Argument

Among the three artful modes of persuasion the most important is rational argument. Aristotle makes that clear in the very first sentence of his work, where he tells us that rhetoric is the counterpart of dialectic (1.1 1354a1). The use of‘counterpart’ recalls Plato’s Gorgias, in which Socrates is made to characterize rhetoric as the counterpart of cookery. It is said to be a form of flattery, and as such it is no art but an irrational habitude that aims at what is pleasant apart from what is best (464b-465d). Aristotle rejects this characterization and in doing so refers to dialectic. We are not to think of the dialectic that Plato trumpets in his Phaedrus, i. e., the method of collection and division (265d-266c). Rather, we are to think of the dialectic that Aristotle discusses in his work called Topics. Baldly stated,1 Aristotle characterizes dialectic as an exercise that proceeds by question and answer between two opponents. One person is presented with a problem like ‘Is ‘‘pedestrian biped animal’’ a definition of man, or not?’ (1.4 101b32-3). He chooses to defend one side or the other, and his opponent attempts to trap him in inconsistency by putting questions that must be answered by a simple yes or no. Some of the questions will involve necessary truths, so that the answer is automatic (assuming the respondent is not a fool or stubbornly dishonest). But most questions will not be of this kind. Rather they will concern acceptable opinions. That is to say, opinions that are acceptable to everyone or most people or wise men or experts in a given art (1.1 101b21-22, 1.10 104a8-9, 33-37).2 That imposes a constraint on the person asking questions. He must take account of what the respondent believes or at least is prepared to concede.3 For only when the questioner elicits the responses that he needs to demonstrate inconsistency, will he be able to achieve his goal.

Exercises of this sort were practiced in Plato’s Academy. As a member of the Academy, the young Aristotle will have participated in them and recognized a connection with rhetorical argument. In saying that, I am not suggesting that Aristotle conceived of rhetorical argument as a school exercise. Clearly he knew the difference between training within the Academy and real debate within the city-state. Nor am I suggesting that Aristotle conceived of rhetorical argument in terms of question and answer. To be sure, he knew that judicial procedure made room for occasional cross-examination (think of Socrates questioning Meletus in Plato’s Apology 24c-28a), but he also knew that orators generally put forward their arguments without directly involving an adversary. Rather, what impressed Aristotle is that both dialectical and rhetorical arguments take two distinct forms. They may have the form of a deduction or an induction. The former draws a conclusion from premises, while the latter adduces similar cases. Moreover, in both dialectic and rhetoric, deduction and induction do not rise to the level of a scientific argument. For typically the premises of dialectical and rhetorical deduction are not known truths but acceptable opinions, and the similar cases involved in induction constitute only a selection and therefore do not rule out the possibility of counter-examples.4 In addition, what is an acceptable opinion varies from group to group, so that not only the questioner in a dialectical exercise but also the orator in a civic setting needs to know what opinions will in fact be accepted. Otherwise, time will be wasted with misdirected questions and arguments that fail to persuade. And if that continues, the questioner will fail to trap his opponent, and the orator will not persuade his audience.

Aristotle calls rhetorical induction ‘example’ (1.2 1356b3-5). Two kinds are recognized. One involves past facts that the orator must search out; the other involves illustrative parallels and fables that the orator produces. The former is illustrated by the aggression of previous Persian kings (2.20 1393a31-b3):

We must prepare for war against the (present) king of Persia and not allow him to capture Egypt. For Darius did not cross over (to attack Greece) until he had first captured Egypt; but having captured (Egypt) he crossed over. And again Xerxes did not attack until he captured (Egypt); but having captured Egypt, he crossed over. It follows that the present (king) too, if he has captured (Egypt) will cross over. Therefore we ought not to allow (the present king to capture Egypt).

Here two past facts are adduced to support a conclusion regarding the present king of Persia: he will attack Greece if he first subdues Egypt. There is nothing wrong with this argument by example, but it should be noted that Aristotle begins and ends with a recommendation for future action: the present king must not be allowed to capture Egypt. To establish that recommendation a further argument is needed:

Either we ought not to allow the present king to capture Egypt in order that we may prevent an attack on Greece, or we ought to permit an attack on Greece by the king. But we ought not to permit an attack on Greece; therefore, we ought not to allow the king to capture Egypt.

This argument is not an induction based on past facts. Rather, it is a hypothetical syllogism that proceeds by way of a separative or disjunctive proposition: Either P or Q; but P; therefore not-Q. Aristotle’s syllogistic did not take account of such arguments, but that of his pupil Theophrastus seems to have done so.5

Illustrative parallels are explained by reference to Socratic comparisons: e. g., we ought not to choose a ruler by lot, for it would be similar to selecting athletes and helmsmen by lot. Fables are illustrated by stories about animals. We are told that

Stesichorus concluded a speech against giving a bodyguard to the dictator Phalaris by relating a story about a horse that permitted a man to bridle and mount him. We also learn that Aesop told the story of a fox that was covered with fleas, in order to defend a demagogue who was on trial for his life (1393b3-1394a2). Such stories are said to be suitable to the popular Assembly and comparatively easy to produce. In contrast, discovering past facts may be difficult, but facts are more useful in deliberation, for future events are most often similar to what has occurred in the past (1394a2-8).

Aristotle tells us that in argument by example the reasoning is from part to part or like to like, when both fall under the same genus, and one is better known than the other. By way of illustration, Aristotle introduces Dionysius of Syracuse, whose motive in asking for a bodyguard is unclear. To establish that he is aiming at a tyranny, two similar cases are adduced - both Peisistratus and Theagenes asked for a bodyguard when they plotted to become a tyrant - and the possibility of other similar cases is indicated, all of which are said to fall under the same general principle: the man who asks for a bodyguard is plotting a tyranny (1.2 1357b25-36). Here Aristotle emphasizes the universal or general principle involved in an argument based on example. The same is true in the Prior Analytics, where Aristotle offers a different illustration of argument by example. If we wish to show that making war on Thebes will have bad consequences for the Athenians, then we must first establish the premise ‘Making war on neighbors has bad consequences.’ Aristotle does that by adducing a single example: the Thebans suffered badly when they attacked their neighbor the Phocians. After that, Aristotle argues deductively to the conclusion that making war on Thebes will have bad consequences (2.24 68b41-69a13). What Aristotle does not tell us is how this plays out in practice. Does an orator regularly establish a general principle before drawing a particular conclusion? Or does he proceed directly from the better-known example(s) to the conclusion he wants to reach? Both are, of course, possible, but in emphasizing the need to establish a universal premise, Aristotle is focusing on the logic of argument by example. However, in the courtroom or the Assembly, an orator may omit that step and make a direct comparison between two particular cases. The listener supplies the universal premise and is pleased not to be told what he can easily supply for himself.

Aristotle calls the rhetorical syllogism an ‘enthymeme’ (1356b4). It is said to be concerned with the contingent, i. e., those matters about which men deliberate (1357a1-7, 13-15). Hence, the majority of its premises are not necessary. They hold only for the most part, and the same is true of its conclusions. That marks off the enthymeme from the scientific syllogism that draws necessary conclusions from premises that are themselves necessary (1357a22-32).6 A different feature of the enthymeme is that it involves few premises and often less than the primary syllogism, i. e., a logically complete syllogism (1357a16-17). There are two points here. The average citizen cannot follow an elaborate argument that proceeds through many steps, each involving its own premises. Therefore, the orator attempts to keep his arguments short and his premises few. That is a concession to the weakness of the audience (1357a3-4, 11-12), but equally the orator is well advised not to treat his audience as dolts. When a premise is obvious, it may be best left unstated, for to state it not only gives the impression of loquacity (2.22 1395b26) but also denies the audience an opportunity to fill a hole. Much as an audience is pleased with itself when it supplies the universal premise in an argument by example, so an audience takes satisfaction in supplying a premise that is unstated but necessary to an enthymeme.

For the sake of clarity, it should be underlined that the omission of a premise is not a defining mark of the enthymeme. It is only a frequent feature of the enthymeme (1357a17), and its frequency is motivated by audience psychology. A fast moving argument holds the attention of the listener, and when the listener supplies something for himself, he is pleased and likely to be favorably disposed to the orator.

It is one thing to understand the logical form of an enthymeme and another thing to be able to construct arguments that are successful in an Assembly or court of law. The latter requires a stock of premises, some of which are general in the sense that they are not subject-specific, while others are subject-specific. Aristotle understands that and draws a distinction between ‘topics’ and ‘species’ (1358a30). The former are illustrated by the topic of the more and the less. It is a general line of argument that can be used in discussing justice, physics, politics and many other subjects. The latter are explained as premises that are specific to each genus (of knowledge). They are the premises that make up most enthymemes and accordingly are discussed at length by Aristotle. The discussion is introduced by a division of rhetoric into three kinds, each of which is distinguished inter alia by its end (telos). Deliberative rhetoric is concerned with establishing the benefit or harm that may be expected from a future course of action. Judicial rhetoric focuses on proving the justice or injustice of some past action. And epideictic aims at showing the noble or shameful qualities of some person or group of persons (1.3 1358a36-1359a29). What follows is organized around these three kinds of rhetoric (1.4 1359a30-1.14 1375a21).

Among the three kinds, deliberative rhetoric is given pride of place. After telling us that men deliberate about things that they can control, Aristotle lists five subjects of deliberation: finances, war and peace, defense, imports and exports, and legislation (1.4 1359b21-23). Each subject is discussed with an emphasis on knowing the facts. Here is the beginning of Aristotle’s discussion of finances:

The man who is going to offer counsel concerning finances must know the number and extent of the city’s revenues, in order that if any source of revenue is missing it may be added, and if any source is deficient it may be increased (1359b24-26).

Here we have the beginnings of two enthymemes whose logical form is that of a mixed hypothetical syllogism. To be more specific, we have two arguments that proceed through a continuous proposition (i. e., the major premise, which is an if-then proposition) and an assumption (the minor premise) to a conclusion. Put schematically, the arguments have the form: If P, then Q; but P; therefore Q. Fleshed out they run as follows:

If any source of revenue is missing it may be added; but some source of revenue is missing; therefore it may be added.

If any source of revenue is deficient, it may be increased; but some source of revenue is deficient; therefore it may be increased.

In each of these two enthymemes, the assumptions are particular facts: ‘Some source of revenue is missing’ and ‘Some source is deficient’. Combined with the two continuous propositions, they create two arguments whose conclusions are action-guiding.

Aristotle appears never to have formulated this kind of hypothetical syllogism, but it is clear from Rhetoric 1.4 and elsewhere that he regarded enthymemes of this form as valid arguments.

In discussing war and peace, Aristotle continues to emphasize knowing the facts. He tells us to observe the wars of other people and how they ended, for similar results naturally follow from similar causes (1360a3-5). To illustrate Aristotle’s point, I refer to a passage (already mentioned above) from Prior Analytics 2.24. Aristotle cites Thebes, which came off poorly when it attacked the neighboring city of Phocis. On the basis of this example, Aristotle asserts the universal principle that making war on neighbors has bad consequences. And having gained that principle, he concludes by deduction that Athens should not attack its neighbor Thebes (68b41-69a13). The application to Rhetoric 1.4 is clear. Aristotle recognizes that facts about the wars of other cities7 can be used to form a general principle, and that principle can be applied to other particular cases. But as already observed the general principle need not be stated. The argument may be presented as a direct comparison.

After discussing the subjects of deliberative oratory and the facts that an orator needs to know, Aristotle turns to more general matters, beginning with the end or ‘target’ (skopos) at which a man individually and all men in common aim in their deliberations. Underlining that he is speaking in summary form, Aristotle tells us that the end is happiness and its parts. He then says that by way of illustration we must ascertain in general what happiness is. That is followed by four definitions of happiness, after which Aristotle adds that everyone would pretty well agree with one or more of these definitions (1.5 1360b4-18). Striking here is the way in which Aristotle dissociates himself from exactitude: we read ‘in summary form’, ‘by way of illustration’, ‘in general’, and we are offered four different definitions of happiness, no one of which is picked out as the correct definition. We should not be surprised, for Aristotle has told us that rhetoric will cease to be rhetoric and become a different art or science if it moves from what is generally acceptable to the first principles of a given art or science (1.2 1358a23-26, 1.4 1359b2-16). Moreover, the orator is concerned with persuading a particular audience and therefore must argue from the beliefs and conceptions actually held by a given audience. One audience may conceive of happiness under one definition, while another has a different idea of happiness. That opens the door to many different definitions of happiness. Aristotle seems to take account of that when he says ‘by way of illustration’. He does not aim at an exhaustive list, but instead satisfies himself with four definitions, each of which seems to have had advocates within Plato’s Academy.8

After discussing happiness, Aristotle takes up the useful or advantageous. He begins by telling us that men do not deliberate about the end but about the means to the end. And the means are what it is useful to do, and what is useful is good (1.6 1362a18-20). Aristotle then lists several ways in which something may be good. It may be something that ought to be chosen for its own sake, something that all creatures possessing sensation or intelligence pursue, that which by its presence creates a good condition, that which is productive or protective of a good condition, and that which prevents or removes the opposite condition (1362a21-29). Aristotle then goes on to draw some general conclusions from this list. For example, pleasure must be something good, for all animals, i. e., creatures possessing sensation, pursue it. And the virtues must be good, for it is by possessing them (by their presence) that men are in a good condition and prepared to produce good works and do good deeds (1362b2-4). How such propositions relate to deliberative rhetoric is straightforward. I limit myself to the second example. The orator who wishes to support the founding of a military academy will argue that such an academy produces courage and courage is a virtue. In other words, it is a good condition of soul that manifests itself in goods deeds. It is, therefore, something good and advantageous that ought to be promoted by the city-state. And that is best accomplished by a military education. An opponent of such an education might argue that courage is over-rated, and a military education all too often creates a character that is ill-suited for peacetime activities. Here there is a dispute, and Aristotle suggests a way in which the supporter of military education might respond. He might argue that the opposite of what is bad is good, and what is advantageous to one’s enemies is bad for the city. And since being cowardly or simply lacking courage is advantageous to one’s enemies, being courageous is an advantage to the city (1362b29-33) that should be promoted through military education.

The involvement of virtue in the preceding example is of some interest, for it suggests that the division of species or subject-specific premises into three classes is not absolute. And that is in fact the case. Within the discussion of happiness, which is the goal of deliberative oratory, virtue is recognized as a constituent part of happiness (1.5 1360b23), and as we have just seen, it may play a role in deliberations concerning the city-state. But virtue is also important for epideictic oratory. Indeed, Aristotle thinks that virtue is most at home in epideictic (speeches of praise), so that he refers from his discussion of the parts of happiness to the discussion of epideictic for an account of virtue (1.5 1362a13-14). There is no difficulty here as long as we recognize that Aristotle’s division of rhetoric into three genera is not rigorously imposed on the premises from which arguments are constructed. Deliberative and judicial oratory may be distinguished by different goals - the former aims at happiness and the latter at justice - but that does not rule out considerations of justice when deliberating about what is best for the city-state.9 A grossly unjust course of action may be rejected even though it would be to the advantage of most citizens and contribute to the long-term happiness of the city-state.

In addition to subject-specific premises, Aristotle recognizes a class of premises that are ‘common’ to the three kinds of rhetoric. They concern the possible and impossible, whether something has happened or will happen or not, and the great and the small. Aristotle tells us that the deliberative, forensic and epideictic orator must possess premises concerning these three subjects (1.3 1359a11-26), and he discusses them at some length, providing examples, many of which take the form of an if-then mixed hypothetical syllogism (2.19). And much as he speaks of the subject-specific premises as most at home in one kind or another, so with the common premises, he recognizes that each is most at home in one of the three kinds: amplification, i. e., the great (and the small) is said to be most at home in epideictic, past fact, i. e., what has happened (and not happened) in forensic oratory, and what is possible and will happen (or not happen) in deliberative oratory (1.18 1392a4-7). One may wonder whether there really are two distinct classes of premises - perhaps one ought to recognize a single class of common premises - but Aristotle sees a significant difference in emphasis. Some premises are largely though not exclusively tied to a particular kind of rhetoric, while others are less closely tied to one of the three kinds and therefore appropriately spoken of as common.

Aristotle also discusses ‘topics’. They are the lines of argument that he discusses at length in the work entitled Topics and more briefly in Rhetoric 2.23-24. In the latter place, twenty-eight valid topics and ten fallacious ones are listed. Twice Aristotle makes explicit reference to the Topics. The first occurs in regard to the topic ‘from different senses’ (1398a28-29) and the second in regard to that ‘from the parts’ (1399a6-7). If there is a problem here, it concerns the composition of the Rhetoric. When Aristotle first introduces the enthymeme and the example in Rhetoric 1.2, he recognizes two sources from which enthymemes are constructed: topics and subject-specific premises (1358a1-32). Topics are spoken of as common, but in this use ‘common’ does not refer to the premises that are common to the three kinds of rhetoric and discussed in 2.19. That is clear from the fact that topics are said to be applicable to many different subjects - all subjects (1358a32) - including physics (1358a10-14), which is not an area covered by rhetoric. However, when Aristotle turns in 1.3 to discuss the three kinds of rhetoric, he introduces the subject-specific premises and ignores topics (1359a11-26). That may be intelligible, for topics are not subject-specific. But later in a transitional passage, Aristotle summarizes what he has accomplished and what he needs to do to complete his project. Here again there is no mention of topics. Rather he tells us that he has discussed subject-specific premises and still needs to discuss common premises, the enthymeme and example. After that, he asserts, his project will be completed (2.18 1391b23-1392a4). That is curious not only because he omits any reference to topics but also because his treatment of topics is still to come. The difficulties here are real and most likely the product of two related factors. Different portions of the Rhetoric were composed at different times, and whoever combined them into the treatise that we know - whether Aristotle himself or a later editor or both - did less than perfect editing.