The subject-matter of dialectic debate, speaking generally, consists of Propositions and Problems, to be propounded as questions by the assailant and to be admitted or disallowed by the defendant. They will relate either to Expetenda and Fugienda, or they must bear, at least indirectly, upon some point of scientific truth or observed cognition.23 They will be either ethical, physical, or logical; class-terms which Aristotle declines to define, contenting himself with giving an example to illustrate each of them, while adding that the student should collect other similar examples, and gradually familiarize himself with the full meaning of the general term, through such inductive comparison of particulars.24

But it is not every problem coming under one of these three heads that is fit for dialectic debate. If a man propounds as subject for debate, Whether we ought to honour the gods or to love our parents, he deserves punishment instead of refutation: if he selects the question, Whether snow is white or not, he must be supposed deficient in perceptive power.25 What all persons unanimously believe, is unsuitable:26 what no one believes is also unsuitable, since it will not be conceded by any respondent. The problem must have some doubts and difficulties, in order to afford scope for discussion; yet it must not be one of which the premisses are far-fetched or recondite, for that goes beyond the limits of dialectic exercise.27 It ought to be one on which opinions are known to be held, both in the affirmative and in the negative; on which either the multitude differ among themselves, the majority being on one side, while yet there is an adverse minority; or some independent authority stands opposed to the multitude, such as a philosopher of eminence, a professional man or artist speaking on his own particular craft, a geometer or a physician on the specialities of his department. Matters such as these are the appropriate subjects for dialectic debate; and new matters akin to them by way of analogy may be imagined and will be perfectly admissible.28 Even an ingenious paradox or thesis adverse to prevailing opinions may serve the purpose, as likely to obtain countenance from some authority, though as yet we know of none.29

These conditions apply both to problems propounded for debate, and to premisses tendered on either side during the discussion. Both the interrogator and the respondent — the former having to put appropriate questions, and the latter to make appropriate answers — must know and keep in mind these varieties of existing opinion among the multitude as well as among the special dissident authorities above indicated. The dialectician ought to collect and catalogue such Endoxa, with the opinions analogous to them, out of written treatises and elsewhere;30 distributing them under convenient heads, such as those relating to good and evil generally, and to each special class of good, &c. Aristotle, however, admonishes him that he is debating problems not scientifically, but dialectically: having reference not to truth, but to opinion.31 If the interrogator were proceeding scientifically and didactically, he would make use of all true and ascertained propositions, whether the respondent conceded them or not, as premisses for his syllogism. But in Dialectic he is dependent on the concession of the respondent, and can construct his syllogisms only from premisses that have been conceded to him.32 Hence he must keep as closely as he can to opinions carrying extrinsic authority, as being those which the respondent will hesitate to disallow.33

Moreover, the form of the interrogation admissible in dialectic debate is peculiar. The respondent is not bound to furnish any information in his answer: he is bound only to admit, or to deny, a proposition tendered to him. You must not ask him, What is the genus of man? You must yourself declare the genus, and ask whether he admits it, in one or other of the two following forms — (1) Is animal the genus of man? (2) Is animal the genus of man, or not? to which the response is an admission or a denial.34

Dialectic procedure, both of the assailant and of the defendant, has to do with propositions and problems; accordingly, Aristotle introduces a general distribution of propositions under four heads. The predicate must either be Genus, or Proprium, or Accident, of its subject. But the Proprium divides itself again into two. It always reciprocates with, or is co-extensive with, its subject; but sometimes it declares the essence of the subject, sometimes it does not. When it declares the essence of the subject, Aristotle calls it the Definition; when it does not declare the essence of the subject, although reciprocating therewith, he reserves for it the title of Proprium. Every proposition, and every problem, the entire material of Dialectic, will declare one of these four — Proprium, Definition, Genus, or Accident.35 The Differentia, as being attached to the Genus, is ranked along with the Genus.36

The above four general heads include all the Predicables, which were distributed by subsequent logicians (from whom Porphyry borrowed) into five heads instead of four — Genus, Species, Differentia, Proprium, Accident; the Differentia being ranked as a separate item in the quintuple distribution, and the Species substituted in place of the Definition. It is under this quadruple classification that Aristotle intends to consider propositions and problems as matters for dialectic procedure: he will give argumentative suggestions applicable to each of the four successively. It might be practicable (he thinks) to range all the four under the single head of Definition; since arguments impugning Genus, Proprium, and Accident, are all of them good also against Definition. But such a simplification would be perplexing and unmanageable in regard to dialectic procedure.37

That the quadruple classification is exhaustive, and that every proposition or problem falls under one or other of the four heads, may be shown in two ways. First, by Induction: survey and analyse as many propositions as you will, all without exception will be found to belong to one of the four.38 Secondly, by the following Deductive proof:— In every proposition the predicate is either co-extensive and reciprocating with the subject, or it is not. If it does reciprocate, it either declares the essence of the subject, or it does not: if the former, it is the Definition; if the latter, it is a Proprium. But, supposing the predicate not to reciprocate with the subject, it will either declare something contained in the Definition, or it will not. If it does contain a part of the Definition, that part must be either a Genus or a Differentia, since these are the constituents of the Definition. If it does not contain any such part, it must be an Accident.39 Hence it appears that every proposition must belong to one or other of the four, and that the classification is exhaustive.

Moreover, each of the four Predicables must fall under one or other of the ten Categories or Predicaments. If the predicate be either of Genus or Definition, declaring the essence of the subject, it may fall under any one of the ten Categories; if of Proprium or Accident, not declaring essence, it cannot belong to the first Category (Οὐσία), but must fall under one of the remaining nine.40

The notion of Sameness or Identity occurs so often in dialectic debate, that Aristotle discriminates its three distinct senses or grades: (1) Numero; (2) Specie; (3) Genere. Water from the same spring is only idem specie, though the resemblance between two cups of water from the same spring is far greater than that between water from different sources. Even Idem Numero has different significations: sometimes there are complete synonyms; sometimes an individual is called by its proprium, sometimes by its peculiar temporary accident.41

Having thus classified dialectic propositions, Aristotle proceeds to the combination of propositions, or dialectic discourse and argument. This is of two sorts, either Induction or Syllogism; of both which we have already heard in the Analytica. Induction is declared to be plainer, more persuasive, nearer to sensible experience, and more suitable to the many, than Syllogism; while this latter carries greater compulsion and is more irresistible against professed disputants.42 A particular example is given to illustrate what Induction is. But we remark that though it is always mentioned as an argumentative procedure important and indispensable, yet neither here nor elsewhere does Aristotle go into any discriminative analysis of the conditions under which it is valid, as he does about Syllogism in the Analytica Priora.

What helps are available to give to the dialectician a ready and abundant command of syllogisms? Four distinct helps may be named:43 (1) He must make a large collection of Propositions; (2) He must study and discriminate the different senses in which the Terms of these propositions are used; (3) He must detect and note Differences; (4) He must investigate Resemblances.

1. About collecting Propositions, Aristotle has already indicated that those wanted are such as declare Endoxa, and other modes of thought cognate or analogous to the Endoxa:44 opinions of the many, and opinions of any small sections or individuals carrying authority. All such are to be collected (out of written treatises as well as from personal enquiry); nor are individual philosophers (like Empedokles) to be omitted, since a proposition is likely enough to be conceded when put upon the authority of an illustrious name.45 If any proposition is currently admitted as true in general or in most cases, it must be tendered with confidence to the respondent as an universal principle; for he will probably grant it, not being at first aware of the exceptions.46 All propositions must be registered in the most general terms possible, and must then be resolved into their subordinate constituent particulars, as far as the process of subdivision can be carried.47

2. The propositions having been got together, they must be examined in order to find out Equivocation or double meaning of terms. There are various ways of going about this task. Sometimes the same predicate is applied to two different subjects, but in different senses; thus, courage and justice are both of them good, but in a different way. Sometimes the same predicate is applied to two different classes of subjects, each admitting of being defined; thus, health is good in itself, and exercise is good as being among those things that promote health.48 Sometimes the equivocal meaning of a term is perceived by considering its contrary; if we find that it has two or more distinct contraries, we know at once that it has different meanings. Sometimes, though there are not two distinct contraries, yet the mere conjunction of the same adjective with two substantives shows us at once that it cannot mean the same in both49 (λευκὴ φωνή — λευκὸν χρῶμα). In one sense, the term may have an assignable contrary, while in another sense it may have no contrary; showing that the two senses are distinct: for example, the pleasure of drinking has for its contrary the pain of thirst; but the pleasure of scientifically contemplating that the diagonal of a square is incommensurable with the side, has no contrary; hence, we see that pleasure is an equivocal term.50 In one sense, there may be a term intermediate between the two contraries; in another sense, there may be none; or there may be two distinct intermediate terms for the two distinct senses; or there may be several intermediate terms in one of the senses, and only one or none in the other: in each of these ways the equivocation is revealed.51 We must look also to the contradictory opposite (of a term), which may perhaps have an obvious equivocation of meaning; thus, μὴ βλέπειν means sometimes to be blind, sometimes not to be seeing actually, whence we discover that βλέπειν also has the same equivocation.52 If a positive term is equivocal, we know that the privative term correlating with it must also be equivocal; thus, τὸ αἰσθάνεσθαι has a double sense, according as we speak with reference to mind or body; and this will be alike true of the correlating privative — τὸ ἀναίσθητον εἶναι.53 Farther, an equivocal term will have its derivatives equivocal in the same manner; and conversely, if the derivative be equivocal, the radical will be so likewise.54 The term must also be looked at in reference to the ten Categories: if its meanings fall under more than one Category, we know that it is equivocal.55 If it comprehends two subjects which are not in the same genus, or in genera not subordinate one to the other, this too will show that it is equivocal.56 The contrary, also, of the term must be looked at with a view to the same inference.57

Again, it will be useful to bring together the same term in two different conjunctions, and to compare the definitions of the two. Define both of them, and then deduct what is peculiar to each definitum: if the remainder be different, the term will be equivocal; if the remainder be the same, the term will be univocal. Thus, λευκὸν σῶμα will be defined, a body having such and such a colour: λευκὴ φωνή, a voice easily and distinctly heard: deduct σῶμα from the first definition, and φωνὴ from the second, the remainder will be totally disparate; therefore, the term λευκόν is equivocal.58 Sometimes, also, the ambiguity may be found in definitions themselves, where the same term is used to explain subjects that are not the same; whether such use is admissible, has to be considered.59 If the term be univocal, two conjunctions of it may always be compared as to greater or less, or in respect of likeness; whenever this cannot be, the term is equivocal.60 If, again, the term is used as a differentia for two genera quite distinct and independent of each other, it must be equivocal; for genera that are unconnected and not subordinate one to the other, have their differentiæ also disparate.61 And, conversely, if the term be such that the differentiæ applied to it are disparate, we may know it to be an equivocal term. The like, if the term be used as a species in some of its conjunctions, and as a differentia in others.62

3. Aristotle has thus indicated, at considerable length, the points to be looked for when we are examining whether a term is univocal or equivocal. He is more concise when he touches on the last two out of the four helps (ὄργανα) enumerated for supplying syllogisms when needed, — viz. the study of Differences and of Resemblances. In regard to the study of Differences, standing third, while he remarks that, where these are wide and numerous, they are sure without any precept to excite our attention, he advises that we should study the differences of subjects that are nearly allied, — those within the same genus, or comprehended in genera not much removed from one another, such as, the distinction between sensible perception and science. But he goes into no detail.63

4. In regard to the study of Resemblances, he inverts the above precept, and directs us to note especially the points of resemblance between subjects of great apparent difference.64 We must examine what is the quality common to all species of the same genus — man, horse, dog, &c.; for it is in this that they are similar. We may also compare different genera with each other, in respect to the analogies that are to be found in each: e.g., as science is to the cognizable, so is perception to the perceivable; as sight is in the eye, so is intellection in the soul; as γαλήνη is in the sea, so is νηνεμία in the air.65