people's wishes and from their openly held opinions. For people do not wish for the same things they say: they state the most respectable-sounding accounts, but they wish for what appears to be advantageous—for instance, they say one ought to die nobly rather than live pleasantly, and be poor justly rather than rich shamefully, but they wish for the opposite. So the man who speaks according to his wishes should be led toward his openly held opinions, and the man who speaks according to those should be led toward his hidden ones: for either way he is bound to say paradoxical things, since he will be saying things contrary either to his openly held opinions or to his hidden ones.

37| The greatest topic for producing paradoxical statements—as Callicles too is portrayed as saying in the Gorgias, and as all the ancients supposed to hold—turns on nature and law: for they held that nature and law are opposed, and that justice is fine according to law but not fine according to nature. So one ought, against the man who speaks in terms of nature, to meet him in terms of law, and against the man who speaks in terms of law, to lead him toward nature: for either way one gets him to say paradoxical things. For them, what accorded with nature was the truth, while what accorded with law was what seems so to the many. So it is clear that those thinkers too, just like people today, tried to make the answerer either be refuted or say paradoxical things. Some questions have the property that either answer is bound to seem unacceptable, for instance whether one ought to obey the wise or one's father, and whether to do what is advantageous or what is just, and whether it is more choiceworthy to suffer wrong or to do harm. One ought to lead a person toward whatever is contrary both to the many and to the wise, according as he speaks like the experts in argument or like the many: if he speaks like the experts, lead him toward the views of the many; if like the many, toward the views of the experts. For some say that the happy man is necessarily just, while to the many it seems unacceptable that a king should not be happy. And bringing someone to unacceptable conclusions of this sort is the same thing as leading him into the opposition between nature and law: for law is the opinion of the many, while the wise speak according to nature and according to truth.

38| Paradoxical statements, then, must be sought out from these topics. As for producing tedious verbosity, we have already said what we mean by "being tediously verbose." All arguments of the following kind aim to produce this effect: if it makes no difference whether one states the name or the account that defines it, and double and double-of-half are the same thing, then, if double is of half, it will be double of half-of-half. And again, if "double" is replaced by "double of half," the phrase will be said three times over: "half of half of half, double." Or again: is appetite a striving for what is pleasant? That is exactly what striving for what is pleasant is; appetite, then, is a striving for what is pleasant, pleasant. Now all arguments of this sort occur among relative terms—namely, all those cases where not only the genera but the terms themselves are said relative to something, and are rendered relative to one and the same thing. For example, both striving is a striving for something and appetite is an appetite for something, and double is a double of something, and double is double of half. And in the case of things whose substance is not relative at all, but of which there are states or affections or something of the kind, this is made additionally clear in their account when they are predicated of these subjects. For example: the odd is a number having a middle term; and an odd number is a number; therefore an odd number is a number having a middle term, a number. And if snub-nosed is a concavity of the nose, and a nose is snub, then a nose is a concave nose, nose.

39| They appear to be doing something, sometimes, when they are not doing it, because they fail to go on and ask whether 'the double,' said just by itself, signifies something or nothing, and if it signifies something, whether the same thing or a different thing, but state the conclusion straightaway. But it appears (to signify the same), because the name is the same, so it appears also to signify the same thing. What sort of thing solecism is has been stated earlier. It is possible both to commit it and appear not to, and to commit it and not be thought to — as Protagoras used to argue, in the case of whether 'wrath' and 'helmet' are masculine. The one who says 'oulomenēn' (feminine) commits a solecism by his account, though it does not appear so to others; the one who says 'oulomenon' (the other form) appears to, but does not commit one. It is clear, then, that this could also be produced by skill. That is why many arguments, without actually reasoning to a solecism, appear to reason to one, as happens in refutations.

40| Nearly all apparent solecisms arise around the word 'this,' and whenever the case-form shows neither the masculine nor the feminine but the form in between. 'Houtos' ('this,' masculine) signifies the masculine, and 'hautē' ('this,' feminine) the feminine; 'touto' ('this,' neuter) is meant to signify the one in between, but often signifies each of the other two as well — for instance, 'What is this (touto)?' 'Calliope,' 'a piece of wood,' 'Koriskos.' Now for the masculine and the feminine all the case-forms differ, but for the one in between some do and some do not. So when 'touto' has repeatedly been granted, people reason as though 'touton' (the masculine accusative) had been stated; and likewise with one case-form substituted for another. The fallacy arises because 'touto' is common to several case-forms; for 'touto' signifies at one time 'houtos' and at another 'touton.' But it ought to signify each in its proper turn — with 'esti' (is) the form 'houtos,' and with 'einai' (to be) the form 'touton' — for instance, 'Koriskos is' and 'that it is Koriskos.' And the same holds for feminine names, and for things called 'implements' that nonetheless carry a feminine or a masculine designation. For only those nouns ending in -on carry the designation proper to an implement, such as 'wood' and 'rope'; those not so ending are masculine or feminine, and some of these we apply to implements — for instance 'wineskin' is masculine in name, while 'couch' is feminine. That is why in such cases too 'esti' and 'einai' will differ in the same way. And in a way solecism is similar to those refutations that arise from calling things alike that are not alike. For just as those fallacies occur with respect to things, these occur with respect to names; for 'man' is both a thing and a name, and so is 'white.'

41| It is clear, then, that one must try to reason to a solecism out of the case-forms just described. These, then, are the species of contentious arguments, and the parts of the species, and the methods that have been stated. And it makes no small difference if the questions are arranged in a certain order with a view to going unnoticed, as happens in dialectical arguments as well. So, following what has been said, these points must be discussed first. Now for refuting there is, first, length; for it is hard to take in many things at once. For length one must use the elements stated earlier. Second, speed; for those who lag behind see less far ahead. Further, anger and contentiousness; for when people are disturbed they are all less able to guard themselves. The elements that produce anger are making it obvious that one means to do the man wrong, and being utterly shameless. Further, putting the questions in alternation, whether one has several arguments bearing on the same point, or arguments both that it is so and that it is not so; for it then happens that one's guard must be kept up against several things at once, or against opposites at once. And in general everything said earlier about concealment is useful also for contentious arguments; for concealment is for the sake of going unnoticed, and going unnoticed is for the sake of deception.

42| Against those who shake their heads at whatever they suspect bears on the argument, one must question by way of the negation, as though wanting the opposite, or else frame the question neutrally; for since it is then unclear what one wants granted, they are less reluctant. And whenever someone has granted the particular cases one by one, one need not, in going on to the universal by induction, often ask for it, but should use it as though it had been granted; for sometimes people think they have granted it themselves too, and it appears so to the listeners as well, because of the mention of the induction, as though the particulars would not have been asked in vain. And in cases where the universal is not signified by a single name, one must make use of the resemblance to one's advantage; for the resemblance often goes unnoticed. And to get a premise granted, one should ask by setting it beside its opposite — for instance, if one needed to get it granted that one must obey one's father in everything, ask whether one must obey one's parents in everything or disobey in everything; and for 'often, many things,' whether one should concede many or few; for it would seem, if anything is necessary at all, more likely to be many — since when opposites are set close together, things appear both greater and great, and both worse and better, to people. And what very often produces, in the strongest degree, the appearance of having been refuted is the most sophistical trick of questioners: without having reasoned anything out, not to put the last point as a question but to state it as a conclusion, as though it had been reasoned out — 'so it is not this-and-that.'

43| It is also sophistical, when a paradoxical thesis has been laid down, to demand that the respondent answer with the ordinary, apparent view, when what is actually at issue from the start is the paradoxical opinion, and to frame the question of such matters thus: 'Does it seem so to you?' For necessarily, if the question is drawn from the premises out of which the deduction is built, the result is either a refutation or a paradox: if he grants it, a refutation follows; if he does not grant it but denies that it seems so to him, an implausibility follows; and if he does not grant it but agrees that it does seem so, something refutation-like follows.

44| Further, just as in rhetorical arguments, so too in refutative ones, one must look for the contradictions either against what the man has himself said, or against those whom he agrees speak or act rightly, further against those thought to be such, or against people like them, or against the majority, or against everyone. And just as respondents, when they are about to be refuted, often draw a distinction if it looks as though the refutation is going to succeed, so too when questioning one must sometimes use this same device against those who raise objections — that if the result follows in one way but not in another, it is because he took the term in this particular sense — as Cleophon does in the Mandrobulus. And one must also, when abandoning a line of argument, cut short the remaining attempts at proof; and the respondent, if he senses this coming, should raise his objection and give warning in advance. And sometimes one must attempt an argument against something other than what was actually said, taking that up instead, when one has nothing to argue against the thesis as laid down — which is what Lycophron did when asked to deliver a eulogy of the lyre. And against those who demand that an argument be brought to bear on a single point, since it seems one ought to give the reason, though once certain things have been said the universal outcome is more easily guarded against in refutations, one should state the contradiction — namely that what he affirmed he (elsewhere) denied, or what he denied he affirmed — rather than claim that the same knowledge is, or is not, of contraries. And one must not ask for the conclusion in the form of a premise; and some things must not even be asked at all, but simply used as though agreed.

45| The sources, then, from which the questions are drawn, and how one must question in contentious discussions, have been stated. About answering — how one must solve such arguments, and what one must solve, and for what use arguments of this kind are beneficial — must be discussed next. They are useful for philosophy on account of two things. First, because, arising for the most part around diction, they leave one better equipped to see in how many ways each thing is said, and which things hold alike and which differently, both in the case of things and in the case of names. Second, with a view to inquiries carried on by oneself; for a man easily paralogized by another, and who does not notice it, might often suffer the same thing at his own hands. And third, still remaining, with a view to reputation — namely, to appear to have been exercised in everything and to be inexperienced in nothing. For when people who take part in discussions find fault with arguments without being able to specify anything about what is wrong with them, this gives rise to the suspicion that they are thought to object not on account of the truth but on account of inexperience.

46| How one ought to answer, when responding to arguments of this kind, is clear, provided we have correctly stated earlier what fallacies are made of, and have adequately distinguished the advantages sought in questioning. But grasping an argument and seeing how to resolve its badness is not the same as being able to answer quickly when questioned. For we often fail to recognize what we know once it has been reformulated. Further, just as in other pursuits the being quicker or slower comes about more from practice, so it is too with arguments: so that if the point is clear to us but we are unpracticed, we often fall behind the opportunity. And something happens here, as it does with geometrical diagrams: there too, having sometimes analyzed a figure, we are unable to put it together again; so too in refutations, though we know the point on which the argument turns, we are at a loss to trace back and undo the argument.

47| First, then, just as we say one ought to reason sometimes from reputable premises rather than true ones when that is one's aim, so too one ought sometimes to resolve an argument in the way that is more reputable rather than strictly according to the truth. For in general one must fight against contentious arguers not as people who are refuting but as people who merely appear to refute; for we say they are not really reasoning, so the correction must be aimed at not appearing to be refuted. For if a refutation is a contradiction derived from certain premises without equivocation, there would be no need at all to draw distinctions against ambiguous and equivocal terms — for these do not produce a valid deduction. But the only reason for drawing the further distinction is that the conclusion appears refutation-like. So it is not being refuted but merely seeming to be that one must guard against, since asking ambiguous or equivocal questions, and all such tricks of deception, both obscure genuine refutation and make it unclear whether the person has been refuted or not. For since it is possible, once the conclusion has been drawn, to say that one is denying not what one actually asserted but something merely homonymous with it, even if it happens to point to exactly the same thing, it is unclear whether the person has been refuted; for it is unclear whether he is now speaking the truth. But if he had asked his question by first distinguishing the equivocal or ambiguous term, the refutation would not have been unclear. And what the contentious arguers now demand less, but formerly demanded more — that the person questioned answer either yes or no — would then have come about. As it is, because questioners do not put their questions well, the person answering is forced to add something further, correcting the badness of the premise, since if the questioner divides his premise adequately, the answerer is bound to say either yes or no.

48| But if someone supposes that a refutation by equivocation is a genuine refutation, in a certain way the person answering will not be able to escape being refuted; for in the case of visible things it is necessary that what he affirmed he must also deny, and what he denied he must also affirm. For the way some people try to correct this is of no use. They do not say Coriscus is both musical and unmusical, but rather this Coriscus is musical and this Coriscus is unmusical. For the claim that this Coriscus is unmusical or musical will amount to the same statement as Coriscus, which is at once to affirm and to deny the same thing. But perhaps it does not signify the same thing; for the name does not signify the same thing there either — so that there is some difference. But if to the one he assigns simply saying 'Coriscus' while to the other he adds 'a certain' or 'this', that is absurd; for it is no more true of the one than of the other — for whichever one he takes, it makes no difference. All the same, since it is unclear, when someone has not distinguished the ambiguity, whether he has been refuted or not, while it is granted in argument that one may draw such distinctions, it is plain that to give one's answer to the question without distinguishing, but simply, is simply a mistake — so that even if the person himself has not been refuted, still the argument is like one that has refuted him. It happens often, however, that people who see the ambiguity are reluctant to draw the distinction, because of how frequently such premises are put forward, so as not to seem to be objecting to everything; then, not expecting the argument to turn on this point, they are often met with a paradox. So since it is granted that one may draw the distinction, one must not hesitate to do so, just as was said before.

49| If someone does not treat two questions as one question, then the fallacy that arises through equivocation and ambiguity would not occur either — the result would be either a genuine refutation or none at all. For what difference is there between asking whether Callias and Themistocles are musical, and asking whether the same one name belonged to two different people? For if it signifies more than one thing, he has asked more than one question. If then it is not right to demand a single answer, simply given, to two questions, it is plain that it is proper to give a simple answer to none of the equivocal terms, not even if the thing is true of all of them, as some people claim. For this is no different from asking whether Coriscus and Callias are at home or not at home, whether both are present or neither is present; for in both cases the premises are more than one — the fact that it is true to say so does not make the question one. For it is possible for countless other questions too, asked one after another, all to be truly answerable by yes or no; but nevertheless one should not answer with a single answer, for that would abolish dialectical discussion. This is like assigning the same name to different things. If then one ought not to give a single answer to two questions, it is plain that in the case of equivocal terms too, one should not say simply yes or no. For the one who does so has not really answered but merely spoken. Still, in the course of discussion, people are somehow permitted to do this, because what is happening escapes notice.

50| As we said, then, since even some things that are not really refutations seem to be refutations, in the same way some things that are not really resolutions will seem to be resolutions; and these, we say, one should sometimes bring forward in preference to the true ones, in competitive arguments and in meeting the doubled question. One should answer, in the case of things that merely seem so, by saying 'let it be so' — for in this way a spurious refutation would least occur; but if one is forced to say something paradoxical, at that point above all one must add 'it seems so' — for in this way it will seem that neither a refutation nor a paradox has come about. Since it is clear how one begs the original question, and everyone thinks that, when the case is close, one must reject it and not concede certain things as though the questioner were begging the original point — one should say the same thing when someone demands something that necessarily follows from the position but is false or contrary to reputable opinion; for what follows of necessity is thought to belong to the same position. Further, when the universal is taken not by name but by way of an illustrative comparison, one must say that it is not taken as it was granted or as it was put forward — for it is often on this point that a refutation arises. And when shut out from these resources, one must proceed to say that the point has not been properly demonstrated, meeting the argument according to the distinction stated.

51| In the case of terms used in their proper, strict sense, one must answer either simply or by drawing a distinction. But as for the things we posit while also implying something further along with them — that is, whatever is asked not clearly but in truncated form — it is on this account that the refutation arises: for example, is whatever belongs to the Athenians the property of the Athenians? Yes. And similarly in the other cases. But surely a human being belongs among the animals? Yes. Then a human being is the property of the animals. For we say a human being belongs among the animals because he is an animal, and we say Lysander belongs among the Spartans because he is a Spartan. It is plain, then, that in cases where the premise put forward is unclear, one should not concede it simply. And when, of two things, if one holds it seems necessary that the other holds too, but not the reverse, then when questioned one should grant the lesser first; for it is harder to draw the conclusion from more premises. And if the opponent argues that a thing is contrary to one thing but not to another, then, if the argument is true, one should say it is contrary, but that no name has been assigned to the other thing.

52| Since some of the things the many say, they would call false if one refused to concede them, while others they would not — for example, matters over which opinion is divided (for whether the soul of animals is perishable or immortal has not been settled among the many) — in cases, then, where it is unclear in which of two ways the premise put forward is customarily meant, whether as maxims are meant (for people call maxims both true opinions and general assertions), or as the diagonal is said to be incommensurable, then further, wherever the truth is a matter of disputed opinion, one might most easily shift the terms without being noticed. For because it is unclear which way the truth actually lies, the speaker will not seem to be using sophistry, while because opinion is divided, he will not seem to be lying; but the shift of meaning will make the argument impossible to refute. Further, whatever questions one can anticipate in advance, one should raise objections to and give warning of beforehand; for in this way one would most effectively hinder the questioner.

53| Since the correct solution is the exposure of the false syllogism—showing at which of the questions the falsity arises—and since a false syllogism is spoken of in two ways (either it has actually reasoned to a false conclusion, or it is not really a syllogism though it seems to be one), the solution just described, and also the correction of the apparent syllogism by showing at which question it merely appears to be one, would both count as solutions. So it turns out that we solve arguments either by demolishing those that have genuinely reasoned to a conclusion, or by drawing distinctions in those that merely appear to reason. Again, since among arguments that have genuinely reasoned to a conclusion some have a true conclusion and some a false one, those that are false in respect of the conclusion admit of solution in two ways: either by demolishing one of the things asked, or by showing that the conclusion does not in fact stand as stated. But those that are false in respect of their premises can be solved only by demolishing one of the premises, since the conclusion is true. So those who wish to solve an argument must first examine whether it is a genuine syllogism or not a syllogism at all, and then whether the conclusion is true or false, so that we may solve it either by drawing a distinction or by demolition—and if by demolition, in one of the two ways stated earlier. It makes a great difference whether one solves an argument while it is being put to one in questions or afterward: to foresee it is difficult, but to see it at leisure is easier.

54| Now among refutations that turn on ambiguity or amphiboly, some have one of the questions signifying more than one thing, others have the conclusion said in several ways—for instance, in "speaking while silent" the conclusion is twofold, while in "knowing together with one who knows" it is one of the questions that is ambiguous. And the twofold sense is sometimes actually present, and sometimes not actually present but only appears to be, the apparent double sense being in one respect real and in another not.

55| Now wherever the multiple sense occurs at the end, in the conclusion, no refutation results unless the contradiction is added in beforehand—as in "the blind man sees"; for without a contradiction there was no refutation. But wherever it occurs in the questions, it is not necessary to deny the double sense in advance; for the argument is not aimed at this but proceeds by means of it. So at the outset, in the case of a double name and a double account, one must answer in this way: that it is so in one sense and not so in another—just as with "speaking while silent," that it is so in one sense and not in another. And "one must do the things that are needful" holds in some respects and not in others; for "the things that are needful" is said in many senses. But if this escapes notice at first, it must be corrected at the end by adding to the question: "Is it possible to speak while silent?" "No—but this man, while silent, can be spoken of." And likewise in cases where the multiplicity of sense lies in the premises. "So they do not know together that they know?" "Yes—but not those who know in this particular way"; saying that joint-knowing is impossible outright is a different claim from saying that it is impossible for people who know in this particular way. And in general one must insist, even where the argument has reasoned to a conclusion without qualification, that what the answerer denied is not the very thing he had affirmed, but only the word for it; and so there is no refutation.

56| It is also clear how one must solve refutations that turn on division and combination: for if the sentence, when divided and when combined, signifies something different, one must state whichever is the opposite of what is concluded. All arguments of this kind turn on combination or division. "Is it the case that the man by whom you saw this fellow being struck, this fellow was struck by him?" and "the man by whom he was struck, it was by him that you saw it?" This does draw something even from the ambiguous questions, but it turns on combination. For the fault that turns on division is not a double meaning; the sentence, when divided, does not become the same sentence—unless "a boundary" and "a definition," spoken with different accent, also signify something different when written the same way. But in writing the word is the same, when it is written from the same letters in the same way, and there too people already mark it specially, while the things spoken are not the same. So the fault that turns on division is not a double meaning. It is also clear that not all refutations turn on the double meaning, as some people claim.

57| So the answerer must draw the distinction: it is not the same thing to see with one's own eyes a man being struck, and to say that one sees with one's own eyes that he is struck. And there is also Euthydemus's argument: "Do you, being now in Sicily, know that there are triremes in the Piraeus?" And again: "Can a man who is a good shoemaker be bad?" But a man might be a good shoemaker and yet be a bad man; so there will be "a good shoemaker who is bad." "Are the subjects whose sciences are excellent, excellent subjects of study?" But the study of evil is excellent; therefore evil is an excellent subject of study. But evil is both "evil" and "a subject of study," so that evil is an evil subject of study. Yet there is an excellent science of evils. "Is it true to say now that you have come to be?" "Then you have come to be now." Or does the sentence signify something different when divided? For it is true to say now that you have come to be, but you did not come to be now. "Could you do these things too, in the way and to the extent that you have the power to do them?" "But you have, while not playing the lyre, the capacity for lyre-playing; so you could play the lyre while not playing it." Or rather: he does not have the capacity for playing the lyre while not playing it, but the capacity, for when he is not doing it, of doing it at another time. Some solve this argument differently as well. For they say that if he granted it in the sense of "as he is able to do it," it does not follow that a man not playing the lyre plays the lyre; for it was not granted without qualification that he would do it "in every way he is able to do it"—for "as he is able" is not the same as "in every way he is able." But it is clear that this is not a good solution; for the same solution applies to arguments that turn on the same point, and this one will not fit every case nor every way of putting the question, but bears on the questioner, not on the argument.

58| Arguments that turn on accent do not really occur, either in writing or in speech, except perhaps a few, such as this one: "Is that which you are un-lodging a house?" "Yes." "And is 'not lodging' the negation of 'lodging'?" "Yes." "But you affirmed that what you are un-lodging is a house; so the house is a negation." How this must be solved is clear: for the word pronounced with the acute accent does not signify the same as when pronounced with the grave.

59| It is also clear, for those refutations that turn on things not the same being spoken of in a like way, how one must meet them, now that we have the genera of the categories to hand. For one man, on being questioned, granted that something did not belong which signifies "what a thing is"; the other then showed that something belonging to the relative, or to quantity, did belong—things which, because of the wording, seem to signify "what a thing is"—as in the following argument. "Is it possible for the same thing, at the same time, to be doing and to have done a thing?" "No." "But surely it is possible for the same thing, in the same respect and at the same time, to be seeing and to have seen it." "Is any of the things one undergoes the same as doing something?" "No." "Well then, 'being cut,' 'being burned,' 'perceiving' are said in a like way, and all signify undergoing something; and again 'speaking,' 'running,' 'seeing' are said in a like way to one another; but surely seeing is a kind of perceiving, so that it is also undergoing something and doing something at the same time." Now if someone, having granted there that the same thing cannot at the same time be doing and have done a thing, then says that seeing and having seen can hold together, he has not yet been refuted—unless he also says that seeing is not doing something but undergoing something; for this further question is still needed. But the hearer supposes it has been granted, once he grants that "cutting" is doing something and "having cut" is having done something, and all the other things said in a like way; for the hearer himself adds the rest, on the assumption that it is said in a like way—but it is not said in a like way, though it appears to be, because of the wording. The very same thing happens as in cases of ambiguous names; for in the ambiguous cases, one unfamiliar with arguments thinks that the person has denied the very thing he affirmed, not merely the word for it—but this still needs a further question, whether he means the ambiguous term with a single thing in view; for only if he grants that will there be a refutation.

60| Similar to these are also arguments of the following kind: whether someone who later does not have what he had has thrown it away. For someone who has thrown away only one knucklebone will not have ten knucklebones. Or rather: what a person previously had and now does not have, that he has thrown away, but it is not necessary that he has thrown away as many as the number he now lacks. So having asked about what he has, the arguer draws his conclusion about how many; for the ten are a quantity. If, then, one had asked from the start whether, of whatever number a person previously had and now does not have, he has thrown away that many, no one would have granted it — only that many, or some of them. And also that a person might give what he does not have. For he does not have only one knucklebone. Or rather: he has not given what he did not have, but has given the one as he did not have it — that is, alone. For 'only' does not signify a this, nor a such, nor a so-much, but a relation — namely, that it is not together with another. So just as if one had asked whether a person would give what he does not have, and on his denying it had then asked whether someone not having something quickly would quickly give it, and on his assenting had inferred that a person would give what he does not have — it is clear that no valid inference has been drawn; for 'quickly' does not signify giving this, but giving in this manner —

61| just as a person might give in a manner in which he does not have something, for instance a man who has something pleasantly might give it painfully. Similar too are all arguments of the following kind: would a person strike with the hand he does not have? Or see with the eye he does not have? For he does not have only one. Now some solve this by saying that a man who has more than one eye, or anything else, also has only one — in the sense that he has one and no more of a further set. Others say that what he has, he has received; for the other party was giving only one pebble, and this man, they say, has only one pebble from him. Others do away with the question straightaway, on the ground that it is possible to have what one did not receive — for instance, to receive wine sweet and, once it has spoiled in the receiving, to have it sour. But, as was said before, all these people are answering not the argument but the man. For if this were a genuine solution, it would not be possible to solve it once the opposite had been granted, as in other cases; for instance, if a thing both is and is not — the solution, if one lets it be stated without qualification, is drawn out in the inference; but if no inference follows, there is no solution at all — whereas in the cases mentioned above, even granting everything, we say no syllogism results.

62| Further, the following also belong among these arguments. Did someone write what has been written? Now it has been written just now that you are sitting, which is a false statement; but it was true when it was being written; so it was being written both false and true at once. For to say that an account or an opinion is false or true signifies not a this but a such — for the same account applies to opinion as well. And again: is what the learner learns the very thing he learns? Now someone learns the slow quickly. So he has said, not what he learns, but the manner in which he learns it. And again: does a person tread what he walks? Now he walks the whole day. Or rather, he has said not what he walks but when he walks; nor, when someone is said to be drinking the cup, does he drink what he drinks, but that from which he drinks. And again: does a person know what he knows either by learning or by discovering? But of the things some he discovered, some he learned; neither applies to both together. Or rather: the one term applies to the whole taken as a unity, the other to all the parts. And so too the argument that there is a third man beside the man himself and the individual men. For 'man,' and every common term, signifies not a this, but a such, or a relation, or a how, or something of that sort. And likewise in the case of Coriscus and musical Coriscus — are they the same or different?

63| For the one signifies a this, the other a such, so that it cannot be set out as a single thing; and it is not the setting-out that produces the third man, but the conceding that it is a this precisely. For it will not be possible for that to be a this precisely which is what Callias is and what man is. Nor will it make any difference if someone says that what is set out is not a this precisely but a such precisely; for there will still be some one thing beside the many, for instance man. It is clear, then, that one must not grant that what is predicated in common of all of them is a this, but rather that it signifies either a such, or a relation, or a quantity, or something of that sort.

64| In general, in arguments that depend on the expression, the solution will always lie in the opposite of that on which the argument turns. For instance, if the argument turns on composition, the solution is division; if on division, composition. Again, if it turns on an acute accent, the grave accent is the solution; if on a grave accent, the acute. If it turns on ambiguity of a word, one can solve it by stating the opposite term — for instance, if it results in saying that something is ensouled, one denies this by asserting that it is, showing how it is ensouled; and if he said 'soulless' but the other inferred 'ensouled,' one should say how it is soulless. Likewise too with amphiboly. And if it turns on similarity of expression, the opposite will be the solution. Would a person give what he does not have? Rather, not what he does not have, but as he does not have it — for instance, only one knucklebone. Does a person, when he knows something, know it by learning or discovering it? But not the things he knows. And if a person treads what he walks — rather, not what, but when. Likewise too in the other cases.

65| Against those arguments that depend on accident, there is one and the same solution for them all. Since it is indeterminate when one ought to say something holds of the thing, given that it holds of the accident, and in some cases it seems and people say it must, in others they say it need not — one must say, having brought it round in the same way for all cases, that it is not necessary. But one must be ready to bring forward the particular example. All arguments of the following kind depend on accident. Do you know what I am about to ask you? Do you know the man approaching, or the man muffled up? Is the statue your work, or is the dog your father? Are few times few, few? For it is clear in all these cases that what is true of the accident need not also be true of the thing; for it is only for things that do not differ in substance and are one that everything seems to hold in common. But for the good, being good and being about to be asked a question are not the same, nor for the approaching or muffled-up man, being approaching and being Coriscus; so that it is not the case that, if I know Coriscus but do not know the approaching man, I both know and do not know the same man; nor, if this is mine and is a work, is it my work — rather it is a possession, or a thing, or something else. And in the same way too with the other cases.

66| Some solve this by destroying the question itself: they say it is possible to know and not know the same thing, but not in the same respect; so, not knowing the man approaching but knowing Coriscus, they say the same thing is both known and not known, but not in the same respect. And yet, first, as we have already said, the correction must be the same for arguments that turn on the same point; but this will not hold if one takes the same premise not with respect to knowing but with respect to being, or being in some way — for example, whether this man is a father, and is yours. For if in some cases this is true, and it is possible to know and not know the same thing, still here what has been said has nothing in common with that case. And nothing prevents the same argument from having several faults. But not every exposure of an error is a solution: for it is possible to show that someone has drawn a false conclusion, without showing the point on which he went wrong — for example, Zeno's argument, that motion is impossible. So too if someone tries to conclude that it is impossible, he is in error, even if he has argued the syllogism ten thousand times; for this is not a solution. For the solution was the exposure of a false syllogism — the point at which it is false; so if no conclusion has actually been drawn, or he is trying to draw a true or false conclusion without one, pointing that out is the solution. Perhaps in some cases nothing prevents this too from happening; but in these cases at least it would not seem to. For one knows that Coriscus is Coriscus, and that the man approaching is approaching. It does seem possible to know and not know the same thing — for example, to know that it is white, but not recognize that it is musical; for in this way one knows and does not know the same thing, but not in the same respect. But the man approaching, and Coriscus — both that he is approaching and that he is Coriscus — one knows.

67| Similarly those also err who solve the puzzle that every number is small, as we mentioned of those before: for if, passing over the fact that no conclusion has been drawn, they say the conclusion has truly been reached — that everything is both many and few — they are in error. Some also solve syllogisms by appeal to double meaning, for example, that this man is your father, or son, or slave. And yet it is clear that if the refutation appears to arise from a word being said in several ways, the name or the account must properly belong to more than one thing; but no one properly calls this man the child of that man merely because he is master of the child — rather the combination rests on accident. Is this yours? Yes. And this is a child. Then this is your child — no: because it happens to be both yours and a child, but not that it is your child.

68| And likewise the claim that something of the evils is good; for practical wisdom is knowledge of evils. But 'this belongs to these' is not said in several ways, but rather as a possession is said. If it were indeed said in several ways (for we also say man is of the animals, yet not at all as a possession), and if something is said in relation to evils as belonging to some class, then on that account it is 'of the evils' — but this thing itself is not thereby 'of the evils.' So the appearance rests on the qualified and the unqualified. And yet perhaps it is possible for something to be good of the evils in two ways — though not in this argument, but rather if some slave were good, good belonging to a bad man, in that sense more so. Perhaps not even that: being good and belonging to a given thing does not, by itself, make something good-of-that-thing at the same moment. Nor is saying that man is of the animals said in several ways; for it is not the case that whenever we signify something by leaving part out, that expression is thereby said in several ways — for indeed, by speaking half the line, we signify 'Give me the Iliad,' for instance by saying 'Sing, goddess, the wrath.' Those fallacies that turn on something being said properly as 'this,' or in a certain respect, or place, or manner, or in relation to something, and not without qualification, must be solved by examining the conclusion against the contradictory, to see whether it admits of being affected in one of these ways. For it is impossible for contraries and opposites, and affirmation and negation, to belong to the same thing without qualification; but nothing prevents each from belonging in a certain respect, or in relation to something, or in a certain manner — or the one without qualification, the other in a certain respect. So if this belongs without qualification and that in a certain respect, there is not yet a refutation. And this must be examined by looking at the conclusion against the contradictory.

69| All such arguments share this feature. Is it possible for what is not to be? But surely there is something that is not. Likewise, what is will also not be; for it will not be one of the things that are. Is it possible for the same man at the same time both to keep his oath and to break it? Is it possible for the same man at the same time both to obey and disobey the same thing? Or is 'being something' not the same as 'being' without qualification? And with regard to what is not: it does not follow, if something is something, that it also is without qualification; nor, if a man keeps his oath in this or that respect, is it necessary that he keeps it without qualification — the man who swore that he would break his oath keeps his oath only in this, that he breaks it, but he does not keep his oath without qualification; nor does the man who disobeys thereby obey — rather, he obeys in some respect. The argument is similar too concerning the same man both lying and speaking the truth at the same time; but because it is not easy to see how, in either direction, one should assign 'speaking the truth without qualification' or 'lying without qualification,' it appears difficult. But nothing prevents him from being false without qualification, yet true in a certain respect or of something — and from being true of some things while not being true without qualification. Similarly too in the case of relatives, and place, and time; for all such arguments turn on this point. Is health, or wealth, a good?

70| But to the fool, and to one who does not use it rightly, it is not good; so it is both good and not good. Is being healthy, or holding power in the city, a good? But there are times when it is not better; so the same thing is both good and not good for the same man. Or does nothing prevent a thing that is good without qualification from not being good for this man — or being good for this man, but not now, or not good here? Is what the practically wise man would not wish for bad? But he does not wish to throw away the good; so the good is bad. For it is not the same to say that the good is bad, and that throwing away the good is bad. Similarly too the argument about the thief. For it is not the case that, if the thief is bad, the taking is also bad; so he does not wish for the bad, but the good — for taking a good thing is good. And disease is bad, but getting rid of disease is not. Is the just more choiceworthy than the unjust, and acting justly than acting unjustly? But dying unjustly is more choiceworthy. Is it just for each man to have what is his own? But whatever someone judges according to his own opinion, even if false, is authoritative by law; so the same thing is both just and not just. And which ought one to judge right — the one who states the just things, or the one who states the unjust things?

71| But surely it is also just for the man who has been wronged to state adequately what he suffered — and these things were unjust. For it is not the case that, if suffering something unjustly is choiceworthy, the unjust is thereby more choiceworthy than the just; rather, without qualification the just is more choiceworthy, though nothing prevents this particular thing from being more choiceworthy unjustly or justly, depending on the case. And having what is one's own is just, and having what belongs to another is not just; yet nothing prevents this judgment from being just — for example, if it accords with the judge's own opinion — for it is not the case that, if this or that is just, it is also just without qualification. Likewise too, though the things said are unjust, nothing prevents calling the saying of them just; for it is not the case that, if it is just to say them, they must therefore be just — just as, if it is beneficial to say something, it is not thereby beneficial. Likewise too in the case of just things. So it is not the case that, if the things said are unjust, the one who says the unjust things wins; for he says what it is just to say, though the things themselves are, without qualification, also unjust to suffer.

72| For refutations that arise in violation of the definition of a refutation, as was sketched out earlier, one must respond by examining the conclusion in relation to the contradiction, to see whether it will be the same thing, in the same respect, in relation to the same thing, in the same way, and at the same time. But if this is put forward as a premise right at the start, one must not concede that it is impossible for the same thing to be both double and not double; rather, one must grant it — though not in the way in which being refuted was once agreed to hold. All these arguments turn on a defect of this kind. Does someone who knows of each thing that it is that thing thereby know the thing? And does someone who is ignorant likewise not know it? Now someone who knows Coriscus, in that he knows he is Coriscus, might not know that he is musical, so that he both knows and does not know the same thing. Is the four-cubit thing greater than the three-cubit thing? A thing three cubits long might become four cubits long; and what is greater is greater than what is smaller; therefore the thing is itself greater than itself, and smaller than itself. As for those refutations that turn on begging and assuming the very point at issue: if this is asked and it is obvious that this is what is being asked, it must not be granted, even if it is a reputable opinion, and the truth must be stated instead. But if it goes unnoticed, then, on account of the badness of arguments of this kind, the fault must be turned back onto the questioner, as one who has not properly argued; for a refutation must proceed without assuming the point at issue. Then one must say that it was granted not on the supposition that he would use it as he did, but on the supposition that he would draw the opposite conclusion from it — or else the case must be treated as one of the merely apparent refutations.

73| As for those who force an agreement through the fallacy of the consequent, this must be exposed on the argument itself. The following of consequents is twofold. Either it is as the universal follows the particular, as animal follows man; for it is claimed that if this goes with that, then that too goes with this. Or it is in accordance with opposites: for if this follows that, then the opposite follows the opposite. It is on account of this that Melissus's argument too proceeds: for if what has come to be has a beginning, he claims that what has not come to be has none, so that if the heaven is ungenerated, it is also unlimited. But this is not so; for the consequence runs the other way round. As for those who draw their conclusion by adding something, one must examine whether, when it is removed, the impossibility results none the less. And then this must be pointed out, and it must be said that he granted it not because it seemed true but for the sake of the argument, while the other has made no use of it for the argument.

74| Against those who make several questions into one, the distinction must be drawn right at the start. For a question is single when there is a single answer to it, so that one must neither say several things of one subject nor one thing of several subjects, but must affirm or deny one thing of one thing. Just as, in the case of equivocal terms, a predicate sometimes belongs to both senses and sometimes to neither, so that when the question is not simple, no harm comes to those who answer it simply, the same holds in these cases too. So then, when several things belong to the one thing, or the one thing belongs to several, no contrary result comes about for the person who has answered simply and so made this kind of mistake. But when it belongs to the one and not the other, or several things are predicated of several subjects, and in one way both belong to both while in another way again they do not, then one must be on guard against this. As, for example, in the following arguments. If one thing is good and the other bad, the sophist argues that it is true to say that these, taken together, are good and bad, and again neither good nor bad; for neither of them is the other, so that the same pair is both good and bad and neither good nor bad. And if each thing is the same as itself and different from something else, then, since they are not the same as other things but as themselves, and are different from themselves, it follows that the same things are, to themselves, both different and the same. Further, if the good becomes bad, and the bad is good, then two things would come to be out of one. And, being two and unequal, each of them is equal to itself, so that they are, to themselves, both equal and unequal.

75| Now these arguments fall under other solutions as well; for both 'both' and 'all' signify more than one thing, so that it is not the same thing, except in name, that turns out to be both affirmed and denied — and this was not a refutation. But it is clear that, when the question is not a single one but several are being asked, and one thing is affirmed or denied of one thing, the impossibility will not result.

76| Concerning those arguments that lead one to say the same thing many times over, it is clear that one must not concede, in the case of relative terms, that the predicates signify anything when separated and taken by themselves — for instance, that 'double' signifies something apart from 'double of half' — on the ground that the half is already implied within it. For indeed 'ten' is implied in 'ten short of one', and 'doing' is implied in 'not doing', and in general the affirmation is implied in the negation; but all the same, if someone says that this thing is not white, he is not thereby saying that it is white. But perhaps 'double' by itself signifies nothing at all, any more than the term embedded within 'half' does; and if after all it does signify something, it is not the same thing as what it signifies when combined with its correlate. Nor is 'knowledge' already contained in the species — for example, it does not follow that if medicine is knowledge, it is knowledge in the sense of the general term; for that, the general term, was knowledge of a knowable object. In the case of predicated terms that are made clear by means of other terms, this must be said: that what is indicated is not the same taken separately as it is within the full account. For 'concave' signifies the same thing in general in the case of 'snub' and of 'bandy-legged', but there is nothing to prevent it, once something is added, from differing; rather, one applies to the nose and the other to the leg, for in the one place it signifies 'snub', in the other 'bowed'; and it makes no difference to say 'snub nose' or 'concave nose'. Further, one must not concede the expression taken straightforwardly; for it is false. For 'snub' is not 'a concave nose', but a certain something belonging to a nose, an affection of it, so to speak, so that there is nothing absurd in the fact that the snub nose is a nose having concavity of nose.

77| Concerning solecisms, we have already said earlier what it is on account of which they appear to occur; how they are to be resolved will become clear from the arguments themselves. For all arguments of this sort aim to establish the following. Is what you truly say also truly the case? Now you say that something is 'a stone'; therefore something is 'a stone'. Or rather, to say 'a stone' is not to say 'what' (neuter) but 'whom' (masculine), nor 'this' (neuter) but 'this' (masculine). If, then, someone were to ask, 'Is he whom you truly speak of, this one (masculine)?' he would not seem to be speaking correct Greek — any more than if he asked, 'Is she whom you say to be, this one (masculine)?' But it makes no difference if he had said 'wood' (neuter), or named anything else that signifies neither feminine nor masculine. That is why no solecism results when one asks: 'Is what (neuter) you say to be, this (neuter)?' — 'You say wood to be the case'; 'therefore it is wood.' But 'stone' and 'this' both take the masculine form. So if someone were to ask, 'Is this (masculine) that (feminine)?' and then again, 'Well then — is this (masculine) not Coriscus?' and then said, 'therefore this (masculine) is that (feminine)', he has not validly inferred the solecism — not even if 'Coriscus' signifies the very thing that 'that' (feminine) does, since the answerer does not grant this in his reply, but it must further be put to him as a question. And if it neither is so nor is granted, he has drawn no valid inference either in reality or in relation to what was asked.

78| Likewise, then, there too 'stone' must signify 'this' (masculine). And if it neither is so nor is granted, the conclusion must not be stated; the appearance of validity arises because a dissimilar case-form of the noun looks like a similar one. Is it true to say that this (feminine) is what you say it to be? Now you say it to be 'a shield'; therefore this (feminine) is a shield. Or rather, it is not necessary — since 'this' (nominative) does not signify 'shield' (accusative) but 'shield' (nominative), while 'shield' (accusative) signifies 'this' (accusative). Nor does it follow, if what (neuter) you say to be this (accusative, masculine) is this (nominative, masculine), and you say it to be Cleon (accusative), that therefore this (nominative) is Cleon (accusative). For 'this' (nominative) is not 'Cleon' (accusative); for what has been said is that what I say to be this (accusative) is this (nominative), not this (accusative); indeed the question, put in that way, would not even be correct Greek. 'Do you know this (accusative)?' — 'And this (nominative) is a stone (nominative)'; 'therefore you know a stone (nominative).' Or rather, 'this' does not signify the same thing in 'do you know this?' as in 'this is a stone' — in the first case it means 'this' (accusative), in the second, 'this' (nominative). 'Do you know that of which you have knowledge?' — 'And you have knowledge of a stone (genitive)'; 'therefore you know of a stone (genitive).' Or rather, one thing is meant by 'of this, of a stone' (genitive), another by 'this, a stone' (accusative); and what was granted was that you know — with the accusative — that of which you have knowledge (genitive), not 'of this', but 'this' — so that what you know is not 'of a stone' but 'the stone' (accusative).

79| It is clear, then, from what has been said, that arguments of this kind do not validly infer a solecism but only appear to, and why they appear to, and how one must respond to them.

80| We must also understand that, among all arguments, some are easier and some harder to see through -- that is, to see through what point and in what respect they mislead the hearer -- even though they are often the same arguments as one another. For we must call an argument the same when it arises from the same source. But the same argument may seem to some to be a fallacy of language, to others a fallacy of accident, and to others a fallacy of some other kind, because each term, when transferred to a new case, is not equally clear. So then, just as with fallacies of ambiguity -- which seem to be the silliest kind of fallacy -- some are obvious even to ordinary people (indeed nearly all jokes turn on language: for example, 'A man carried a chair down the stairs'; 'And where are you sailing to?' 'To the yard-arm'; 'And which of the two cows will calve first?' 'Neither -- both calve from behind'; 'And is the North Wind pure?' 'Not at all -- it has killed the beggar and the buyer'; 'Is that Euarchus?' 'No, it's Apollonides' -- and most of the rest work the same way), others evidently escape the notice even of the most experienced people. A sign of this is that they often dispute about names -- for instance, whether 'being' and 'one' signify the same thing in every case, or something different. For to some it seems that 'being' and 'the one' signify the same thing; while others solve the argument of Zeno and Parmenides by saying that 'the one' and 'being' are said in many ways. Likewise, concerning accident and each of the other kinds, some arguments will be easier to see through and others harder; and it is not equally easy in every case to grasp under which genus a fallacy falls, and whether it is a refutation or not.

81| An argument is sharp when it produces the greatest perplexity; for such an argument bites hardest. Perplexity is of two kinds: one occurs in arguments that have been validly syllogized -- the perplexity of which of the questions one should reject; the other occurs in contentious arguments -- how one is to answer what has been proposed. That is why, among syllogistic arguments, the sharper ones make one search harder. A syllogistic argument is sharpest when, from premises that are as reputable as possible, it destroys a conclusion that is as reputable as possible. For being a single argument, when the contradictory is substituted, it will have all its syllogisms alike; since it will always, from reputable premises, likewise either destroy or establish a reputable conclusion, which is why one is bound to be perplexed. The sharpest argument of this kind, then, is the one that makes its conclusion equal in reputability to its questions; second is the one proceeding from premises that are all alike in reputability, for this too will likewise produce perplexity as to which of the questions must be rejected. And this is difficult: for one of them must be rejected, but which one must be rejected is unclear. Among contentious arguments, the sharpest is the one where it is unclear right from the start whether it has been validly syllogized or not, and whether the solution lies in denying a falsehood or in drawing a distinction; second among the rest is the one that is clearly to be solved by distinction or denial, but where it is not evident by rejecting or distinguishing which of the questions asked it must be solved -- that is, whether the fault lies in the conclusion or in one of the premises.

82| Now sometimes an argument that has not been validly syllogized is simple-minded, if its premises are excessively implausible or false; but sometimes it does not deserve to be dismissed. For whenever one of the relevant questions is missing -- the one the argument concerns and proceeds through -- and the arguer has not added this and has not validly syllogized, the syllogism is silly. But whenever what is missing is one of the external premises, the argument is by no means to be despised; rather the argument itself is sound, and it is the questioner who has not asked his questions well. And just as one can solve an argument sometimes with reference to the argument itself, sometimes with reference to the questioner and the question, and sometimes with reference to neither of these, so likewise one can also question and construct syllogisms with reference to the thesis, with reference to the answerer, and with reference to the time -- whenever the solution requires more time than the present occasion allows for arguing one's way through to it. From how many sources, then, and of what kinds, fallacies arise for people engaged in discussion, and how we shall show someone to be speaking falsely and make him say paradoxical things; further, from what premises a syllogism results, and how one must put questions, and what the order of the questions should be; further, for what purpose all arguments of this kind are useful, and about answering -- every kind of answering in general -- and how arguments and syllogisms are to be solved: let this be all we have to say on these points. It remains, having reminded ourselves of our original purpose, to say something brief about it and to put an end to what has been said.

83| We set out, then, to discover a certain capacity for reasoning syllogistically about a given problem, from the premises available that are as reputable as possible; for this is the function of dialectic in its own right, and of the art of testing. But since, because of its kinship with sophistic, something further gets built onto this capacity -- namely, that it can not only test dialectically but can also speak as one who knows -- for this reason we set down as the task of this treatise not only the function just mentioned, being able to get hold of an argument, but also how, in submitting to be questioned, we shall defend our thesis in the same way, through premises as reputable as possible. We have already stated the reason for this: it was for this same reason that Socrates used to ask questions but not answer them, for he admitted that he did not know. It has been shown in what precedes both with a view to how many things, and from how many premises, this will be achieved, and where we shall find an ample supply of these; further, how the whole line of questioning must be put or arranged, and about answers and about solutions to syllogisms. It has also been shown concerning the other matters that belong to this same method of argument. In addition to this we have gone through the fallacies, as we have already said before. That the tasks we set out to do, then, have reached a sufficient conclusion is clear; but we must not fail to notice what has happened in the course of this undertaking.

84| Of all discoveries, those taken over from others and previously worked on in part have been advanced bit by bit by those who received them later; but discoveries made from the very beginning generally receive only a small advance at first, though one far more useful than the growth that comes from it later. For perhaps the beginning of anything is its greatest part, as the saying goes, and that is also why it is the hardest; for the more powerful it is in capacity, the smaller it is in size, and so the harder it is to see. But once this beginning has been found, it is easier to add to it and build up the rest -- which is exactly what has happened with rhetorical speeches, and indeed with practically all the other crafts as well. For those who found the starting points advanced them only a little way in all; but those held in high repute today, having received the subject from many predecessors who had advanced it bit by bit in succession, so to speak, have built it up in this way -- Tisias coming after the first pioneers, Thrasymachus after Tisias, Theodorus after him, and many people having contributed many parts. That is why it is no wonder that the craft has attained some considerable bulk. But of this undertaking of ours, it was not the case that part had been worked out before and part had not -- rather, nothing at all existed beforehand. Indeed, the training given by those who worked for hire in contentious arguments was something like the practice of Gorgias.

85| For some of them gave their pupils rhetorical speeches to memorize, others gave them speeches in question-and-answer form -- the very speeches into which each side thought the other's arguments would most often fall. That is why the teaching they gave their students was quick, but unsystematic; for they supposed they were educating them by handing over, not a craft, but the products of a craft -- just as if someone, claiming he would hand down a science for keeping the feet from ever suffering pain, were then not to teach the craft of shoemaking, nor the means by which one will be able to procure such things for oneself, but were instead to give many kinds of shoes of every sort; for such a person has helped meet the need, but has not handed down a craft. Now concerning rhetoric there already existed much that had been said, and of long standing; but concerning syllogistic reasoning we had nothing whatever to say beforehand -- instead we labored for a long time, searching by trial and error. So if, on examining it, our method appears to you to hold up well enough, starting from such beginnings, by comparison with the other studies that have grown through a long tradition, then it remains for all of you, or for those who have heard these lectures, to grant forgiveness for what has been left out of the method, and much gratitude for what has been discovered.

← All of Aristotle: The Complete Works of Aristotle