Aristotle takes the four heads — Accident, Genus, Proprium, and Definition, in the order here enumerated. The thesis of which the predicate is enunciated as Accident, affirms the least, is easiest to defend, and hardest to upset.79 When we enunciate Genus or Proprium, we affirm, not merely that the predicate belongs to the subject (which is all that is affirmed in the case of Accident), but, also something more — that it belongs to the subject in a certain manner and relation. And when we enunciate Definition, we affirm all this and something reaching yet farther — that it declares the whole essence of the definitum, and is convertible therewith. Accordingly, the thesis of Definition, affirming as it does so very much, presents the most points of attack and is by far the hardest to defend.80 Next in point of difficulty, for the respondent, comes the Proprium.

Beginning thus with the thesis enunciating Accident, Aristotle enumerates no less than thirty-seven distinct loci or argumentative points of view bearing upon it. Most of them suggest modes of assailing the thesis; but there are also occasionally intimations to the respondent how he may best defend himself. In this numerous list there are indeed some items repetitions of each other, or at least not easily distinguishable.81 As it would be tedious to enumerate them all, I shall select some of the most marked and illustrative.

1. The respondent has enunciated a certain predicate as belonging in the way of accident, to a given subject. Perhaps it may belong to the subject; yet not as accident, but under some one of the other three Predicables. Perhaps he may have enunciated (either by explicit discrimination, or at least by implication contained in his phraseology) the genus as if it were an accident, — an error not unfrequently committed.82 Thus, if he has said, To be a colour is an accident of white, he has affirmed explicitly the genus as if it were an accident. And he has affirmed the same by implication, if he has said, White (or whiteness) is coloured. For this is a form of words not proper for the affirmation of a genus respecting its species, in which case the genus itself ought to stand as a literal predicate (White is a colour), and not to be replaced by one of its derivatives (White is coloured). Nor can the proposition be intended to be taken as affirming either proprium or definition; for in both these the predicate would reciprocate and be co-extensive with the subject, whereas in the present case there are obviously many other subjects of which it may be predicated that they are coloured.83 In saying, White is coloured, the respondent cannot mean to affirm either genus, proprium, or definition; therefore he must mean to affirm accident. The assailant will show that this is erroneous.

2. Suppose the thesis set up by the respondent to be an universal affirmative, or an universal negative. You (the interrogator or assailant) should review the particulars contained under these universals. Review them not at once as separate individuals, but as comprised in subordinate genera and species; beginning from the highest, and descending down to the lowest species which is not farther divisible except into individuals. Thus, if the thesis propounded be, The cognition of opposites is one and the same cognition; you will investigate whether this can be truly predicated respecting all the primary species of Opposita: respecting Relata and Correlata, respecting Contraries, respecting Contradictories, respecting Habitus and Privatio. If, by going thus far, you obtain no result favourable to your purpose,84 you must proceed farther, and subdivide until you come to the lowest species:— Is the cognition of just and unjust one and the same? that of double and half? of sight and blindness? of existence and non-existence? If in all, or in any one, of these cases you can show that the universal thesis does not hold, you will have gained your point of refuting it. On the other hand, if, when you have enumerated many particulars, the thesis is found to hold in all, the respondent is entitled to require you to grant it as an universal proposition, unless you can produce a satisfactory counter-example. If you decline this challenge, you will be considered an unreasonable debater.85

3. You will find it useful to define both the accident predicated in the thesis, and the subject respecting which it is predicated, or at least one of them: you will see then whether these definitions reveal anything false in the affirmation of the thesis. Thus, if the thesis affirms that it is possible to do injustice to a god, you will define what is meant by doing injustice. The definition is — hurting intentionally: you can thus refute the thesis by showing that no injustice to a god can possibly be done; for a god cannot be hurt.86 Or let the thesis maintained be, The virtuous man is envious. You define envy, and you find that it is — vexation felt by reason of the manifest success of some meritorious man. Upon this definition it is plain that the virtuous man cannot feel envy: he would be worthless, if he did feel it. Perhaps some of the terms employed in your definition may themselves require definition; if so, you will repeat the process of defining until you come to something plain and clear.87 Such an analysis will often bring out some error at first unperceived in the thesis.

4. It will be advisable, both for assailant and respondent, to discriminate those cases in which the authority of the multitude is conclusive from those in which it is not. Thus, in regard to the meaning of terms and in naming objects, we must speak like the multitude; but, when the question is as to what objects deserve to be denominated so and so, we must not feel bound by the multitude, if there be any special dissentient authority.88 That which produces good health we must call wholesome, as the multitude do; but, in calling this or that substance wholesome, the physician must be our guide.

5. Aristotle gives more than one suggestion as to those cases in which the terms of the thesis have a double or triple sense, yet in which the thesis is propounded either as an universal affirmative or as an universal negative. If the respondent is himself not aware of the double sense of his thesis, while you (the questioner) are aware of it, you will prove the point which you are seeking to establish against him in one or other of the two senses, if you cannot prove it in both. If he is aware of it in the double sense, he will insist that you have chosen the sense which he did not intend.89 This mode of procedure will be available to the respondent as well as to you; but it will be harder to him, since his thesis is universal. For, in order to make good an universal thesis, he must obtain your assent to a preliminary assumption or convention, that, if he can prove it in one sense of the terms, it shall be held proved in both; and, unless the proposition be so plausible that you are disposed to grant him this, he will not succeed in the procedure.90 But you on your side, as refuting, do not require any such preliminary convention or acquiescence; for, if you prove the negative in any single case, you succeed in overthrowing the universal affirmative, while, if you prove the affirmative in any single case, you succeed in overthrowing the universal negative.91 Such procedure, however, is to be adopted only when you can find no argument applicable to the equivocal thesis in all its separate meanings; this last sort of argument, wherever it can be found, being always better.92

In cases where the double meaning is manifest, the two meanings must be distinguished by both parties, and the argument conducted accordingly. Where the term has two or more meanings (not equivocal but) related to each other by analogy, we must deal with each of these meanings distinctly and separately.93 If our purpose is to refute, we select any one of them in which the proposition is inadmissible, neglecting the others: if our purpose is to prove, we choose any one in which the proposition is true, neglecting the others.94

6. Observe that a predicate which belongs to the genus does not necessarily belong to any one of its species, but that any predicate which belongs to one of the species does belong also to the genus; on the other hand, that any predicate which can be denied of the genus may be denied also of all its contained species, but that any predicate which can be denied of some one or some portion of the contained species cannot for that reason be denied of the genus. You may thus prove from one species to the genus, and disprove from the genus to each one species; but not vice versÂ. Thus, if the respondent grants that there exist cognitions both estimable and worthless, you are warranted in inferring that there exist habits of mind estimable and worthless; for cognition is a species under the genus habit of mind. But if the negative were granted, that there exist no cognitions both estimable and worthless, you could not for that reason infer that there are no habits of mind estimable and worthless. So, if it were granted to you that there are judgments correct and erroneous, you could not for that reason infer that there were perceptions of sense correct and erroneous; perceiving by sense being a species under the genus judging. But, if it were granted that there were no judgments correct and erroneous, you might thence infer the like negative about perceptions of sense.95

7. Keep in mind also that if there be any subject of which you can affirm the genus, of that same subject you must be able to affirm one or other of the species contained under the genus. Thus, if science be a predicate applicable, grammar, music, or some other of the special sciences must also be applicable: if any man can be called truly a scientific man, he must be a grammarian, a musician, or some other specialist. Accordingly, if the thesis set up by your respondent be, The soul is moved, you must examine whether any one of the known varieties of motion can be truly predicated of the soul, e.g., increase, destruction, generation, &c. If none of these special predicates is applicable to the soul, neither is the generic predicate applicable to it; and you will thus have refuted the thesis. This locus may serve as a precept for proof as well as for refutation; for, equally, if the soul be moved in any one species of motion, it is moved, and, if the soul be not moved in any species of motion, it is not moved.96

8. Where the thesis itself presents no obvious hold for interrogation, turn over the various definitions that have been proposed of its constituent terms; one or other of these definitions will often afford matter for attack.97 Look also to the antecedents and consequents of the thesis — what must be assumed and what will follow, if the thesis be granted. If you can disprove the consequent of the proposition, you will have disproved the proposition itself. On the other hand, if the antecedent of the proposition be proved, the proposition itself will be proved also.98 Examine also whether the proposition be not true at some times, and false at other times. The thesis, What takes nourishment grows necessarily, is true not always, but only for a certain time: animals take nourishment during all their lives, but grow only during a part of their lives. Or, if a man should say that knowing is remembering, this is incorrect; for we remember nothing but events past, whereas we know not only these, but present and future also.99

9. It is a sophistical procedure (so Aristotle terms it) to transfer the debate to some point on which we happen to be well provided with arguments, lying apart from the thesis defended. Such transfer, however, may be sometimes necessary. In other cases it is not really but only apparently necessary; in still other cases it is purely gratuitous, neither really nor apparently necessary. It is really necessary, when the respondent, having denied some proposition perfectly relevant to his thesis, stands to his denial and accepts the debate upon it, the proposition being one on which a good stock of arguments may be found against him; also, when you are endeavouring to disprove the thesis by an induction of negative analogies.100 It is only apparently, and not really, necessary, when the proposition in debate is not perfectly relevant to the thesis, but merely has the semblance of being so. It is neither really nor apparently necessary, when there does not exist even this semblance of relevance, and when some other way is open of bringing bye-confutation to bear on the respondent. You ought to avoid entirely such a procedure in this last class of cases; for it is an abuse of the genuine purpose of Dialectic. If you do resort to it, the respondent should grant your interrogations, but at the same time notify that they are irrelevant to the thesis. Such notification will render his concessions rather troublesome than advantageous for your purpose.101

10. You will recollect that every proposition laid down or granted by the respondent carries with it by implication many other propositions; since every affirmation has necessary consequences, more or fewer. Whoever says that Sokrates is a man, has said also that he is an animal, that he is a living creature, biped, capable of acquiring knowledge. If you can disprove any of these necessary consequences, you will have disproved the thesis itself. You must take care, however, that you fix upon some one of the consequences which is really easier, and not more difficult, to refute than the thesis itself.102

11. Perhaps the thesis set up by the respondent may be of such a nature that one or other of two contrary predicates must belong to the subject; e.g., either health or sickness. In that case, if you are provided with arguments bearing on one of the two contraries, the same arguments will also serve indirectly for proof, or for disproof, of the other. Thus, if you show that one of the two contraries does belong to the subject, the same arguments prove that the other does not; vice versÂ, if you show that one of them does not belong, it follows that the other does.103

12. You may find it advantageous, in attacking the thesis, to construe the terms in their strict etymological sense, rather than in the sense which common usage gives them.104

13. The predicate may belong to its subject either necessarily, or usually, or by pure hazard. You will take notice in which of these three ways the respondent affirms it, and whether that which he chooses is conformable to the fact. If he affirms it as necessary, when it is really either usual or casual, the thesis will be open to your attacks. If he affirms it without clearly distinguishing in which of the three senses he intends it to be understood, you are at liberty to construe it in that one of the three senses which best suits your argument.105

14. Perhaps the thesis may have predicate and subject exactly synonymous, so that the same thing will be affirmed as an accident of itself. On this ground it will be assailable.106

15. Sometimes the thesis will have more than one proposition contrary to it. If so, you may employ in arguing against it that one among its various contraries which is most convenient for your purpose.107 Perhaps the predicate (accidental) of the thesis may have some contrary: if it has, you will examine whether that contrary belongs to the subject of the thesis; and, should such be the case, you may use it as an argument to refute the thesis itself.108 Or the predicate of the thesis may be such that, if the thesis be granted, it will follow as a necessary consequence that contrary predicates must belong to the same subject. Thus, if the thesis be that the Platonic Ideas exist in us, it follows necessarily that they are both in motion and at rest; both perceivable by sense, and cogitable by intellect.109 As these two predicates (those constituting the first pair as well as the second pair) are contrary to each other, and cannot both belong to the same subject, this may be used as an argument against the thesis from which such consequence follows.

16. We know that whatever is the recipient of one of two contraries, is capable also of becoming recipient of the other. If, therefore, the predicate of the thesis has any contrary, you will examine whether the subject of the thesis is capable of receiving such contrary. If not, you have an argument against the thesis. Let the thesis be, The appetitive principle is ignorant. If this be true, that principle must be capable of knowledge.110 Since this last is not generally admitted, you have an argument against the thesis.

17. We recognize four varieties of Opposita: (1) Contradictory; (2) Contrary; (3) Habitus and Privatio; (4) Relata. You will consider how the relation in each of these four varieties bears upon the thesis in debate.

In regard to Contradictories, you are entitled, converting the terms of the thesis, to deny the predicate of the converted proposition respecting the negation of the subject. Thus, if man is an animal, you are entitled to infer, What is not an animal is not a man. You will prove this to be an universal rule by Induction; that is, by citing a multitude of particular cases in which it is indisputably true, without possibility of finding any one case in which it does not apply. If you can prove or disprove the converted obverse of the thesis — What is not an animal is not a man — you will have proved or disproved, the thesis itself, Man is an animal. This locus is available both for assailant and respondent.111

In regard to Contraries, you will study the thesis, to see whether the contrary of the predicate can be truly affirmed respecting the contrary of the subject, or whether the contrary of the subject can be truly affirmed respecting the contrary of the predicate. This last alternative occurs sometimes, but not often; in general the first alternative is found to be true. You must make good your point here also by Induction, or by repetition of particular examples. This locus will serve either for the purpose of refutation or for that of defence, according to circumstances. If neither of the two alternatives above-mentioned is found correct, this is an argument against the thesis.112

In regard to Habitus and Privatio, the rule is the same as about Contraries; only that the first of the two above alternatives always holds, and the second never occurs.113 If sensible perception can be predicated of vision, insensibility also can be predicated of blindness; otherwise, the thesis fails.

In regard to Relata, the inference holds from the correlate of the subject to the correlate of the predicate. If knowledge is belief, that which is known is believed; if vision is sensible perception, that which is visible is sensibly perceivable. Some say that there are cases in which the above does not hold; e.g., That which is sensibly perceivable is knowable; yet sensible perception is not knowledge. But this objection is not valid; for many persons dispute the first of the two propositions. This locus will be equally available for the purpose of refutation — thus, you may argue — That which is sensibly perceivable is not knowable, because sensible perception is not knowledge.114

18. You will look at the terms of the proposition, also, in regard to their Derivatives, Inflections, &c., and to matters associated with them in the way of production, preservation, &c. This locus serves both for proof and for refutation. What is affirmable of the subject, is affirmable also of its derivatives: what is not affirmable of the derivatives, is not affirmable of the subject itself.115

19. Arguments may often be drawn, both for proof and for refutation, from matters Similar or Analogous to the subject or predicate of the thesis. Thus, if one and the same cognition comprehends many things, one and the same opinion will also comprehend many things. If to possess vision is to see, then also to possess audition is to hear. If to possess audition is not to hear, then neither is to possess vision to see. The argument may be urged whether the resemblance is real or only generally supposed. Sometimes, however, the inference will not hold from one to many. Thus, if to know is to cogitate, then to know many things should be to cogitate many things. But this last is impossible. A man may know many things, but he cannot cogitate many things; therefore, to know is not to cogitate.116

20. There are various loci for argument, arising from degrees of Comparison — more, less, equally. One is the argument from concomitant variations, which is available both for proof and for disproof. If to do injustice is evil, to do more injustice is more evil. If an increase in degree of the subject implies an increase in degree of the predicate, then the predicate is truly affirmed; if not, not. This may be shown by Induction, or repetition of particular instances.117 Again, suppose the same predicate to be affirmable of two distinct subjects A and B, but to be more probably affirmable of A than of B. Then, if you can show that it does not belong to A, you may argue (À fortiori) that it does not belong to B; or, if you can show that it belongs to B, you may argue (À fortiori) that it belongs also to A. Or, if two distinct predicates be affirmable respecting the same subject but with unequal degrees of probability, then, if you can disprove the more probable of the two, you may argue from thence in disproof of the less probable; and, if you can prove the less probable, you may argue from thence in proof of the more probable. Or, if two distinct predicates be affirmable respecting two distinct subjects but with unequal degrees of probability, then, if you can disprove the more probable you may argue from thence against the less probable; and, if you can prove the less probable, you are furnished with an argument in proof of the more probable.118 If the degrees of probability, instead of being unequal, are equal or alike, you may still, in the cases mentioned, argue in like manner from proof or disproof of the one to proof or disproof of the other.119

117 Ibid. b. 37-p. 115, a. 5: e?s? d? t?? ????? t?p?? t?ssa?e?, e?? ?? e? ???????e? t? ????? t? ?????, — ???s??? d? p??? ?f? ? t?p??· e? ?? ??? ???????e? t? t?? ?p??e????? ?p?d?se? ? t?? s?e???t?? ?p?d?s??, ?a??pe? e???ta?, d???? ?t? s????e?, e? d? ? ???????e?, ?? s????e?. t??t? d’ ?pa???? ??pt???.