Aristotle · a new plain-English translation from the original language
1| The commonplaces, then, by means of which we shall be well supplied for arguing about each of the problems, have been enumerated fairly adequately. Book Theta. After this we must speak about order, and about how one ought to put questions. First, before one is going to question, one must find the commonplace from which the argument is to be drawn; second, one must put the questions and arrange them, each in relation to the others, for oneself; and the remaining, third, task is to state these, now, to another person. Up to the point of finding the commonplace, the inquiry of the philosopher and of the dialectician is alike, but the arranging and putting of questions is peculiar to the dialectician; for all of this is directed toward another person, whereas the philosopher, inquiring on his own, does not care at all if the answerer refuses to grant them, provided that the premises through which the reasoning proceeds are true and known, because they lie too close to the starting point and he can foresee what will result; rather he would perhaps even be eager for the axioms to be as knowable and as close as possible, since it is from these that scientific syllogisms proceed. The commonplaces, then, from which one must take one's premises have been stated earlier; about order and about questioning we must speak, first distinguishing the premises that must be taken besides the necessary ones. The necessary premises are those through which the syllogism comes about. Those taken besides these are of four kinds: either for the sake of induction, so that the universal may be granted, or for bulk of the argument, or to conceal the conclusion, or to make the argument clearer.
2| Besides these no other premise should be taken, but one must try to amplify and to question by means of these. Those aimed at concealment are for the sake of contest; but since all such activity is directed toward another person, one is bound to make use of these too. Now the necessary premises, through which the syllogism proceeds, should not be put forward straightaway, but one should hold off from them as far as possible, for example, not requiring that the same science covers contraries, if one wishes to secure this, but rather that it covers opposites; for once this is granted, one will also be able to infer that the same science covers contraries, since contraries are opposites. And if the answerer will not grant this, it must be taken by induction, putting it forward in the case of particular contraries. For the necessary premises must be taken either by syllogism or by induction, or some by induction and some by syllogism, while those that are exceedingly obvious may simply be put forward as they are; for what will result is always less clear in the case of holding off and of induction, and at the same time, putting forward the useful premises themselves is ready to hand for one who cannot obtain them in the other way. The premises spoken of besides these must be taken for the sake of the above, and each is to be used as follows: leading by induction from particulars to the universal and from the known to the unknown; and more known are the things according to perception, either absolutely or to the many.
3| In concealing, one should reason out beforehand the things through which the syllogism regarding the original point is going to come about, and as many of these as possible. This would happen if one reasoned out not only the necessary premises but also some of those useful in relation to these. Further, one should not state the conclusions but reason them out later all together; for in this way one would remove them furthest from the original position. To put it generally, one who inquires by way of concealment ought to question in such a way that, when the whole argument has been asked and the conclusion stated, it is sought why it is so. This will happen best by the method described before; for when only the final conclusion has been stated it is unclear how it comes about, because the answerer does not foresee from what it results, since the earlier syllogisms have not been articulated. The syllogism of the conclusion would be least articulated if we do not lay down its premises directly, but rather those from which the syllogism producing them comes about. It is also useful not to take the axioms from which the syllogisms proceed in a continuous series, but to alternate between arguments directed toward one conclusion and toward another; for when the related premises are placed side by side, what will result from them is more evident. One should also, wherever it is possible, take the universal premise by definition, not applied to the things themselves but to what is coordinate with them;
4| for people deceive themselves when the definition is taken in application to what is coordinate, thinking that they are not conceding the universal. For example, if one needed to secure that the angry man desires vengeance because of an apparent slight, and it were secured that anger is desire for vengeance because of an apparent slight, it is clear that once this is secured we would have, universally, what we intended to establish. But those who put the premise forward in application to the things themselves often find the answerer refusing it, because his objection has more purchase there — for example, that the angry man does not desire vengeance, since we are angry with our parents but do not desire vengeance on them. Now perhaps the objection is not true, for with some people it is enough vengeance simply to pain them and make them regret it; nonetheless it has some plausibility, so as not to seem to be refusing what is put forward without reason. But in the case of the definition of anger it is not equally easy to find an objection. Further, one should put a premise forward not as being wanted for its own sake but for the sake of something else, since people are wary of what is useful for the thesis. Generally speaking, one should make it as unclear as possible whether one wants to secure the premise put forward or its opposite; for when it is unclear which is useful for the argument, people are more inclined to lay down what seems true to themselves.
5| Further, one should inquire by way of resemblance; for this is persuasive, and the universal is more likely to pass unnoticed. For example, that just as knowledge and ignorance of contraries are the same, so too perception of contraries is the same, or conversely, since perception of them is the same, so too is knowledge. This is like induction, yet not the same thing; for there, the universal is taken from the particulars, whereas in the case of resemblances what is taken is not the universal under which all the resembling things fall. One should also sometimes bring an objection against oneself; for answerers are unsuspicious toward those who seem to be arguing fairly. It is also useful to add that such a point is customary and commonly said; for people are reluctant to disturb what is habitual when they have no objection to it. At the same time, by using such points themselves they are on guard against disturbing them. Further, one should not press the point earnestly, even if it is generally useful; for people resist more against those who press earnestly. And one should put things forward as though in a comparison; for what is put forward for the sake of something else, and not for its own sake, people set down as useful more readily. Further, one should not put forward the very point that needs to be taken, but that from which it follows of necessity; for people concede this more readily, because what will result is not equally evident from it, and once this is taken, that other point has also been taken along with it.
6| And ask last the thing you most want to get granted; for people are most likely to shake their heads at the first questions, because most questioners state first the points they are most eager about. But against some people, propose such things first; for the difficult ones grant the first points most readily, if what will follow is not entirely obvious, and grow difficult at the end. Similarly also with those who think they are sharp in answering; for having granted most of the points, at the end they quibble that the conclusion does not follow from what has been laid down; but they grant things carelessly, trusting their own cleverness and supposing they will suffer nothing from it. Further, there is drawing the argument out and inserting things of no use for it, as those who draw false diagrams do; for when there are many things present, it is unclear in which of them the falsity lies. Hence questioners sometimes go unnoticed when, in the midst of the rest, they insert points which, if proposed by themselves, would not be granted. For concealment, then, one should use the devices mentioned; for ornament, induction and division of things akin to the subject. Now what sort of thing induction is, is clear; and division is of this sort: that one science is better than another either by being more exact or by being about better things, and that of the sciences some are theoretical, some practical, some productive.
7| For each of these things helps adorn the argument, but they are not necessary for stating the conclusion. For clarity one should bring in examples and comparisons — examples that are familiar and drawn from what we know, such as Homer's, not such as Choerilus's; for so the point proposed will be clearer. In discussion one must use deduction more against dialecticians than against ordinary people, and induction, on the contrary, more against ordinary people; this has also been said before. Now in some cases, by inducing, one can ask for the universal, but in others it is not easy, because there is no name lying ready that is common to all the similar cases; but when one needs to take the universal, they say 'so in all such cases' — and it is one of the hardest things to determine which of the things brought forward are of this kind and which are not. And on this account people often mislead one another in arguments: some claiming that things not alike are alike, others disputing that things alike are not alike. Hence one should try, in all such cases, to coin a name for oneself, so that the answerer cannot dispute that what is being added is not said in the same manner, nor can the questioner play the sycophant by claiming it is said in the same manner, since many things not said in the same manner appear to be said so. But when, in inducing over many cases, the answerer does not grant the universal, then it is fair to demand an objection from him. But if he himself has not said in which cases it holds, it is not fair to demand of him in which cases it does not hold; for the one inducing must first demand the objection in this way.
8| One should also require that objections not be brought against the very thing proposed, unless that is the only instance of its kind — as the dyad is the only even number that is prime; for the objector must bring the objection against some other instance, or say that this alone is of this kind. Against those who object to the universal, not bringing the objection against the universal itself but against the equivocal term — for instance, that one might have a color, or a foot, or a hand, that is not one's own (for a painter might have a color, and a cook a foot, that is not his own) — one should make a distinction and put the question accordingly in such cases; for when the equivocation goes unnoticed, the objection will seem to be a good one against the premise. But if the objector, not bringing the objection against the equivocal term but against the thing itself, blocks the question, then one must take away that in which the objection lies and propose the remainder as universal, generalizing it, until one gets what is useful. For instance, in the case of forgetting and having forgotten: they do not concede that one who has lost knowledge has forgotten, on the ground that, the situation having changed, he has lost the knowledge but has not forgotten it. One should, then, take away that in which the objection lies and state the remainder — for instance, that if, the situation remaining the same, he has lost the knowledge, he has forgotten it.
9| Similarly also against those who object that a greater evil is opposed to a greater good; for they bring forward that to health, which is a lesser good than good condition, a greater evil is opposed — since disease, they say, is a greater evil than bad condition. One must, then, take away, in this case too, that in which the objection lies; for once it is taken away, the answerer would more readily grant the point — for instance, that a greater evil is opposed to a greater good, provided neither of the two entails the other, as good condition entails health. And this must be done not only when an objection is raised, but even if, without an objection, the answerer denies the point because he foresees something of this kind; for once that in which the objection lies has been taken away, he will be forced to grant it, because he cannot foresee, in what remains, a case where it is not so; and if he does not grant it, when an objection is demanded of him he will not be able to produce one. Premises of the following sort are such as to be false in one respect and true in another; for in these cases it is possible to take away the false part and leave the remainder true. But if, when one is proposing something that holds over many cases, the answerer brings no objection, one should require him to grant it; for a dialectical premise is one against which, holding thus in many cases, there is no objection. And when it is possible to draw the same conclusion both without the impossible and through the impossible, for one who is demonstrating and not conversing in dialectic it makes no difference whether he draws it one way or the other; but for one conversing in dialectic with another, he should not use the deduction through the impossible.
10| For when one has drawn a conclusion without the impossible, there is no room to dispute it; but when one draws the impossible conclusion, unless it is an extremely manifest falsehood, people say it is not impossible, so that questioners do not get what they want. One should propose such things as hold in this way in many cases, where either there is no objection at all, or it is not easy to see one at a glance; for the answerer, unable to see the cases where it does not hold, posits it as true. One should not put the conclusion itself as a question; otherwise, if the answerer shakes his head, it will seem that no deduction has taken place. For often, even when one does not ask but states it as a consequence, they deny it — and in doing this they do not seem to be refuted, in the eyes of those who do not see what follows from what has been laid down. So when one asks without even having stated that the point follows, and the answerer denies it, it seems that no deduction at all has taken place. Nor does every universal question seem to be a dialectical premise — for instance, 'What is man?' or 'In how many ways is the good spoken of?'; for a dialectical premise is one to which it is possible to answer yes or no, but to the questions just mentioned it is not. Hence such questions are not dialectical, unless the questioner himself, having drawn a distinction or made a division, puts it that way — for instance, 'Is the good spoken of in this way, or in that way?'
11| for to such questions the answer is easy, either by affirming or by denying. One should therefore try to propose premises of this sort in that form. And at the same time it is perhaps also fair to ask the other person in how many ways the good is spoken of, when, though he himself divides it up and proposes it, he grants it in no way at all. Whoever asks about one and the same argument at great length inquires badly. For if it is because the one being questioned keeps answering what is asked, clearly the questioner is asking many questions, or asking the same things many times, so that he is either babbling or has no deduction — for every deduction is drawn from few premises; but if it is because the other does not answer, it is because the questioner either fails to press the point or gives up.
12| It is hard to argue for, and easy to maintain, the very same hypotheses. Such are both the things that are first by nature and the things that are last. For the first need a definition, while the last are worked out through many steps by anyone who wants to take the continuous chain from the first things — otherwise the arguments appear sophistical; for it is impossible to demonstrate anything without starting from its proper principles and stringing the chain together as far as the last terms. Now answerers do not think it right to give definitions, nor, even when the questioner does define something, do they pay attention to it; and since it has not been made clear what the thing proposed actually is, it is not easy to argue about it. This happens most of all in the case of principles; for the other things are proved by means of these, while these cannot be proved by means of anything else, but each of them must necessarily be known through definition. Also hard to argue for are the things very close to a principle; for it is not possible to procure many arguments bearing on them, since there are few intermediate steps between them and the principle, through which the things that come after must be shown. And of definitions, the hardest of all to argue about are those that employ such terms as are, in the first place, unclear as to whether they are said simply or in many ways, and, besides this, not even known whether they are used by the one who framed the definition in their proper sense or metaphorically.
13| For because they are unclear, such problems have no arguments to bring against them, and because it is unknown whether they are of this kind on account of being spoken metaphorically, they admit of no censure. In general, whenever a problem is hard to attack, we must suppose either that it needs a definition, or that it belongs to the things that are said in several ways or spoken metaphorically, or that it is not far from the principles, or that the reason is that it is not yet clear to us in which of the stated ways the difficulty it presents arises. For once the way is clear, it is plain that either a definition must be given, or a division made, or the intermediate premises must be supplied; for it is through these that the extreme terms are shown. In the case of many positions, too, when the definition has not been correctly given, it is not easy to argue and to attack them—for instance, whether one thing is contrary to one thing, or to more than one. But once the contraries have been properly defined, it is easy to establish agreement whether it is possible for more than one thing to be contrary to the same thing or not. And the same holds in the other cases that need a definition. It seems that even in mathematics some things are not easily proved for want of a definition—for instance, that the line drawn parallel to the side, cutting the plane figure, divides the line and the area alike.
14| But once the definition is stated, what is meant becomes immediately clear; for the areas and the lines undergo the same reciprocal subtraction, and this is the definition of ratios being the same. In general, when the definitions of the primary elements are laid down—for instance, what a line is and what a circle is—it is easiest to demonstrate things, except that not many arguments can be brought to bear on each of these, because there are not many intermediate steps; but if the definitions of the principles are not laid down, it is difficult, and perhaps altogether impossible. The same holds, similarly, in the case of arguments concerned with accounts. So we must not fail to notice, whenever a position is hard to attack, that it is affected by one of the things just mentioned. But when it is a greater task to argue against the axiom and the premise than against the position itself, one might be at a loss whether such things should be granted or not. For if the answerer will not grant it but demands that the questioner argue for this point too, he will be requiring more than what was laid down at the start; but if he does grant it, he will be believing on the strength of things less credible than what is at issue. If, then, the problem must not be made more difficult, it should be granted; but if the argument ought to proceed through things better known, it should not be granted. Or rather: to the learner it should not be granted, unless it is better known, but to the one in training it should be granted, provided only that it appears true.
15| So it is clear that one ought not to grant premises in the same way to a questioner as to a teacher. On how one ought to put questions and arrange them, then, what has been said is more or less sufficient. Concerning answering, first we must determine what the function of the good answerer is, just as we did for the good questioner. It is the function of the questioner to lead the argument in such a way as to make the answerer say the things that are least reputable among those necessitated by his position; and it is the function of the answerer to make it appear that the impossibility or the paradox does not come about through him, but through the position—for perhaps it is one mistake to have laid down at the start what one should not have laid down, and another to have failed to maintain properly what one did lay down.
16| Since these matters are left undetermined for those who carry on arguments for the sake of practice and trial—for the aims are not the same for teachers and learners as for competitors, nor for these as for those who spend time discussing together for the sake of inquiry. For the learner, one should always grant what seems to be the case; for no one even attempts to teach falsehood. But among those competing, the questioner must clearly be seen to be doing something at all costs, while the answerer must be seen to be suffering nothing. In dialectical meetings, however, among those who carry on arguments not for the sake of competition but for trial and inquiry, it has not yet been worked out at what the answerer should aim, and what sorts of things he should grant and what sorts not, with a view to maintaining his position well or badly. Since, then, we have nothing handed down to us by others, let us ourselves try to say something. Now the answerer must, of necessity, sustain his argument having laid down a position either reputable or unreputable or neither, and either reputable or unreputable simply, or in a qualified way—for instance, reputable to some particular person, either to himself or to another. It makes no difference in what way it is reputable or unreputable; for the manner of answering well will be the same, whether to grant or not to grant what is asked. If, then, the position is unreputable, the conclusion must necessarily be reputable, but if the position is reputable, the conclusion must be unreputable;
17| for the questioner always concludes the opposite of the position. But if what is laid down is neither unreputable nor reputable, the conclusion too will be of that same kind. Since the one who reasons well demonstrates the problem proposed from things more reputable and better known, it is clear that if what is laid down is simply unreputable, the answerer should grant neither what is simply not held to be true, nor what is held to be true but less so than the conclusion. For since the position is unreputable, the conclusion is reputable, so that everything taken as premises must be reputable, and more reputable than the point at issue, if the less known thing is to be established by way of the better known. So if any of the things asked is not of this kind, the answerer should not grant it. But if the position is simply reputable, it is clear that the conclusion is simply unreputable. He should therefore grant all the things that seem to be true, and, of those that do not, whatever are less unreputable than the conclusion; for in this way he would seem to have argued adequately. Similarly, if the position is neither unreputable nor reputable; for in this case too he must grant everything that appears to be so, and, of the things that do not seem so, whatever are more reputable than the conclusion; for in this way the arguments will turn out more reputable. If, then, what is laid down is simply reputable or unreputable, the comparison must be made against what seems to be true simply;
18| but if what is laid down is reputable or unreputable not simply but to the answerer, then he must judge what seems and does not seem to be true by reference to himself, and grant or not grant accordingly. But if the answerer is maintaining someone else's opinion, it is clear that he must grant or deny each point with an eye to that person's way of thinking. This is why those who introduce other people's opinions—for instance, that good and bad are the same, as Heraclitus says—do not grant that the contraries cannot be present in the same thing at the same time, not because this does not seem so to them, but because, on Heraclitus's view, it must be said in that way. Those who take up positions from one another do the same thing; for they aim at saying what the one who laid down the position would say.
19| It is clear, then, at what the answerer should aim, whether what is laid down is reputable simply or reputable to someone in particular. Since everything asked must be either reputable or unreputable or neither, and must be either relevant to the argument or not relevant to it, if it is held to be true and not relevant to the argument, it should be granted, with the answerer saying that it seems so; but if it is not held to be true and not relevant to the argument, it should be granted, but he should note that it is not held to be true, as a safeguard against appearing simple-minded. But if it is relevant to the argument and held to be true, he should say that it does seem so, but that it is very close to the point at issue from the start, and that if this is granted, the position laid down is destroyed. But if it is relevant to the argument, and the claim is very unreputable, he should say that it does follow if this is granted, but that what is being proposed is very simple-minded. But if it is neither unreputable nor reputable, then if it is not at all relevant to the argument, it should be granted without qualification; but if it is relevant to the argument, he should note that if it is granted, the point at issue from the start is destroyed. For in this way the answerer will not be thought to suffer anything through his own fault, if he foresees each point and lays it down accordingly, and the questioner will get his syllogism, since everything more reputable than the conclusion has been granted him. But those who attempt to reason from premises less reputable than the conclusion clearly do not reason well; hence such things should not be granted to questioners.
20| One must respond in the same way in the case of things said unclearly or in several ways. For since it is granted to the answerer, when he does not understand, to say "I do not understand," and, when a thing is said in several ways, he is not compelled either to agree or to deny it, it is clear, first, that if what has been said is not clear, one should not hesitate to say that one does not follow; for often, from having granted something asked unclearly, one runs into some difficulty. But if the thing is familiar yet said in several ways, then if what is said is true or false of all the senses alike, one should grant or deny it without qualification; but if it is false in one sense and true in another, one should point out that it is said in several ways, and that it is false in one sense but true in another. For if the distinction is drawn only later, it is unclear whether the ambiguity was seen from the start as well. And if one has not foreseen the ambiguity but, looking to one of the senses, has granted it, one should say, to the person who tries to lead one to the other sense, that it was not with a view to that sense that the concession was granted but to the other. For when several things fall under the same name or account, dispute arises easily. But if what is asked is both clear and simple, one should answer either yes or no.
21| Since every premise used in argument either is one of the things a syllogism is drawn from, or is asked for the sake of one of these — and it is clear when something is being taken for the sake of something else, by the fact that several similar things are being asked, since it is by induction, or by likeness, that people usually get the universal — one should grant all the particular cases, provided they are true and reputable, but should try to bring an objection against the universal; for to block the argument without an objection, whether real or apparent, is to be quarrelsome. If, then, when many things are apparent, someone does not grant the universal and has no objection, it is clear that he is being quarrelsome. Further, if he cannot even counter-argue that it is not true, he would seem to be being quarrelsome all the more. Yet even this is not sufficient; for we have many arguments contrary to what is believed, which are hard to resolve — as with Zeno's argument that motion is impossible, and that one cannot even traverse the stadium; but this is no reason not to grant the opposite of such conclusions. If, then, someone neither raises an objection nor is able to counter-argue, and yet does not grant the point, it is clear that he is being quarrelsome; for quarrelsomeness in arguments is answering in a way contrary to the ways stated, one that destroys the syllogism.
22| One who is about to defend a thesis and a definition should first examine it himself, on his own account; for it is clear that one must set oneself against the very points from which questioners overturn a position laid down. One should be wary of defending an unreputable premise. A premise may be unreputable in two ways: either because absurd things follow from stating it — for instance if someone were to say that everything is in motion, or that nothing is — or because it is such as would be chosen only by a worse character and is contrary to what people wish, for instance that pleasure is the good, or that doing wrong is better than suffering it. For people will hate someone not as defending it for the sake of argument but as saying what he actually believes.
23| As for arguments that infer something false, the resolution must consist in demolishing the point through which the falsity arises. For not just any demolition counts as a resolution, even if what is demolished is false. For the argument may contain several falsehoods — for instance, if someone assumes that the one who is sitting is writing, and that Socrates is sitting; for from these it follows that Socrates is writing. Demolishing the claim that Socrates is sitting does not resolve the argument any more, even though that premise is false; but it is not on account of this premise that the argument is false. For if someone happens to be sitting but not writing, the same resolution will no longer fit such a case. So this is not what must be demolished, but rather the claim that the one who is sitting is writing; for not everyone who is sitting writes. So whoever demolishes the point through which the falsity arises has resolved the argument in every case, but only the one who knows that the argument turns on this point knows the resolution, as with cases of false diagrams; for it is not enough to raise the objection, even if what is demolished is false, but one must also demonstrate why it is false; for only then would it be clear whether the objector has foreseen something or not. There are four ways of preventing an argument from reaching its conclusion. Either by demolishing the point through which the falsity arises, or by stating an objection against the questioner —
24| for often one has not really resolved the argument, but the questioner is nonetheless unable to carry it any further. A third way is directed against the questions asked; for it may happen that what the questioner wants does not follow from the questions asked, because they were badly framed, but that the conclusion would follow if something were added. If, then, the questioner can no longer advance, the objection would be directed against the questioner; but if he can, it should be directed against the questions asked. A fourth kind of objection, and the worst, is the one directed against the time available; for some people raise objections of a kind that would take longer to discuss than the present session allows. The objections, then, as we have said, arise in four ways; but of these, only the first is a resolution, while the rest are merely certain hindrances and obstructions to the conclusions.
25| Criticism of an argument in itself, and criticism of it as it is being put through questioning, are not the same thing; for often the answerer is himself responsible for the argument's not having been conducted well, by not conceding the points from which it was possible to argue well against the thesis; for it does not rest with one party alone to see the common task carried through well. It is sometimes necessary, then, to argue against the speaker and not against the thesis, when the answerer keeps opposing whatever the questioner says, out of sheer spite. People who do this, being quarrelsome, turn their discussions into contests rather than dialectical exchanges. Further, since arguments of this kind exist for the sake of practice and testing rather than teaching, it is clear that one should draw syllogisms not only from truths but also from falsehoods, and not always through true premises but sometimes through false ones as well. For often, when a true premise has been laid down, the person arguing dialectically is compelled to demolish it, so that false premises must be put forward instead. And sometimes, even when a false premise has been laid down, it must be demolished by means of falsehoods; for nothing prevents someone from believing things that are not so more readily than things that are true, so that if the argument proceeds from what he believes, he will be more convinced than benefited. But one who shifts ground well must do so dialectically and not eristically, just as the geometer does so in a geometrical manner, whether the conclusion reached is false or true —
26| what kinds of syllogisms are dialectical has been stated earlier. Since a partner who obstructs the common task is a poor partner, this clearly holds in argument as well. For here too there is some common goal set before the parties, except in the case of contestants; for these two cannot both attain the same end, since it is impossible for more than one to win. And it makes no difference whether someone does this through answering or through questioning; for the one who asks questions in an eristic manner argues badly, and so does the one who, in answering, will not grant what is apparent nor accept whatever it is the questioner wants to establish. It is clear, then, from what has been said, that one must not criticize the argument in itself and the questioner in the same way; for nothing prevents the argument from being poor while the questioner has conducted the discussion against the answerer as well as could be done. For against quarrelsome people it is perhaps not possible to construct the syllogisms one wants straightaway, but only such as are possible. And since it is undefined when people are taking the contrary of what was granted and when they are taking the point at issue itself (for often, speaking on their own, they say contrary things, and having first denied something, they later grant it — which is why, when questioned, they often assent both to the contrary and to the point at issue), the arguments are bound to turn out poor.
27| The answerer is responsible—by not granting some things, and by granting others that are of that sort. It is clear, then, that questioners and arguments should not be criticized in the same way. In its own right, there are five criticisms that apply to an argument. The first is when nothing is concluded from the questions asked—neither the thesis originally proposed nor anything else at all—though none of the premises on which the conclusion depends are false or lacking reputation (either all of them or most), and the conclusion does not come about even if certain premises are taken away, or added, or some taken away and others added. The second is if the syllogism does not bear on the thesis, even though it is constructed from such premises and in the way stated before. The third is if a syllogism comes about only when certain things are added, and these are worse than what was asked and less reputable than the conclusion. Again, the fourth is if it comes about only when certain things are taken away; for sometimes people take more premises than are necessary, so that the syllogism does not come about because of these particular premises being present. Further, the fifth is if it proceeds from premises less reputable and less credible than the conclusion, or from premises that are true but require more work to demonstrate than the problem itself. One should not demand that the syllogisms for every problem be equally reputable and persuasive; for by nature some of the things under investigation are straightaway easier, others harder, so that if someone draws his conclusion from the most reputable premises available, he has argued well.
28| It is clear, then, that even in the case of an argument, the same criticism does not apply both in relation to the problem proposed and in the argument's own right. For nothing prevents an argument from being blameworthy in its own right but praiseworthy in relation to the problem, and again, conversely, praiseworthy in its own right but blameworthy in relation to the problem, whenever it is easy to draw a conclusion from many true and reputable premises. Indeed, an argument that has actually reached a conclusion could sometimes be worse than one that has not—when the former is concluded from silly premises though the problem itself is not of that sort, while the latter still needs premises that are reputable and true, and the deficiency does not lie in the argument itself but in what remains to be assumed. It is not fair to criticize arguments that reach a true conclusion through false premises; for a false conclusion must always be reached through a false premise, but a true conclusion can sometimes also be reached through false premises. This is clear from the Analytics. Whenever the argument stated is a demonstration of something, then if some other point has no bearing at all on the conclusion, there will be no syllogism from that point; but if there appears to be one, it will be a sophism, not a demonstration. A philosopheme is a demonstrative syllogism, an epicheireme is a dialectical syllogism, a sophism is a contentious syllogism, and an aporema is a dialectical syllogism concerning a contradiction. If something is proved from two premises that are both reputable but not equally so, nothing prevents the thing proved from being more reputable than either premise.
29| But if one premise is reputable and the other neither reputable nor disreputable, or if one is reputable and the other not reputable, then, if they are equally matched, the conclusion will likewise be equally matched; but if one is more so, the conclusion will follow the stronger one. There is also this further error concerning syllogisms: when one proves a point through more steps than necessary, though it was possible to prove it through fewer, using premises already present in the argument. For example, that one opinion is more truly an opinion than another: suppose someone claims that the thing-itself-in-each-case exists most fully, and that it is truly an object of opinion, so that it is more fully an object of opinion than particular things; and that what corresponds to the more is itself more; and that there is also a true opinion-itself, which will be more exact than particular opinions; but it has already been claimed both that opinion-itself is true and that the thing-itself-in-each-case exists most fully; so that this opinion, being most true, is more exact. What, then, is the fault here? It is that it causes the actual reason the argument works to go unnoticed.
30| An argument is clear in one way, the most obvious way, if it is so concluded that nothing further need be asked; and in another way, the one most properly so called, when the premises taken are such that the conclusion necessarily follows from them, and they are drawn together through intermediate conclusions; further, an argument is unclear if it seriously falls short of reputable premises. An argument is called false in four ways. In one way, when it appears to reach a conclusion without actually doing so, which is called a contentious syllogism. In another way, when it does reach a conclusion, but not one bearing on the point at issue—which happens especially with arguments leading to an impossibility—or when it reaches a conclusion bearing on the point at issue, but not by the method proper to it. This last happens when someone who is not a doctor seems to argue as a doctor would, or someone who is not a geometer seems to argue as a geometer would, or someone who is not skilled in dialectic seems to argue as a dialectician would, whether the resulting conclusion is false or true. In yet another way, when the conclusion is reached through false premises. Of this kind, the conclusion will sometimes be false, sometimes true; for a false conclusion is always reached through false premises, but a true conclusion may come about even when not drawn from true premises, as has been said before. Now, an argument's being false is a fault of the speaker rather than of the argument, and not even always of the speaker, but only when it escapes his own notice; since, in its own right, we accept such an argument more readily than many true ones, if it does away with something true by proceeding from the most reputable premises possible.
31| For, being of this sort, it is a demonstration of other truths; for it must be that one of the premises laid down does not hold at all, and so there will be a demonstration of this. But if a true conclusion is reached through premises that are false and excessively silly, it would be worse than many arguments that conclude falsely; and such an argument might in fact also conclude falsely. So it is clear that the first thing to examine about an argument in its own right is whether it reaches a conclusion at all; the second, whether the conclusion is true or false; the third, from what sort of premises it proceeds. For if it comes from premises that are false but reputable, it is a dialectical argument; if from premises that are true but not reputable, it is a poor one. And if the premises are both false and utterly disreputable, it is clearly a poor argument, either without qualification or with respect to the matter at hand.
32| How the questioner begs the original point, and how he begs contraries, has been discussed with respect to truth in the Analytics; it must now be discussed with respect to reputation. People appear to beg the original point in five ways. Most obviously, and first, if someone posits the very thing that needs to be proved. This is not easy to do unnoticed when the very same thing is stated outright, but it is more likely with synonyms, and in cases where the name and the definition signify the same thing. Second, when, though it is a particular case that needs to be proved, someone posits the universal—for example, in attempting to show that there is one science of contraries, he claims outright that there is one science of opposites in general; for what needed to be proved on its own is thereby seen to be claimed together with a number of other things. Third, if, though the task set is to prove the universal, someone posits the particular case—for example, if, though the task is to prove it of all contraries, he claims it merely of certain particular ones; for this person too is seen to claim, on its own and separately, what needed to be proved together with a number of others. Again, if someone divides up the problem proposed and then claims it—for example, if, though it is necessary to show that medicine is concerned with both the healthy and the diseased, he claims each of these separately. Or if, of two things that necessarily follow from each other, someone claims one of them—for example, claiming that the side is incommensurable with the diagonal, when what needs to be proved is that the diagonal is incommensurable with the side.
33| Contraries are begged in an equal number of ways as the original point. First, if someone claims both an opposed affirmation and denial. Second, if he claims contraries according to their opposition—for example, that good and bad are the same. Third, if, having claimed the universal, someone then claims the contradictory of it in a particular case—for example, if, having taken it that there is one science of contraries, he then claims that there is a different science of the healthy and the diseased; or, having claimed this particular case, he tries to take the opposite of it as the universal. Again, if someone claims the contrary of what necessarily follows from the premises laid down, and likewise if, without taking the opposites themselves, he claims two such things from which the opposite contradiction will follow. Taking contraries differs from begging the original point in this: in the latter case, the fault lies with respect to the conclusion (for it is with an eye to the conclusion that we speak of begging the original point), whereas contraries lie in the premises, in virtue of how these stand in relation to one another.
34| For exercise and practice in arguments of this kind, the first thing to make a habit of is converting arguments. This will make us more resourceful with regard to what is said, and will let us come to know many arguments from a few. For converting an argument means taking the conclusion together with the remaining premises and doing away with one of the premises granted; for it is necessary, if the conclusion is not so, that one of the premises be done away with, given that when all were laid down the conclusion had to hold. And for every thesis one must examine the argument both for its being so and for its not being so, and one must examine the argument ... and when one has found it, one must at once seek the solution; for in this way we shall at the same time have become practiced both in questioning and in answering. And if we have no one else to argue with, we must do it with ourselves. And one should set arguments side by side, picking out arguments bearing on the opposition; for this creates great resourcefulness for forcing an opponent's hand, and is a great help toward refutation, whenever one is well supplied both for the position that it is so and that it is not so; for then the defense comes to be made against opposites. And for knowledge and for the practical wisdom that comes with philosophy, the ability to see together, and to have seen together, the consequences of each of the two hypotheses is no small instrument; for what remains is to choose the right one of the two correctly.
35| For this one must have a natural gift; and this is what natural aptitude with regard to truth is, the ability to choose well what is true and avoid what is false; which is what those who are well endowed by nature can do; for by loving and hating well what is presented to them, they judge well what is best. And one must know by heart the arguments that bear on the problems that come up most often, and above all those concerning the first theses; for it is with these that answerers most often lose heart. Further, one must be well supplied with definitions, and have those in current repute and the primary ones ready to hand; for it is through these that syllogisms come about. And one must try to get a grip on the topics into which other arguments most often fall. For just as in geometry it is useful to have been trained about the elements, and in arithmetic to have ready to hand the multiplication tables, which makes a great difference for also knowing the rest of number when it is multiplied out, likewise also in arguments it is useful to have the principles and the premises ready, to know them by heart; for just as in the art of memory the mere placing of the topics at once makes people remember the things themselves, so too this will make one more capable of syllogizing, because one looks to premises that are fixed and countable. And one should set down in memory a common premise rather than an argument;
36| for it is hard to be adequately supplied with a starting point and a hypothesis. Further, one must make a habit of turning one argument into many, disguising it as unrecognizably as possible. This would be achieved if one removed it as far as possible from its family relation to the subject the argument is about. Those arguments will be capable of undergoing this most of all which are most universal, for instance that there is not a single knowledge of several things; for this holds both in the case of relatives and in the case of contraries and of things in the same coordinate series. And one must also make one's records of arguments universal, even if the argument was discussed in a particular case; for in this way it will be possible to turn the one into many. Likewise also in rhetoric, in the case of enthymemes. And one should oneself, as far as possible, avoid carrying syllogisms toward the universal. And one must always examine whether arguments are discussing common matters; for all who discuss particulars are also discussing universals, since the demonstration of the universal is present in the particular, because it is not possible to draw a syllogism about anything without the universal. Exercise in inductive arguments should be assigned to the young, exercise in syllogistic ones to the experienced. And one must try to get premises from those trained in syllogistic argument, and comparisons from those trained in induction;
37| for it is in these that each of the two groups is practiced. And in general, from practicing in discussion one must try to carry away either a syllogism about something, or a solution, or a premise, or whether a given question was asked correctly or not, whether by oneself or by someone else, and in virtue of what each was so. For it is from these that capacity comes, and the practicing is for the sake of capacity, above all concerning premises and objections; for, to put it simply, the dialectician is the one skilled at putting forward premises and at raising objections. Now to put forward a premise is to make the many into one (for what the argument is directed at must be grasped as a whole), while to raise an objection is to make the one into many; for it either divides or does away with it, granting one part of what is put forward and not the other. One should not engage in discussion with everyone, nor practice with just anyone one meets. For with some people the arguments are bound to turn out worthless; for against someone who tries by every means to appear to escape, it is right to try by every means to draw the syllogism, but it is not a graceful thing to do. That is why one should not readily enter into argument with chance people; for base talk is bound to result; and indeed even those who are in training find it impossible to keep from arguing in a contentious way. And one must also have arguments already prepared for problems of the following sort, in which, by being well supplied with very few, we shall have arguments useful for very many.
38| These are the universal ones, and it is against these that it is harder to provide resources out of what lies ready to hand.