eBook of George Grote, Aristotle

In the next two Fallacies — those of Composition and Division, or Conjunction and Disjunction — when the questioner draws up his refutative syllogism as if one of the two had been conceded, the respondent will retort by saying that his concession was intended only in the other construction of the words. This fallacy is distinct from Equivocation; and it is a mistake to try (as some have tried) to reduce all fallacies to Equivocation or Amphiboly.108 The respondent will distinguish, in each particular case, that construction of the words which he intended in his admission, from that which the questioner assumes in his pretended refutation.109

108 Soph. El. xx. p. 177, a. 33-b. 9. οὐ πάντες οἱ ἔλεγχοι παρὰ τὸ διττόν, καθάπερ τινές φασιν.

This is another of the evidences showing that there were theorists prior to Aristotle on logical proof; and that his declaration of originality (in the concluding chapter of Sophist. Elenchi) must be taken with reserve.

109 Soph. El. xx. p. 177, b. 10-26: διαιρετέον οὖν τῷ ἀποκρινομένῳ· &c.

The Fallacies of Accent rarely furnish sophistical refutations,110 but those of Figura Dictionis furnish a great many. When two words have the like form and structure, it may naturally be imagined that the signification of one belongs to the same Category as that of the other. But this is often an illusion; and in such cases a sophistical refutation may be founded thereupon. The respondent will solve it by denying the inference from similarity of form to similarity of meaning, and by distinguishing accurately to which among the ten Categories the meaning of each several word or each proposition belongs. When two words thus seem, by their form, to belong to the same Category, the questioner will often take it for granted, without expressly asking, that they do belong to the same, and will found a confutation thereupon; but the respondent must not admit the confutation to be valid, unless this question has been explicitly put to him and conceded.111 A question is put which, in its direct and obvious meaning, bears only on the category of Quantity, of Quality, of Relation, of Action, or of Passion; but the respondent, not aware of the equivocation, answers it in such a manner as to comprehend the Category of Substance, and is so understood by the questioner when he constructs his refutative syllogism. The respondent will secure himself from being thus confuted, by keeping constantly in view to which of the Categories his answer is intended to refer.112

110 Ibid. xxi. p. 177, b. 35.

111 Ibid. xxii. p. 178, a. 4-28. τὸ γὰρ λοιπὸν αὐτὸς προστίθησιν ὁ ἀκούων ὡς ὁμοίως λεγόμενον· τὸ δὲ λέγεται μὲν οὐχ ὁμοίως, φαίνεται δὲ διὰ τὴν λέξιν.

112 Several illustrative examples of this mode of sophistical refutation, founded on the Fallacy called Figura Dictionis, are indicated in this chapter by Aristotle. The indication however, is often so brief and elliptical, that there is great difficulty in restoring the fallacies in full, and still greater difficulty in translating them into any modern language.

1. Is it possible at the same time to do and to have done the same thing? — No. To see something is to do something; to have seen something is to have done something? — Yes. Is it possible at the same time to see and to have seen the same thing? — Yes.

The respondent has thus contradicted himself. The form of the word ὁρᾶν appears to rank it under the Category ποιεῖν. However, I think that the mistake really made here was, that the respondent returned an answer universally negative to the first question.

2. Does anything coming under the Category Pati come under the Category Agere? — No. But τέμνεται, καίεται, αἰσθάνεται, all show by their form that they belong to the Category Pati? — Yes. Again, λέγειν, τρέχειν, ὁρᾶν, show by their form that they belong to the Category Agere? — Yes. You will admit, however, that τὸ ὁρᾶν is αἰσθάνεσθαί τι? — Certainly. Therefore something that belongs to the Category Agere belongs also to that of Pati.

If we turn back to Aristot. Categ. viii. p. 11, a. 37, we shall find that he admits the possibility that the same subject may belong to two distinct Categories.

3. Did any one write that which stands here written? — Yes. It stands here written that you are standing up — a false statement; but when it was written the statement was true? — Yes. Therefore the writer has written a statement both true and false? — Yes.

Here true and false belong to the Category Quality; the statement or matter written belongs to that of Substance. What the writer wrote had nothing to do with the former of the two Categories; and no contradiction has been made out by admitting that the statement was once true and is now false.

4. Does a man tread that which he walks? — Yes. But he walks the whole day? — Yes. Therefore he treads the whole day.

Here the Category of Quando is confused with that of Substance.

5. But the most interesting illustration of this confusion of one Category with another, is furnished by Aristotle in respect of the difference between himself and Plato as to Ideas or Universals. According to Plato the universal term denoted a separate something apart from the particulars, yet of which each of these particulars partook. According to Aristotle it denoted nothing separate from the particulars, but something belonging (essentially or non-essentially) to all and each of the particulars. In the Platonic theory it was an Hoc Aliquid (τόδε τι), or had an existence substantive and separate: in the Aristotelian it was a Quale or Quale Quid (ποιόν), having an existence merely adjective or predicative. Aristotle maintains that Plato or the Platonists placed it in the wrong Category — in the Category of Substance instead of in that of Quality.

Now it is by rectifying this confusion of Categories that Aristotle solves two argumentative puzzles which he ranks as sophistical:— (1) The argument concluding in what was called the ‘Third Man;’ (2) The following question: Koriskus, and the musical Koriskus — are these the same, or is the second different from the first?

What is called the ‘Third Man’ was a refutation of the Platonic theory of Ideas. Because Plato recognized a substantive existence, corresponding to each common denomination connoting likeness, apart from all the similar particulars denominated, e.g., a Self-man, or separate self-existent man, corresponding to the Idea, and apart from all individual men, Caius, &c. — opponents argued against him, saying:— If this is recognized, you must also recognize that the Self-man, and the individual man called Caius, have also a common denomination and similarity, which (upon your principles) corresponds to another Ideal Man, or a Third Man. You must, therefore, go on inferring upwards to a Fourth Man, a Fifth Man, &c., and so onwards to an indefinite number of Ideal Men, one above the other. This was intended as a refutation, by Reductio ad Impossibile, of the Platonic view of Ideas as separate Entities, each of them One and Universal. But Aristotle here treats it as a Sophistical Refutation; and he indicates what he calls the solution of it by saying that it confounds the Categories of Substance and Quality, putting the Universal (which ought to be under the Category of Quality) under the Category of Substance. He has no right, however, to include this among Sophistical Refutations, which are (as he himself defines them) not real but fallacious refutations, invented by a dishonest money-getting profession called Sophists, and which are solved by pointing out the precise seat of the fallacy. The refutation called the ‘Third Man’ is so far from being fallacious, that it is valid, and is recited as such elsewhere by Aristotle himself (Metaphs. A. ix. p. 990, b. 17); while the solution tendered by Aristotle, instead of being a solution, is a confirmation, pointing out, not where the fallacy of the refutation resides but, where the fallacy of the doctrine refuted resides. Moreover, if we are to treat the refutation called the ‘Third Man’ as sophistical, we must number Plato himself among the dishonest class called Sophists. Here is one among the many proofs that the strong line drawn by Aristotle between the Dialectician and the Sophist is quite untenable. The argument is distinctly enunciated in the Platonic Parmenides (pp. 131-133).

The meaning of the Universal (Aristotle maintains) must be considered as predicative only, tacked on to some Hoc Aliquid, and belonging to Quale or some other of the nine latter Categories. It may be set out as a distinct subject for logical consideration and reasoning: but it cannot be set out as a distinct existence beyond and apart from its particulars (παρὰ τοὺς πολλοὺς ἕν τι). It is ποιόν, and it cannot even be recognized as ὅπερ ποιόν or αὐτο-ποιόν, for this would put it apart from all the other ποιά, and would be open to the refutation above noticed called the ‘Third Man.’ Such is the drift of the very difficult passage of the Sophistici Elenchi (xxii. p. 178, b. 37-p. 179, a. 10). I differ from Mr. Poste’s translation (p. 71) of part of this passage, and still more from the explanation given in the latter part of his note (p. 155). I think that the doctrine of τὸ ἓν παρὰ τὰ πολλά is produced by Aristotle here and elsewhere in his work as untrue and inadmissible, not as his own doctrine. Mr. Poste understands this passage differently from the previous translators, with whom I agree for the most part, though M. Barthélemy St. Hilaire appears to me to have missed the hinge upon which Aristotle’s argument turns, by translating ὅπερ ποιόν — id ipsum, quod quale est (J. Pacius) — “une qualité:” the argument turns upon the distinction between ὅπερ ποιόν and ποιόν.

I come now to the second sophistical refutation given by Aristotle: Koriskus, and the musician Koriskus — are the two the same or different? This is what Aristotle calls a sophistical or fallacious argument (compare Metaphys. E. ii. p. 1026, b. 15); but it can hardly be so called with propriety, for the only solution that Aristotle himself gives of it is, that the two are idem numero, but in an improper or secondary sense (Topic. I. vii. p. 103, a. 30); i. e., that they are in one point of view the same, in another point of view different — they are ἓν κατὰ συμβεβηκός. See Arist. Metaph. Δ. vi. p. 1015, b. 16; Scholia, p. 696, a. 22, seq.; and Alexand. Aphrodis. ad Metaph. pp. 321, 322, 414, 415, ed. Bonitz. I understand Aristotle to say that Κόρισκος μουσικός cannot be properly set out or abstracted (οὐκ ἔστιν αὐτὸ ἐκθέσθαι), because it includes two Categories (Substance and Quality) in one; wherefore it cannot be properly compared either with Κόρισκος simply (Category of Substance) or with μουσικός simply (Category of Quality). It seems strange that Aristotle does not notice this argumentative difficulty in the discussion which he bestows on ταὐτόν in the Seventh Book of the Topica. The subtle reasonings, very hard to follow, which Aristotle employs (Physic. V. iv. p. 227) might have made him cautious in treating the difficulties of opponents as so many dishonest cavils. It is curious that Alexander, in reciting the sophistical argument, assumes as a matter of course that ὁ γραμματικὸς Σωκράτης is ὁ αὐτὸς τῷ Σωκράτει (Schol. ad Metaphys. p. 736, b. 26, Brand.).

As a general rule, in all the refutations founded on the seven Fallacies In Dictione, the respondent will solve the refutation by distinguishing the double meaning of the words or of the phrase, and by adopting as his own the one opposite to that which the questioner proceeds upon. If the Fallacy is of Conjunction and Disjunction, and if the questioner assumes Conjunction, the respondent will adopt Disjunction; if it be a Fallacy of Accent, and if the questioner assumes the grave accent, the respondent will adopt the acute.113

113 Soph. El. xxiii. p. 179, a. 11-25.

Passing to the Fallacies Extra Dictionem, where the sophistical refutation is founded upon a Fallacy of Accident, the respondent ought to apply one and the same solution to all. He will say: “The conclusion does not necessarily follow from the premisses”; and he will be prepared with an example, in which the conclusion obtained under this fallacy is notoriously untrue.114 “Do you know Koriskus?” — “Yes.” “Do you know the distant person coming this way?” — “No.” “That distant person is Koriskus: therefore you know, and you do not know, the same person.” The inference here is not necessary. To be coming this way — is an accident of Koriskus; and, because you do not know the accident, we cannot infer that you do not know the subject; such may or may not be the case.115

114 Soph. El. xxiv. p. 179, a. 30: ῥητέον οὖν συμβιβασθέντας ὁμοίως πρὸς ἅπαντας ὅτι οὐκ ἀναγκαῖον· ἔχειν δὲ δεῖ προφέρειν τὸ οἷον.

115 Ibid. a. 35-b. 7.

The major premiss upon which the preceding sophistical refutation must rest, is, That it is impossible both to know and not to know the same thing. This must be put as a direct question by the questioner, and must be conceded by the respondent, before the intended refutation can be made good. Now there are some persons who solve the refutation by answering this question in the negative, and by saying that it is possible both to know and not to know the same thing, only not in the same respect: such is the case when we know Koriskus, but do not know Koriskus approaching from a distance.116 Aristotle disapproves this mode of solution, as well as another mode which refers the fallacy to equivocation of terms. He points out that there are many other sophistical refutations, coming under the general head of Fallaciæ Accidentis, to which such solution will not apply; and that there ought to be one uniform mode of solution applicable to every fallacy coming under the same general head; though he admits at the same time that particular sophistical refutations may be vicious in more than one way. He says, moreover, that this contradiction or negation of the premiss is no true solution; for a solution ought to bring to view clearly the reason why the fallacious refutation appears to be a real refutation. Thus the Fallacia Accidentis consists in an inference that what is true of an accident is true also of the subject thereof: you explain that such inference, though apparently cogent, has no real cogency, and in that explanation consists the only proper solution of the fallacy.117

116 Ibid. b. 7, 18, 37: λύουσι δέ τινες ἀναιροῦντες τὴν ἐρώτησιν· φασὶ γὰρ ἐνδέχεσθαι ταὐτὸ πρᾶγμα εἰδέναι καὶ ἀγνοεῖν, ἀλλὰ μὴ κατὰ ταὐτό.

Mr. Poste (pp. 152-157) translates ἀναιροῦντες τὴν ἐρώτησιν — “contradicting the thesis,” and he expresses his surprise at the assertion, observing (very truly) that contradiction of the thesis is the very opposite of a solution; it helps in the very work which the refutation aims at accomplishing. But I cannot think that ἐρώτησις does mean “the thesis,” either here or in the other passage to which Mr. Poste refers (xxii. p. 178, b. 14). I think it means a premiss which the respondent has conceded, or must be presumed to have conceded, essential to the validity of the refutation. The term ἐρώτησις cannot surely, with any propriety, be applied to the thesis. It means either a question, or what is conceded in reply to a question; and the thesis cannot come under either one meaning or the other, being the proposition which the respondent sets out by affirming and undertakes to defend.

117 Soph. El. xxiv. p. 179, b. 23: ἦν γὰρ ἡ λύσις ἐμφάνισις ψευδοῦς συλλογισμοῦ, παρ’ ὃ ψευδής.

In like manner, all those Fallacies which come under the general head of A dicto Secundum Quid ad dictum Simpliciter, can only be solved by pointing out, in each particular case, in what terms this confusion is concealed — wherein resides the inference apparently cogent which is mistaken for one really cogent. The respondent is driven to an apparent contradiction, by having granted premisses from which the inference is derivable that both sides of the Antiphasis are true — that the same predicate A may be both affirmed and denied of the same subject B. He solves the contradiction by analysing the Antiphasis, and by showing that affirmation is secundum quid, while denial is simpliciter; and that there is a contradiction not real, but only apparent, between the two.118

118 Ibid. xxv. p. 180, a. 23-31.

In like manner, the Fallacy Ignoratio Elenchi will be solved by analysing the two supposed counter-propositions of the Antiphasis, and by showing that there is no real contradiction or inconsistency between them.119

119 Ibid. xxvi. p. 181, a. 1-14.

In regard to the Fallacies under Petitio Principii, the respondent if he perceives that the premiss asked of him involves such a fallacy, must refuse to grant it, however probable it may be in itself. If he does not perceive this until after he has granted it, he must throw back the charge of mal-procedure upon the questioner; declaring that an Elenchus involving assumption of the matter in question is null, and that the concession was made under the supposition that some separate and independent syllogism was in contemplation.120

120 Ibid. xxvii. p. 181, a. 15-21.

There are two distinct ways in which the Fallacia Consequentis may be employed. The predicate may be an universal, comprehending the subject: because animal always goes along with man, it is falsely inferred that man always goes along with animal; or it is falsely inferred that not-animal always goes along with not-man. The fallacy is solved when this is pointed out. The last inference is only valid when the terms are inverted; if animal always goes along with man, not-man will always go along with not-animal.121

121 Ibid. xxviii. p. 181, a. 22-30. ἀνάπαλιν γὰρ ἡ ἀκολούθησις.

If the sophistical refutation includes more premisses than are indispensable to the conclusion, the respondent, after having satisfied himself that this is the fact, will point out the mal-procedure of the questioner, and will say that he conceded the superfluous premiss, not because it was in itself probable but, because it seemed relevant to the debate; while nevertheless the questioner has made no real or legitimate application of it towards that object.122 This is the mode of solution applicable in the case of the Fallacies coming under the head Non Causa pro Causâ.123

122 Soph. El. xxix. p. 181, a. 31-35.

123 Schol. p. 318, a. 36, Br.

Where the sophistical questioner tries to refute by the Fallacia Plurium Interrogationum (i.e., by putting two or more questions as one), the respondent should forthwith divide the complex question into its component simple questions, and make answer accordingly. He must not give one answer, either affirmative or negative, to that which is more than one question. Even if he does give one answer, he may sometimes not involve himself in any contradiction; for it may happen that the same predicate is truly affirmable, or truly deniable, of two or more distinct and independent subjects. Often, however, the contrary is the case: no one true answer, either affirmative or negative, can be given to one of these complex questions: the one answer given, whatever it be, must always be partially false or inconsistent.124 Suppose two subjects, A and B, one good, the other bad: if the question be, Whether A and B are good or bad, it will be equally true to say — Both are good, or, Both are bad, or, Both are neither good nor bad. There may indeed be other solutions for this fallacy: Both or All may signify two or more items taken individually, or taken collectively; but the only sure precaution is — one answer to one question.125

124 Soph. El. xxx. p. 181, a. 38: οὔτε πλείω καθ’ ἑνὸς οὔτε ἓν κατὰ πολλῶν, ἀλλ’ ἓν καθ’ ἑνὸς φατέον ἢ ἀποφατέον.

Suppose that, instead of aiming at a seeming refutation, the Sophist tries to convict the respondent of Tautology. The source of this embarrassment is commonly the fact that a relative term is often used and conveys clear meaning without its correlate, though the correlate is always implied and understood. The respondent must avoid this trap by refusing to grant that the relative has any meaning at all without its correlate; and by requiring that the correlate shall be distinctly enunciated along with it. He ought to treat the relative without its correlate as merely a part of the whole significant expression — as merely syncategorematic; just as ten is in the phrase — ten minus one, or as the affirmative word is in a negative proposition.126 Thus he will not recognize double as significant by itself without its correlate half, nor half without its correlate double; although in common parlance such correlate is often understood without being formally enunciated.

126 Soph. El. Xxxi. p. 181, b. 26: οὐ δοτέον τῶν πρός τι λεγομένων σημαίνειν τι χωριζομένας καθ’ αὑτὰς τὰς κατηγορίας.

Mr. Poste observes in his note:— “The sophistic locus of tautology may be considered as a caricature of a dialectic locus. One fault which dialectic criticism finds with a definition is the introduction of superfluous words.” He then cites Topic. VI. ii. (p. 141, a. 4, seq.); but in this passage we find that the repetition of the same word is declared not to be an argumentative impropriety, so that the Sophist would gain nothing by driving his opponent into tautology.

Lastly, another purpose which Aristotle ascribes to the Sophist, is that of driving the respondent into a Solecism — into some grammatical or syntactical impropriety, such as, using a noun in the wrong case or gender, using a pronoun with a different gender or number from the noun to which it belongs, &c. He points out that the solution of these verbal puzzles must be different for each particular case; in general, when thrown into a regular syllogistic form, even the questioner himself will be found to speak bad Greek. The examples given by Aristotle do not admit of being translated into a modern language, so as to preserve the solecism that constitutes their peculiarity.127

127 Soph. El. xxxii. p. 182, a. 7-b. 5.

After having thus gone through the different artifices ascribed to the Sophist, and the ways of solving or meeting them, Aristotle remarks that there are material distinctions between the different cases which fall under one and the same general head of Sophistical Paralogism. Some cases there are in which both the fallacy itself, and the particular point upon which it turns, are obvious and discernible at first sight. In other cases, again, an ordinary person does not perceive that there is any fallacy at all; or, if he does perceive it, he often does not detect the seat of the fallacy, so that one man will refer the case to one general head, and another, to a different one.128 Thus, for example, Fallacies of Equivocation are perhaps the most frequent and numerous of all fallacies; some of them are childish and jocular, not really imposing upon any one; but there are others again in which the double meaning of a word is at first unnoticed, and is disputed even when pointed out, so that it can only be brought to light by the most careful and subtle analysis. This happens especially with terms that are highly abstract and general: which are treated by many, including even philosophers like Parmenides and Zeno, as if they were not equivocal at all, but univocal.129 Again, the Fallaciæ Accidentis, and the other classes Extra Dictionem, are also often hard to detect. On the whole, it is often hard to determine, not merely to which of the classes any case of fallacy belongs, but even whether there is any fallacy at all — whether the refutation is, or is not, a valid one.130

128 Ibid. xxxiii. p. 182, b. 6-12.

129 Soph. El. xxxiii. p. 182, b. 13-25: ὥσπερ οὖν ἐν τοῖς παρὰ τὴν ὁμωνυμίαν, ὅσπερ δοκεῖ τρόπος εὐηθέστατος εἶναι τῶν παραλογισμῶν, τὰ μὲν καὶ τοῖς τυχοῦσίν ἐστι δῆλα — τὰ δὲ καὶ τοὺς ἐμπειροτάτους φαίνεται λανθάνειν· σημεῖον δὲ τούτων ὅτι μάχονται πολλάκις περὶ ὀνομάτων, οἷον πότερον ταὐτὸ σημαίνει κατὰ πάντων τὸ ὂν καὶ τὸ ἓν ἢ ἕτερον.

130 Ibid. b. 27: ὁμοίως δὲ καὶ περὶ τοῦ συμβεβηκότος καὶ περὶ τῶν ἄλλων ἕκαστον, οἱ μὲν ἔσονται ῥᾴους ἰδεῖν οἱ δὲ χαλεπώτεροι τῶν λόγων· καὶ λαβεῖν ἔν τινι γένει, καὶ πότερον ἔλεγχος ἢ οὐκ ἔλεγχος, οὐ ῥᾴδιον ὁμοίως περὶ πάντων.

The pungent arguments in debate are those which bite most keenly, and create the greatest amount of embarrassment and puzzle.131 In dialectical debate a puzzle arises, when the respondent finds that a correct syllogism has been established against him, and when he does not at once see which among its premisses he ought to controvert, in order to overthrow the conclusion. In the eristic or sophistic debate the puzzle of the respondent is, in what language to enunciate his propositions so as to keep clear of the subtle objections which will be brought against him by the questioner.132 It is these pungent arguments that most effectually stimulate the mind to investigation. The most pungent of all is, where the syllogistic premisses are highly probable, yet where they nevertheless negative a conclusion which is also highly probable. Here we have an equal antithesis as to presumptive credibility, between the premisses taken together on one side and the conclusion on the other.133 We do not know whether it is in the premisses only, or in the conclusion, that we are to look for untruth: the conclusion, though improbable, may yet be true, while we may find that the true conclusion has been obtained from untrue premisses; or the conclusion may be both improbable and untrue, in which case we must look for untruth in one of the premisses also — either the major or the minor. This is the most embarrassing position of all. Another, rather less embarrassing, is, where our thesis will be confuted unless we can show the confuting conclusion to be untrue, but where each of the premisses on which the conclusion depends is equally probable, so that we do not at once see in which of them the cause of its untruth is to be sought. These two are the most pungent and perplexing argumentative conjunctures of dialectical debate.

131 Ibid. 32: ἔστι δὲ δριμὺς λόγος ὅστις ἀπορεῖν ποιεῖ μάλιστα· δάκνει γὰρ οὗτος μάλιστα.

132 Soph. El. xxxiii. p. 182, b. 33: ἀπορία δ’ ἐστὶ διττή, ἡ μὲν ἐν τοῖς συλλελογισμένοις, ὅ τι ἀνέλῃ τις τῶν ἐρωτημάτων, ἡ δ’ ἐν τοῖς ἐριστικοῖς, πῶς εἴπῃ τις τὸ προταθέν. The difficulty here pointed out, of finding language not open to some logical objection by an acute Sophist, is illustrated by what he himself states about the caution required for guarding his definitions against attack; see De Interpret. vi. p. 17, a. 34: λέγω δὲ ἀντικεῖσθαι τὴν τοῦ αὐτοῦ κατὰ τοῦ αὐτοῦ, μὴ ὁμωνύμως δέ, καὶ ὅσα ἄλλα προσδιοριζόμεθα πρὸς τὰς σοφιστικὰς ἐνοχλήσεις. What is here meant by σοφιστικαὶ ἐνοχλήσεις is expressed elsewhere by πρὸς τὰς λογικὰς δυσχερείας — Metaphys. Γ. iii. p. 1005, b. 21; N. i. p. 1087, b. 20. See the Scholia (pp. 112, 651, Br.) of Ammonius and Alexander upon the above passages of De Interpr. and Metaphys.

133 Soph. El. xxxiii. p. 182, b. 37-p. 183, a. 4: ἔστι δὲ συλλογιστικὸς μὲν λόγος δριμύτατος, ἂν ἐξ ὅτι μάλιστα δοκούντων ὅτι μάλιστα ἔνδοξον ἀναιρῇ· εἷς γὰρ ὢν ὁ λόγος, μετατιθεμένης τῆς ἀντιφάσεως, ἅπαντας ὁμοίους ἕξει τοὺς συλλογισμούς· ἀεὶ γὰρ ἐξ ἐνδόξων ὁμοίως ἔνδοξον ἀναιρήσει [ἢ κατασκευάσει]· διόπερ ἀπορεῖν ἀναγκαῖον. μάλιστα μὲν οὖν ὁ τοιοῦτος δριμύς, ὁ ἐξ ἴσου τὸ συμπέρασμα ποιῶν τοῖς ἐρωτήμασι. I transcribe this text as it is given by Bekker, Waitz, Bussemaker, and Mr. Poste. The editions anterior to Bekker had the additional words ἢ κατασκευάζῃ after ἀναιρῇ in the fourth line; and M. Barthélemy St. Hilaire in his translation defends and retains them. Bekker and the subsequent editors have omitted them, but have retained the last words ἢ κατασκευάσει in the seventh line. To me this seems inconsistent: the words ought either to be retained in both places or omitted in both. I think they ought to be omitted in both. I have enclosed them in brackets in the fifth line.

This difficult passage (not well explained by Alexander, Schol. p. 320, b. 9) requires the explanations of Waitz and Mr. Poste. The note of Mr. Poste is particularly instructive, because he expands in full (p. 164) the three “similar syllogisms” to which Aristotle here briefly alludes. The phrase μετατιθεμένης τῆς ἀντιφάσεως is determined by a passage in Analyt. Priora, II. viii. p. 59, b. 1: it means “employment of the contradictory of the conclusion, in combination with either one of the premisses, to upset the other.” The original syllogism is assumed to have two premisses, each highly probable, while the conclusion is highly improbable, being the negation of a highly probable proposition. The original syllogism will stand thus: All M is P; All S is M; Ergo, All S is P: the two premisses being supposed highly probable, and the conclusion highly improbable. Of course, therefore, the contradictory of the conclusion will be highly probable — Some S is not P. We take this contradictory and employ it to construct two new syllogisms as follows:— “All M is P; Some S is not P; Ergo Some S is not M. And again, Some S is not P: All S is M; Ergo, Some M is not P. All these three syllogisms are similar in this respect: that each has two highly probable premisses, while the conclusion is highly improbable.

But in eristic or sophistic debate our greatest embarrassment as respondents will arise when we do not at once see whether the refutative syllogism brought against us is conclusive or not, and whether it is to be solved by negation or by distinction.134 Next in order as to embarrassment stands the case, where we see in which of the two processes (negation or distinction) we are to find our solution, yet without seeing on which of the premisses we are to bring the process to bear; or whether, if distinction be the process required, we are to apply it to the conclusion, or to one of the premisses.135 A defective syllogistic argument is silly, when the deficient points are of capital importance — relating to the minor or to the middle term, or when the assumptions are false and strange; but it will sometimes be worthy of attention, if the points deficient are outlying and easily supplied; in which cases it is the carelessness of the questioner that is to blame, rather than the argument itself.136 Both the line of argument taken by the questioner, and the mode of solution adopted by the respondent, may be directed towards any one of three distinct purposes: either to the thesis and main subject discussed; or to the adversary personally (i.e., to the particular way in which he has been arguing); or to neither of these, but simply to prolong the discussion (i.e., against time). The solution may thus be sometimes such that it would take more time to argue upon it than the patience of the auditors will allow.137