Doubtless there are rules that require to be observed in the dialectical attack and defence, as there are rules for all other matches such as chess or fencing. I should have been glad if Aristotle had given a precise and tenable explanation what these rules were. He describes the Sophist as one who plays the game unfairly; but we have already seen that the ends pursued by the Dialectician generally are hardly at all distinguishable from those aimed at by the Sophist. If we look to the account of the means employed by one and the other, we shall in like manner fail to see how any real line can be drawn between them.

Thus, one proceeding declared to be characteristic of the Sophist is — that he puts multiplied questions apparently at random, without any visible bearing on the thesis; practising a sort of fishing examination, in order to obtain some answer of which he may take advantage.67 But, when we turn to the Eighth Book of the Topica, we find Aristotle expressly recommending the like manœuvre to the Dialectician; advising him to conceal as much as possible the scheme and intended series of his questions — to begin as far as possible apart from the thesis, to put the questions in a succession designedly incoherent and unintelligible, and to obtain (what, if obtained, ensured complete success) the full extent of premisses necessary for his final refutative syllogism, without the respondent being aware that he had conceded them.68 The questioner is farther advised to throw the respondent off his guard by affecting indifference whether each question is answered affirmatively or negatively, and by occasionally taking objection against himself, in order that he may create the impression of a strictly honest purpose.69 If we compare the interrogative procedure which Aristotle recommends to the Dialectician with that which he blames in the Sophist, we shall find that the former is even a greater refinement of deception than the latter.

The next trick which we find ascribed to the Sophist is — that he conducts the train of interrogation in such manner as to bring it upon a ground on which his memory is abundantly furnished with topics. Aristotle adds that this may be done well and honourably, or ill and dishonourably.70 From his own admission we see that this practice was not peculiar to Sophists, but was common also to those whom he calls Dialecticians: like every other part of the procedure, it might be done well or ill; but wherein this difference consisted he does not further explain. Indeed, when we recollect that the elaborate details and classification of the Topica are mainly intended to furnish the memory with an abundant store of premisses well-arranged and ready for interrogation,71 we may be sure that every Dialectician who had gone through the trouble of learning them would be impatient to apply them; and would make an opportunity for doing so, if none were spontaneously tendered to him. But, if the answers obtained were totally irrelevant to his final purpose of refuting the thesis, they would be nothing but embarrassment to him.72 We must, therefore, understand that the questions put would be such as tended ultimately to introduce that refutative Syllogism which the questioner was bound to conclude with. If they were not, he was of course punished by failure.

A third manœuvre treated as peculiar to the Sophist is — that he takes account of the particular philosophical sect to which the respondent belongs, and endeavours to bring out by interrogations whatever there may be paradoxical in the tenets of that sect.73 But would not any expert Dialectician do just the same? What else would be done by Sokrates, if cross-examining an Anaxagorean or a Herakleitean? or by Aristotle himself, if interrogating a Platonist?

Another proceeding treated as peculiar to the Sophist is — that he seeks to drive the respondent into a paradox, by bringing out in cross-examination certain well-known antitheses or contradictions which subsist together in the opinions of mankind. Thus, men profess in their public talk high principles of virtue; but secretly and at the bottom of their hearts they desire to get wealth or power per fas et nefas. Again, there are two kinds of justice: one, that which is just by nature and in truth, such as wise men or philosophers approve; the other, that which is just according to law or custom, such as the multitude in this or in in some other society approve. There is, also, conflict between the authority of a father, and that of the wise; between justice and expediency; and as to whether it is more eligible to suffer wrong or to do wrong.74 All these antitheses are presented to us in the Platonic Gorgias, to which (i.e., to the speech of Kallikles therein) Aristotle here makes reference; and he numbers it among the vices distinguishing the Sophist from the genuine Dialectician — to dwell upon such antitheses for the purpose of forcing the respondent into paradoxical answers. But, surely, the antitheses here fastened upon that obnoxious name are of a class utterly opposed to the class of pseudo-probabilia, which he tells us are the peculiar game of the litigious Sophist, though every man of ordinary intelligence detects them at first sight as fallacies. They are all real and serious issues,75 having plausible arguments pro and con, debateable without end, and settled by every man for himself according to his own sentiment and predisposition. They are exactly the subject-matter best fitted for the acute Dialectician. No man would be allowed by Aristotle to deserve that title, if he omitted to raise and argue them, the thesis being supposed suitable.76 Aristotle himself speaks often of the equivocal sense of the term justice — of the distinction between what is just by nature and what is just according to some local or peculiar sentiment.77 The manœuvre which Aristotle imputes to the Sophist being exactly the same as that which Kallikles imputes to Sokrates in the Platonic Gorgias,78 it is Sokrates, and not Kallikles, who serves here as illustrating what Aristotle calls a Sophist. Indeed, if we read the Gorgias, we shall find the Platonic Sokrates there represented as neglecting the difference between what is probable (conformable to received opinion) and what is paradoxical. He admits that he stands alone in his opinion, against all the world, and his opponents even imagine that he is bantering them; but he confides in his own individual reason and consistency, so as to be able to reduce all opponents dialectically to proved contradiction with themselves.79 Himself maintaining a paradox, he constrains his respondent by acute dialectic to assent to it; which is exactly what Aristotle imputes to the Sophists of his day as a reproach.

Some predecessors of Aristotle had distinguished arguments or discourses into two separate classes — those addressed to the name, and those addressed to the thought.80 This distinction Aristotle disapproves, denying certainly its pertinence and almost its reality. There can be no arguments addressed to the thought only, apart from the name: all of them must be addressed to the name, and through it to the thought.81 Whether an argument is addressed to the thought or not, depends not upon any thing in the argument itself, but upon the meaning which one respondent or other may happen to attach to the words: if the respondent understands it as the questioner intended, it is addressed to the thought; if not, not.82 To require that the questioner shall distinguish accurately the sense in which he puts the question, would, according to Aristotle, convert him into a teacher — would confound the line between Dialectic and Didactic.83 And this may be granted; but not less, if Dialecticians are to refrain from all those proceedings which Aristotle notes and condemns as peculiar to the Sophist, must they be held to pass into the attitude of teacher and learner; the questioner doing what he can, not to embarrass but, to enlighten and assist the respondent. The purpose of victory, and the stimulus of competition in the double function of question and answer (while entirely absent from Didactic), are quite as essential to the Dialectician as to the Sophist. That the Sophist seeks victory unscrupulously and at all cost, while the Dialectician respects certain rules and limits of the procedure — is a difference well deserving to be noticed; yet not a differentia giving name and essence to a new species. The unfair Dialectician is a Dialectician still; all his purposes remain the same, though the means whereby he pursues them are altered. This distinction of means between the two, Aristotle has taken very insufficient pains to point out. Rude and provocative manner, either on the part of questioner or respondent, and impudent assumption of concessions which have neither been asked nor granted, — these are justly enumerated as illustrations of unfair Dialectic.84 But the enumeration is most incompletely performed; because Aristotle, in his anxiety to erect Sophistic into an art or procedure by itself, distinct from and alongside of Dialectic, has transferred to it much that belongs to fair and and admissible Dialectic. Hence the really unfair and objectionable means are not often brought into the foreground.

Though Aristotle speaks so contemptuously about Sophistic, he nevertheless indicates Loci (or general heads of subjects) to assist the sophistical questioner in attacking, and precepts to the sophistical respondent for warding off attack. On the whole, these precepts are not materially different from those laid out in the Topica for Dialectic; except that he gives greater prominence to Solecism and Tautology, as thrusts practised by the sophistical questioner. He insists upon the intellectual usefulness of practice in sophistical debate, hardly less than in what he calls dialectical, and, as was remarked, upon similar grounds.85 He recommends it as valuable not only for imparting readiness and abundance in argument, but also for solitary meditation and for investigation of scientific truths. Without it (he declares) we cannot become familiar with the equivocations of terms and propositions, nor acquire the means of escaping them. If we allow ourselves to be entangled in them, without being aware of it, by others, we shall also be entangled in them when we pursue reflections of our own.86 It is not enough to see generally that there is a fallacy; we must farther learn to detect at once the precise seat of the fallacy, and to point out rapidly how it may be cleared up. This is the more difficult to do, because fallacies that we are thoroughly aware of will often escape our notice under inversion and substitution of words.87 Unless we acquire promptitude by frequent exercise in such debates, we shall find ourselves always unprepared and behind-hand in each particular case of confusion. If we complain and condemn such debates generally, we shall appear to do so upon no better grounds than our own stupidity and incompetence.88

Accordingly the Sophistici Elenchi contains precepts, at considerable length,89 to the respondent in a sophistical debate, how reply or solution is to be given to the fallacies involved in the questions; all the thirteen Fallacies, (the six In Dictione, and the seven Extra Dictionem) being treated in succession. In conducting his defensive procedure, the respondent must keep constantly in mind what the Sophistical Refutation really is. He must treat it not as a real or genuine refutation, but as a mere simulation of such; and he must so arrange his reply as to bring into full evidence this fact of simulation. What he has to guard against is, not the being really refuted but, the seeming to be refuted.90 The refutative syllogism constructed by the sophistical questioner, including as it does Equivocation, Amphiboly, or some other verbal fallacy, and therefore yielding no valid conclusion, does not settle whether the respondent is really refuted or not. If indeed the questioner, in putting his interrogation, discriminates the double meaning of his words, where they have a double meaning, the respondent ought to answer plainly and briefly Yes, or No; either affirming or denying what is tendered. But, if the questioner does not so discriminate, the respondent cannot reply simply Yes, or No: he must himself discriminate the two meanings, and affirm or deny accordingly.91 Unless he guards himself by such discrimination, he cannot avoid falling into a contradiction, at least in appearance. The equivocal wording of the question will be tantamount to the fallacy of putting two questions as one.92

As the questioner may propound as refutation what seems to be such but is not so in reality, so the respondent may meet it by what is an apparent solution but no solution in reality, There occur various cases, in sophistic or agonistic debate, wherein a simulated solution of this kind is even preferable to a real one.93 If the question is plausible, the respondent may answer, “Be it so”; but, if it involves any paradox in answering, he will answer by saying, “So it would appear”: he will thus not be supposed to have granted what amounts to refutation or paradox.94 Where the question put is such that, while involving falsehood or paradox if answered in the affirmative, it is at the same time closely or immediately connected with the thesis set up, — the respondent may treat it as equivalent to a Petitio Principii, and make answer in the negative. Also, where the questioner, trying to establish an universal proposition by Induction, puts the final question, not under an universal term but, as the general result of the particulars conceded (and such like), — the respondent may refuse to admit this last step, and may say that his antecedent concessions have been misunderstood.95

If a question is put in plain and appropriate language, answer must be made plainly or with some clear distinction; but, where the question is put obscurely and elliptically, leaving part of the meaning unexpressed, the respondent must not concede it unreservedly. If he does, fallacious refutation may very possibly be the result:96 he may appear to be refuted by that which is no real refutation. If, of two propositions, the second follows upon the first, but the first does not follow upon the second, the respondent, where he has the choice, ought to grant the second only, and not the first. He ought not to make a greater concession when he can escape with a less;97 e.g., he ought to concede the particular rather than the universal.

Again, among opinions generally received, there are some which the public recognize as matters of more or less doubt and uncertainty; others, on which they are firmly assured that every one who contradicts them speaks falsely. When it is uncertain to which of these two classes the question put is referable, the respondent will be safer in answering neither affirmatively nor negatively, but simply, “I go with the received opinions.”98 In cases where opinions are divided, he may find opportunity for changing the terms, and for substituting a metaphorical equivalent as what he concedes. Such change of terms may pass without protest, in consequence of the doubtful character of the matter; while it will embarrass the questioner in constructing his refutation.99 The respondent may further embarrass him by anticipating questions that seem likely to be put, and by objecting against them beforehand.100

When the questioner has obtained the premisses which he thinks necessary, and has drawn from them a refutative syllogism, the respondent must see whether he can properly solve that syllogism or not.101 A good and proper solution is, to point out on which premiss the fallacy of the conclusion depends. First, he must examine whether it is formally correct, or whether it has only a false appearance of being so: if the last be the case, he must distinguish in which of the premisses and in what way such false appearance has arisen. If on the other hand the syllogism is formally correct, he must look whether the conclusion is true or false. Should it be true, he cannot solve the syllogism except by controverting one or both of the premisses; but should the conclusion be false, two modes of solution are open to him. One mode is, if he can point out an equivocation or amphiboly in the terms of the conclusion; another mode will be, to controvert, or exhibit a fallacy in, one of the premisses.102 The respondent, however, must learn to apply this examination rapidly and unhesitatingly: to do so at once is very difficult, though it may be easily done if he has leisure to reflect.103

Aristotle then proceeds to indicate the modes in which the respondent may provide solutions for each of the thirteen heads of fallacious refutation above enumerated. For these thirteen classes, he pronounces that one and the same solution will be found applicable to all fallacies contained in one and the same class.104

Thus, in the two first of them — Equivocation of Terms and Amphiboly of Propositions — duplicity of meaning must be either in the conclusion, or in the premisses, of the refutative syllogism. If it be in the conclusion, the refutation must at once be rejected, unless the respondent has previously admitted some proposition containing the equivocal word as one of its terms, so that the refutation may appear to contradict it expressly and distinctly. But, if it be in the premisses, then there is no necessity that the respondent should have previously admitted such a proposition; for the equivocal word may form the middle term of the refutative syllogism, and may thus not appear in the conclusion thereof.105 The proper way for the respondent to deal with these questions, involving equivocation or amphiboly, is to answer them, at the outset, with a reserve for the double meaning, thus: “In one sense, it is so; in another sense, it is not.” If he does not perceive the double meaning until he has already answered the first question, he must recover himself, when he answers the second, by pointing out the equivocation more distinctly, and by specifying how much he is prepared to concede.106 Even if he has been taken unawares, and has not perceived the equivocation until the refutative syllogism has been constructed simply and absolutely, he should still contend that he never meant to concede what has been apparently refuted, and that the refutation tells only against the name, not against the thing meant;107 so that there is no genuine refutation at all.