K O D I K A S / C O D E Volume 39 (2016) · No. 1 - 2 Gunter Narr Verlag Tübingen How Deceptive are Rhetorical Arguments from Signs or Likelihoods? - Aristotle on Apparent Rhetorical Syllogisms 1 Markus H. Wörner (Galway/ Ireland) Dividing lines between genuine and apparent rhetorical syllogisms from signs or likelihoods (enthymemes), such as logical necessity or different sets of topoi, do not sufficiently separate informed, reasonable and useful enthymemes from those which are fallacious or deceitful; nor good rhetoric from sham rhetoric. Only some on the spectrum of enthymematic arguments have been taken into account in recent Aristotelian research. The field to be considered needs widening, and deductions need their place within this wider spectrum. Finally, a suggestion ad mentem Aristotelis for classifying enthymematic arguments is made here. Rather than rejecting apparent enthymemes as fallacious or deceitful per se, it exposes their different grades of reasonableness instead. 1. Introduction Rhetoric, its theory, its ubiquitous practice and its claim to be a method capable of supporting reasonableness, specifically in areas of human action in which opinions are divided yet convincing solutions need to be found, frequently raises suspicion; this occurs particularly among philosophers, critics of ideologies or scientists. Immanuel Kant is commonly regarded as the chief ally for those who wish to raise these concerns. In his Critique of the Power of Reason (par. 53), he describes the part of rhetoric termed ‘ Rednerkunst ’ (ars oratoria) as a deceitful art, exploiting others ’ weakness for one ’ s own purposes by perverting audiences ’ judgments - which they might make correctly if given the opportunity for calm reflection. 2 Not least, political or religious judgments - to 1 I should like to thank Marja-Liisa Kakkuri-Knuuttila and Ricca Edmondson for truly helpful discussions of the content of this paper. Errors remain my own responsibility. In this paper I use the term he for historical reasons: Aristotle will have had male orators in mind, though of course this has no implications for the logical questions involved. 2 “ I must confess ” , Kant remarks, “ that a beautiful poem has always given me a pure enjoyment, whereas reading the best speech of a Roman popular speaker or a contemporary speaker in parliament or the pulpit has always been mixed with the disagreeable feeling of disapproval of a deceitful art, which understands how to move people, like machines, to a judgment in important matters which must lose all weight for them in calm re flection. Eloquence and wellspokenness (together, rhetoric) belong to beautiful art; but the art of the orator (Rednerkunst/ ars oratoria) as the art of using the weakness of people for one ’ s own purposes (however wellintentioned or even really good these may be) is not worthy of any respect at all. Further, both in Athens and in Rome it reached its highest level only at a time when the state was rushing toward its ruin and a truly patriotic mention only those referred to by Kant himself - , are distorted in this way unless, he believes, one is able to speak with vigour as a good person experienced in speaking but without artifice. On the other hand, rhetoric appears to be the medium for most of our reasoning when we aim at convincing others (and ourselves) on the basis of knowledge of facts or values which are part and parcel of the area of the contingent, that which might be otherwise. According to H.-G. Gadamer, rhetoric is [. . .] the only advocate of a truth claim which defends verisimilitude, the eikos (verisimile), and whatever makes sense (einleuchtend) to common sense (gemeine Vernunft), against science ’ s claim of proof (Beweis) and certainty (Gewissheit). Convincing (Überzeugen) and making sense in the absence of final proof is obviously the goal and measure of understanding and interpreting as well as of the art of speaking and persuasion [. . .].The ubiquity of rhetoric is unlimited (Gadamer 1971: 62; my translation, M.W.). Even the most rigorous scientists depend on rhetoric when they teach or communicate their results effectively to a wider audience. Kant and Gadamer do not mean quite the same thing when they talk about rhetoric. An art of deceitfully using audiences ’ weakness for one ’ s own purposes by causing them to abandon their sober judgment in favour of what they hear, on the one hand, and acting as an advocate for whatever makes justified sense in the area of the contingent, are far from identical. Unfortunately, they are also not mutually exclusive. The very fact of this divergence of assessments of rhetoric seems to reflect an ambiguity intrinsic to the art itself. This ambiguity does not only affect its treatment of the arousal of emotions (pathos) and character presentation (ethos), traditionally regarded as grounds for conviction typically evoked by rhetorical means (pisteis entechnoi), it also affects the core of the rhetorical method itself, the argument (logos). Rhetoric seems to provide instructions for rational arguments but also for merely apparent ones, for arousing appropriate but also unsuitable emotions and for presenting the speaker ’ s character truthfully or merely apparently truthfully. Evidently, ē thos, pathos and logos can be used for good but also for bad or deceitful ends. This is the case in spite of the fact that grounds for conviction such as these are most cognate and effective aids for human beings capable of communicating; thus they bring to the fore via logoi whatever is the case or what ought to be thought, felt or done. 3 The areas of way of thinking had been extinguished. He who has at his command, along with clear insight into the facts, language in all its richness and purity, and who, along with a fruitful imagination capable of presenting his ideas, feels a lively sympathy for the true good, is the vir bonus dicendi peritus, the speaker without art but full of vigor, as Cicero would have him, though he did not himself always remain true to this ideal ” (Kant 2001, 205 (= 5: 327 f ). However, this is not Kant ’ s last word on the meaning and function of rhetoric since certain rhetorical practices in education and public discourse are capable of becoming vital means to move people to moral improvement without infringing on their personal autonomy (see Stroud 2014). 3 The basic political implication of the fact that human beings are living beings who possess logos is summarised by Aristotle: [. . .] and no animal has speech except a human being [. . .]. But speech is for making clear what is beneficial or harmful, and hence also what is just or unjust. For it is peculiar to human beings, in comparison to the other animals, that they alone have perception of what is good or bad, just or unjust, and the rest. And it is community in these that makes a household and a city-state ” (Ar. Politics 1253 a 10 - 17; tr. Reeve). 144 Markus H. Wörner (Galway/ Ireland) merely apparent arguments in particular, or of arguments dealing with whatever is merely apparent, frequently seem to involve unfair, deceitful or fallacious communication. However, instead of dismissing these areas in their entirety as ‘ merely ’ apparent and hence as fallacious per se, to be avoided by respectable speakers, it also becomes apparent in rhetorical practice that they may provide audiences with the best possible arguments under the specific circumstances of a particular speech situation, without being deceitful or downright fallacious. They may even be of vital importance in practical life. Consequently, criteria are needed which allow us to draw a distinction between reasonable arguments and deceptive or fallacious ones, between constructive and sham rhetoric. A closer look at the logic and pragmatics of persuasion is required to prepare the ground for deciding what the border lines between rhetoric proper and sham rhetoric are. Aristotle ’ s Rhetoric and recent research dealing with his theory of so-called “ fallacious arguments ” , particularly his remarks on merely apparent rhetorical syllogisms/ enthymemes, provide a rich starting point for approaching this intricate set of problems. These remarks concern the core, the “ body of proof ” (s ō ma t ē s piste ō s) of the rhetorical method, the enthymeme. 4 They highlight a set of problems and distinctions affecting our understanding of the rationality of rhetoric and human rationality in general. We want to know what rhetorically configured arguments we can trust and which are faulty, even deceitful. Roughly speaking, three tendencies in the present controversy concerning Aristotle ’ s distinction between good or “ real ” and so-called “ fallacious ” or “ deceptive ” rhetorical arguments and their core Gestalt, the enthymeme, can be identified. 5 1. On the one hand there is the deductionist view that a real rhetorical syllogismos 6 or enthymeme is a deduction that involves logical necessity of the consequence (necessitas consequentiae) and which also has a valid epistemic status, provided that its premises are valid. In contrast, an enthymeme which is not a deduction is merely apparently so because it is a-syllogistic. It is fallacious in this respect. Hence, an enthymematic argument which involves a for-the-most-part validity of the consequence (probabilitas consequentiae, not consequentis), is deceitful. Logical necessity of the rhetorical syllogismos separates real from apparent enthymemes, good rhetoric from potentially sham rhetoric. 1.1. The deductionist view is supported by a related position which maintains that real enthymemes and apparent ones can clearly be distinguished, since they presuppose two mutually exclusive sets of topoi used to generate them, listed and discussed by Aristotle mainly in Rhet. B 23/ 24. The latter set (B 24), it is suggested, tends to generate fallacious and deceitful apparent enthymemes, whereas the former (B 23) is capable of generating real enthymemes. 4 Ar. Rhet. 1354 a 14 f. A discussion of rhetorical induction or paradeigma (epagog ē rhetorike/ paradeigma), the second core Gestalt of rhetorical arguments, will follow in a different paper. 5 Paradigmatic for the deductionist view (in spite of major differences of interpretation in detail) see Sprute 1982; Rapp 2002; Rapp 2013, 122 - 158; for the more relaxed view see Ryan 1984; Wörner 1990; Manetti 1993, 70 - 91; Burnyeat 1996, 88 - 115; Allan 2001. 6 I am using “ syllogismos ” instead of “ syllogism ” in order to indicate that I am using the term in the wider sense of the Top., SE and Rhet., not in the technical sense of An.pr.. Whenever the categorical syllogism of An.pr. is meant, “ syllogism ” is used. How Deceptive are Rhetorical Arguments from Signs or Likelihoods? 145 2. On the other hand, a more relaxed view maintains on the whole that enthymemes involving necessity of consequence and for-the-most-part consequence may count as real rhetorical syllogismoi. Alternatively, enthymemes with a for-the-most-part consequence may be called “ apparent ” since they are a-syllogistic (in a narrow sense of syllogismos) but may nonetheless be informed, reasonable and useful arguments ( “ syllogismoi ” in a wider sense), coming closest - as far as their logical status is concerned, to the case involving necessitas consequentiae. They are not “ merely ” apparent and deceitful nor fallacious per se. Hence, logical necessity does not form a strict dividing line between real and apparent enthymemes, between real and potentially sham rhetoric, as long as these enthymemes represent informed, reasonable and useful arguments. 2.1. The more relaxed view is supported by a position claiming that the apparently straightforward distinction between distinct sets of topoi, one supposedly generating real enthymemes (good rhetoric), and the other one generating deceitful or fallacious apparent enthymemes (sham rhetoric), cannot be upheld in all cases and under all circumstances of uses of topoi mentioned in B 23/ 24. This becomes particularly apparent - and confusing - in the analysis of examples of enthymemes provided by Aristotle himself. 3. Finally, there exists a minority view contending that almost no rhetorical argument is logically valid. 7 Parts 2 - 7 of my paper will critically discuss the evidence for positions 1 and 2, reaching the conclusion that dividing lines between genuine and apparent syllogismoi such as logical necessity (parts 3/ 4) or distinctly different sets of topoi (part 5) do not effectively distinguish informed, reasonable and useful enthymemes from those which are fallacious or deceitful, good rhetoric from sham rhetoric. I shall claim that only a small number of the broad spectrum of informed, reasonable and useful enthymemes have been taken into account systematically in this debate. We need to widen the field of consideration by looking more closely at enthymematic arguments with different epistemic and pragmatic status, and giving deductions their proper place within this wider spectrum of ways in which premises and conclusions are reasonably connected. Taking arguments from likelihoods and signs as paradigmatic examples, I shall make an alternative suggestion for classifying enthymematic arguments without disclaiming all apparent enthymemes as fallacious and deceitful per se, and this suggestion will exhibit their strengths as well as their different grades of reasonableness. Hence, part 8 of this paper makes a proposal ad mentem Aristotelis to cover these cases. Consequently, the borderlines between good (informed, reasonable and useful) and sham (fallacious and deceitful) rhetoric may have to be drawn in a more differentiated way than is frequently assumed. 7 See Garver 1994, 163: “ Considered purely logically, all rhetorical arguments are invalid; among invalid arguments it is hard to find a further distinction between real and apparent enthymemes ” ; Rapp 2013, 156 n. 5. 146 Markus H. Wörner (Galway/ Ireland) 2. Is rhetoric similar to eristic sophistry and hence not to be trusted? Recent research concerning Aristotle ’ s Rhetoric and comparisons with the Topics (Top.) or Sophistical Refutations (SE) focuses on the problem of the nature of enthymemes and “ fallacious rhetorical arguments ” , of “ merely apparent ” rhetorical syllogismoi or syllogismoi demonstrating the “ merely apparent ” (phainomenoi syllogismoi/ phainomena enthym ē mata) (Rapp 2013, 122 - 158). Dialectic as discussed in Top. promises to provide a clear distinction between real and fallacious syllogismoi in general, while SE claims to list all possible types of fallacies used in sophistic and in eristic reasoning. 8 The Rhetoric seems at first sight to use the same distinctions. When introducing the enthymeme in Rhet. A 2 (1356 a 36 - b 25), Aristotle maintains that it is a rhetorical syllogismos and that an apparent enthymeme is an apparent syllogismos. 9 He relies on the reader ’ s knowledge of dialectic to know the distinction between induction, real and apparent syllogismoi, saying that this is similar in rhetoric. 10 Real enthymemes are syllogismoi proper while apparent enthymemes are asyllogistic and refutable. Strictly speaking, the latter appear not to be enthymemes at all. 11 However, he does not say that this distinction is the same as in dialectic. He also does not say explicitly - he will imply it later on in B 24 - that apparent enthymemes may start from apparent reputable opinions (phainomena endoxa) or from apparent likelihoods (phainomena eikota), 12 which might otherwise be expected from the description of types of sophistic or eristic syllogismoi in SE if apparent enthymemes were simply classified in the same way. Still, rhetoric itself “ is a combination of the science of logic and of the ethical branch of politics; and it is partly like dialectic, partly like sophistical reasoning ” (Rhet. 1359 b 9 - 12 tr. Roberts). Commonalities between Aristotle ’ s treatment of apparent enthymemes in the Rhetoric and apparent syllogismoi in SE give rise to the problem to what extent and in what sense the art of rhetoric is “ like sophistical reasoning ” , or eristic, when it comes to allegedly creating the appearance of syllogistic arguments rather than the genuine article. A sophist, Aristotle claims, is a master of apparent-but-not-real arguments. If he engages in eristics using eristic syllogismoi, he either argues from premises which have the character of reputable opinions that are not what they appear to be (sometimes they may even be merely apparent scientific premises); they are not genuinely reputable opinions or genuine principles of a science. In this first type of eristic syllogismoi, the logical status of the 8 On the differences between dialectic, sophistic and eristic reasoning see SE 171 b 3 - 172 b 8. Schreiber 2003, 1 - 3 provides a brief summary of major differences. 9 Kassel ’ s revised text of Rhet. 1356 b 3 f brackets the remark that an apparent enthymeme is an apparent syllogismos as a later addition; see Kassel 1976, 11. The addition makes sense due to the distinction in dialectic to which Aristotle refers. 10 A “ merely apparent induction ” or a “ merely apparent paradeigma ” are not mentioned as a fourth possibility. 11 This appears to be confirmed by Rhet. 1397 a 3 f and 1400 b 34 - 37. If it were indeed the case that all apparent enthymemes are not enthymemes at all, an enthymeme explicitly used by Aristotle as a paradigmatic case would not be an enthymeme: “ No man who is sensible ought to have his children taught to be excessively clever, for, quite apart from becoming idle, they earn jealous hostility from the citizens ” (Rhet. 1393 b 29 f [Euripides, Medea 294 - 297]); see Burnyeat 1996, 100 f. 12 The expression phainomena endoxa does not occur in the Rhetoric, whereas phainomenon eikos does in respect of rhetorical or eristic ways of making the weaker argument stronger (B 24, 1402 a 26). However, when discussing examples of apparent enthymemes in B 24, Aristotle mentions reputable opinions as premises which may be regarded as apparent. See also King 2013, 185 - 199. How Deceptive are Rhetorical Arguments from Signs or Likelihoods? 147 argument seems to be that of logical necessity, which Aristotle mentions shortly before in the definition of a dialectical syllogismos (Top. 100 a 25 ff ), whereas the epistemic status of its premises is not that of truly reputable opinions or truly scientific statements. He adds that this type of eristic syllogismos should still be called syllogismos because it “ reasons ” (syllogizetai). Alternatively, an eristic speaker appears to produce logically valid arguments from real or apparent reputable opinions when, in fact, he does not (Top. 100 b 23 - 101 a 4). Two further types of eristic syllogismoi are referred to here. In the first case, the logical status of the argument is not that of logical necessity but is based on the appearance of it, while the epistemic status of its premises is that of reputable opinions. The second case is a syllogismos whose logical status is equally not that of logical necessity but based on the mere appearance of it, while the epistemic status of its premises is not that of reputable opinions. Aristotle adds that the latter type of syllogismos should still retain its name “ eristic ” syllogismos; nonetheless he conveys that it should no longer be termed syllogismos in the strict sense, because it appears to reason but does not (phainetai men syllogizesthai, syllogizetai de ou). Finally, eristic procedure may produce valid arguments that appear to be, but are not, relevant to the topic under discussion (SE 169 b 22 f ). When an argument is refuted, an eristic speaker does not provide an argument as a dialectician does, namely by using a real refutation, i. e. a syllogismos whose conclusion is the contradictory of an opponent ’ s thesis (cf. SE 165 a 2 f; 168 a 36 f; cf. An.pr. 66 b 11)); he merely gives the impression that he does (Bolton 2012, 270 - 295). Must the same be said of different types of enthymemes a rhetorician may use when he speaks in the assembly (genos symbouleutikon) or at court (genos dikanikon) in order to defend his case, advising, dissuading, accusing or defending against possible opponents or by praising or blaming publicly-known people on festive occasions of the polis (genos epideiktikon), possibly counting on the ignorance of an audience incapable of detecting nuances in argumentation? If something is said ambiguously, for instance, an audience which is not experienced in detecting nuances (mikra synhor ā n dynamenoi) may be incapable of drawing distinctions which would resolve the confusion (SE 169 a 22 - 25). Failing to make these distinctions or being offered paraphrases of arguments that make more sense than what was said originally may prevent listeners from unravelling inherent confusion, inconsistency, vagueness or falsity. Kant may have been right - in this case. But there are more issues that need to be taken into account. It is obviously the case that not everything which appears to be endoxical is truly endoxical, as Aristotle points out (Top. 100 b 26); yet some of what merely appears to be endoxical may nonetheless be (or turn out to be) truly endoxical. 13 The fact that premises are apparent reputable opinions may not diminish their epistemic status tout court. Might something similar be said of the logical status of apparent syllogizing? Rather than denying any possibility of qualifying in what sense apparent syllogizing might still be seen as reasonable (a denial which frequently seems to be involved in claiming that an apparent syllogismos is a-syllogistic and hence 13 An apparently endoxical opinion (prejudice) assuming, for instance, that women are less reliable workers in a market economy, is not truly endoxical but “ merely ” apparent in stating what is supposedly the case for the most part. Conversely, a truly reputable opinion that, for instance, those who love are loved in return, may also be apparent, but not “ merely ” apparent. 148 Markus H. Wörner (Galway/ Ireland) fallacious), there will remain a possibility that its status can still be clarified positively. An affirmative answer to this problem would show that the realm of good, reasonable and useful rhetorical enthymematic arguments and, hence, the range of a type of rhetoric which is good, informed, reasonable and useful is considerably wider than originally expected, provided that it can be distinguished sufficiently from intrinsically fallacious or deceitful rhetoric. Gadamer may be right after all. But we must proceed with caution. There exist textual reasons for the view that whatever is true of the epistemic and logical status marking the difference between an eristic syllogismos and a syllogismos fulfilling its definition as outlined at the beginning of Top. (100 a 25 - 101 a 4), is also true of the difference between real and apparent rhetorical syllogismoi. These textual reasons seem, at least initially, to make it appear less likely that we shall be able to produce an affirmative answer that “ softens ” the logical status of many, if not the majority of good, reasonable and useful enthymemes. Despite this, a further reason for caution is based on the fact that major textual divisions in SE dealing with modes (tropoi) of sophistic syllogismoi, also found in the way the material of topoi for apparent enthymemes and rhetorical refutations is organised in Rhet. B 24/ 5, seem to form a further objection to a stance which potentially softens the logical status of good, reasonable and useful enthymemes: 14 Most types of tropoi for fallacious arguments mentioned in the list of SE find their analogues in the list of topoi in Rhet. B 24. 15 Obviously, lines of argument closely resembling each other are found in Rhet. and SE, although they are mostly explained by using different examples that come partly from different (literary, oratorical [rhetoric], philosophical or proto-scientific [dialectic]) backgrounds. 16 Yet the 14 Rhet. B 24, for instance, provides topoi and examples of apparent enthymemes; B 25 tells us how to refute them. Similarly, going through the same material twice in SE IV - VII with differing emphasis and detail of analysis, Aristotle presents tropoi for generating the appearance of syllogismoi (IV/ V), followed by explanations showing how to dissolve them (VI/ VII). Tropoi or modes for apparent syllogismoi in SE are divided into those that depend on the particular language used (para t ē n lexin; fallacia dictions) and those outside the particular language (ex ō t ē s lexe ō s; fallacia extra dictionem). Lexis, here, means not so much written as spoken language. Equally, Rhet. B 24 divides topoi generating apparent enthymemes into those that depend on the particular language used and those which do not. 15 These are: form of expression (fallacia figurae dictionis) - compare Rhet. 1401 a 3 - 12 with SE 174 b 8 - 11 - , use of similar words for different things (fallacia aequivocationis/ homonymy) - comp. Rhet. 1401 a 13 - 25, SE 165 b 30 - 166 a 6 - , asserting of the whole what is true of the parts (fallacia compositionis) - comp. Rhet. 1401 a 25 - 01 b 3, S. E. 166 a 23 - 32 - or of the parts what is true of the whole (fallacia divisionis) - comp. Rhet. 1401 a 25 - 01 b 3, SE 166 a 33 - 38. This first group generates fallacies or apparent enthymemes depending on the particular language used. The topoi mentioned in the Rhetoric as generating apparent enthymemes outside diction are: argument from sign (fallacia signi), mentioned only briefly in SE in the context of fallacies from consequences - comp. Rhet. 1401 b 9 - 14, SE 167 b 8ff - , representing the accidental as essential (fallacia accidentis) - comp. Rhet. 1401 b 14 - 19, SE 166 b 28 - 36 - , argument from consequent (fallacia consequentis) - comp. Rhet. 1401 b 20 - 29, SE 167 b 1 - 20 - , representing as causes things which are not causes (fallacia propter non causam ut causam) - comp. Rhet. 1401 b 29 - 34 with SE 167 b 21 - 36 - , leaving out any mention of time and circumstances (fallacia ellipsis) - comp. Rhet. 1401 b 34 - 02 a 3, SE 166 a 3 f: leaving out any mention of time is also implied in ‘ The same person sits and stands ’ - , confusion of the absolute with that which is not absolute but particular (fallacia secundum quid et simpliciter) - comp. Rhet. 1402 a 3 - 28 with SE 166 b 37 - 67 a 20. 16 Nonetheless, analogues to some of the types of fallacies mentioned in SE are not mentioned in the Rhetoric, for instance accent, begging the question, making one question out of many, whereas an equivalent of one of the topoi in B 24 cannot be found in SE. Exaggerating a case in order to influence the audience so that it accepts a conclusion, is mentioned in Rhet. B 24, but not in SE. How Deceptive are Rhetorical Arguments from Signs or Likelihoods? 149 mere fact that differences can be found in the treatment of sources of appearance mentioned in SE and in the Rhetoric will not calm the fears of those who are generally in favour of a positive assessment of the art of rhetoric as far as its capacity to provide a method for finding and exposing truth or verisimilitude in a convincing manner is concerned. 3. Is logical necessity an unambiguous dividing line between real/ acceptable and merely apparent (most likely fallacious/ deceitful) enthymemes? Any approach to solving the puzzle of how to distinguish good from sham rhetoric by distinguishing real rhetorical syllogismoi from merely apparent ones in terms of the presence or lack of necessitas consequentiae will have to start from an acceptable understanding of what, in Aristotle ’ s view (and possibly ours), may rightfully count as a “ real ” enthymeme. In contrast to what he states categorically in other places in the Rhetoric (1397 a 3 f, 1400 b 34 - 37), the definition of the enthymeme he gives us in this text which is disputed in contemporary research 17 appears to leave room for enthymemes which involve necessitatem consequentiae as well as probabilitatem consequentiae. It appears to embrace not only real syllogismoi, but also some enthymemes which would otherwise be counted as asyllogistic and hence merely apparent and, for some commentators, fallacious or deceitful: [. . .]when, certain things being the case, something different results in addition to them by virtue of their being the case, either universally or for the most part, there it is called syllogismos, here enthymeme. (Rhet. 1356 b 16ff; tr. Burnyeat) An enthymeme is characterised as a kind of oratorically presented “ going together ” (symbainein). In contradistinction to the definitions of syllogismos in Top., SE and An.pr., it is not termed a logos in which something applies by necessity to something else, 18 although this characterisation is not excluded. According to the definition, the rhetorical syllogism does not start from something posited (thetent ō n tin ō n), like a thesis in a dialectical disputation, but from some (plural) existing states of affairs (tin ō n ont ō n). Real states of affairs and their interconnection, rather than a connection between a thesis assumed for the sake of argument and its conclusion, are emphasised. The mode of “ going together ” is determined by virtue of these states of affairs (t ō tauta einai). These on their own are sufficient in order to go together with a different state of affairs (heteron ti t ō n keimen ō n). The formula “ either universally or for the most part ” ( ē katholou ē h ō s epi to poly) used, here, in which “ universally ” is equivalent with “ in all cases and at all times ” , replaces the less complex “ by necessity ” formula (ex anagk ē s) of the other three definitions. The formulation 17 For a critical summary of this dispute see Rapp 2002, 161 - 167, 191 - 198. 18 Compare the following statements with Rhet. 1356 b 16ff: (1) “ A syllogism is an argument/ account in which, certain things being posited, something else, different from the things which are posited, comes together [with them] by necessity because of them (An.pr. 24 b 18 - 20; my translation, M.W.) ” ; (2) “ Now reasoning is an argument in which, certain things being laid down, something other than these necessarily comes about through them ” (Top. 100 a 25 - 27; tr. Pickard-Cambridge); (3) “ Reasoning rests on certain statements such that they involve necessarily the assertion of something other than what has been stated, through what has been stated ” (SE 164 b 27; tr. Pickard-Cambridge). 150 Markus H. Wörner (Galway/ Ireland) “ for the most part ” does not occur in any of the others. 19 In the latter cases the formula “ by necessity ” qualifies the consequentia (N[(P1 ᴧ P2) → C]), not the modality of the consequens ([P1 ᴧ P2] → N[C]). Hence, the modal operator does not specify the conclusio but the inference itself. Since the formula is used in these cases to refer to the inference itself rather than to a conclusion allegedly stating a necessary fact, it seems more than reasonable to suppose that the same occurs in the case of the definition in the Rhetoric, unless it suddenly and in unannounced fashion were to shift its function from an indicator of a mode of inference to an indicator of the modality of a conclusio. Provided that this shift does not occur, and Aristotle replaces the “ by necessity ” formulation of definitions of syllogismos of the rest of the Corpus Aristotelicum deliberately when it comes to rhetorical argumentation, then he appears to allow for rhetorical syllogismoi a kind of “ coming together ” of premises and conclusion that may apply for the most part as well as by necessity. 20 Moreover, if the definition of the Rhetoric stresses states of affairs and their connectedness rather than the specific argumentative move as such, then it points to the fact that the connection of symbainein between the states of affairs referred to in the premises and the different state of affairs referred to in the conclusion is a fact rather than merely a logical connection between propositions. As a connection between states of affairs it is a state of affairs in itself, not merely an inferential sequence of propositions. There is no compulsion by which this kind of state of affairs, one which joins other states of affairs, should exist necessarily only in order to be presented in a reasonable rhetorical argument, nor must it exist for the most part only. It may even exist rarely (ep ’ elatton/ forthe-most-part not) or else with a likelihood of something that is equally likely to exist or not to exist (possibilitas ad utrumlibet). However, if such a connection exists by necessity or - what comes closest to it - for the most part, i. e. almost always or in almost all cases, it is most likely that an audience understands that it does exist and that it is most reasonable to think that it does. This may apply notwithstanding the fact that some of the states of affairs mentioned in the premises may be of the “ necessary ” type ( “ Human beings are mortal ” ) or that some may be likelihoods (eikota) ( “ Thieves slink about in the hours of darkness ” ), including ‘ for-the-most-part ’ cases ( “ Those who love are loved ” ). Some things conveyed in the premises may also be signs, either necessarily or likely linked to something different, signified in the conclusion, to which they refer. Audiences at court or in the assembly, when listening to enthymemes, understand that such connections between states of affairs brought to their attention are not only more or less pellucid but they themselves may not have thought of them before; such audiences do not confine themselves simply to admiring logical skills of an orator presenting connections that are evident to them as soon as they 19 A strikingly similar formula is used in Aristotle ’ s description of the sequence of events (plot structure) on which a poet should focus (Poetics 1451 a 36 ff ); see Bittner 1992 , 101; Wörner 2003, 41 - 69. 20 Why not call rhetorical syllogisms which do not involve logical necessity “ rhetorical inductions ” instead and continue to discuss enthymemes and (rhetorical) inductions separately? Given Aristotle ’ s verdict that there are only two ways of showing something to be used in rhetorical speeches, namely either rhetorical syllogismoi or rhetorical inductions (paradeigmata) (Rhet. 1356 b 5 - 7), this new treatment of “ for-the-mostpart ” enthymemes as kinds of induction seems to be adopted by recent debates concerning probabilistic inferences (see Hempel 1965, 394 ff ). However, this might seemingly solve a Gordian knot but is clearly at variance with Aristotle ’ s definition of the enthymeme itself as well as his own understanding of rhetorical induction; see Burnyeat 1996, 97 f. How Deceptive are Rhetorical Arguments from Signs or Likelihoods? 151 understand that the conclusion is necessarily entailed in the premises. Arguments revealing the connectedness of states of affairs evoke corresponding thoughts in the audience. It is these thoughts and their adequacy which carry the force of conviction. 4. The test: enthymemes from sign or likelihood Critical tests for the validity of the “ necessary ” or the “ universally or for-the-most-part ” formulae as demarcation lines between real and apparent enthymemes are provided by syllogismoi from signs (s ē meia) or likelihoods (eikota). 21 Enthymemes based on signs form a substantial group among real as well as apparent enthymemes. According to An.pr., a sign is “ something such that, when it exists, another thing exists, or, when it has happened, the other has happened before or after ” (An.pr. 70 a 7 ff ). 22 Signs used in syllogismoi also have a close relation to likelihoods even though they are not identical with them, since “ a sign is meant to be a demonstrative proposition either necessary or reputable ” (protasis apodeiktik ē ē anagkaia ē endoxos) and a likelihood is a reputable proposition: “ What men know to happen or not to happen, to be or not to be, for the most part thus and thus, is a likelihood, e. g. envious men hate, those who are loved show affection ” (An.pr. 70 a 4 - 6). Enthymemes from necessary, infallible signs (tekm ē ria; s ē meia in a narrow rather than a generic sense), seem to be the least controversial ones. Arguments such as these, for instance “ She has milk; hence she has given birth, ” provide real syllogismoi which Aristotle sees as irrefutable if the premises are true. From a logical and epistemological point of view they appear to be the best possible cases in terms of which all the remaining enthymemes should be judged. Enthymemes from non-necessary, fallible signs (s ē meia anonyma), 23 for instance “ She is pale; hence she is pregnant ” or “ The wise are just, because Socrates was wise and just, ” in spite of being initially introduced (in Rhet. A 2) in a way that raises no prima facie suspicion that they might not be real enthymemes, appear to provide us with merely apparent enthymemes. They seem to represent a-syllogistic and refutable arguments. 24 On closer inspection, however, tekm ē rion-enthymemes and s ē meion-enthymemes do not form mutually exclusive sets, since some enthymemes from non-necessary signs (or from likelihoods) are also capable of representing logically valid enthymemes. In relation to their logical structure, some of them can be constructed in the same way as can enthymemes from necessary signs, although others cannot. There is no reason for strict deductionists to reject those among the ranks of real enthymemes (and good rhetoric) which are capable of involving logical necessity and impeccable epistemic status. Some of those which cannot be constructed in the same way as enthymemes from necessary signs, may nonetheless be capable of being constructed analogously to logically valid syllogismoi, whereas others may 21 Wörner 1982,73 - 98; Weidemann 1989, 343ff; Manetti 1993,70 - 91; Allan 2001, 13 - 79. 22 Aristotle does not mention signs which signify the absence of a fact (the fact that the house was locked as a sign for a thief that its inhabitants are not at home), whereas the Rhetorica ad Alexandrum does: “ That which has happened is a sign not only of what has happened but also of what has not happened; and similarly of what has not happened is a sign not only of what does not exist but also of what does exist ” (Rhet. ad Alex. 12). Aristotle ’ s notion of sign appears to be narrower than the one used by Anaximenes. 23 The term anonymous is used in Rhet. 1357 b 4 f for the fallible kind of sign. 24 Cf. An.pr. 70 a 28 - 38. Here, Aristotle draws a distinction between syllogistic and a-syllogistic enthymemes but does not call the latter “ merely apparent ” . 152 Markus H. Wörner (Galway/ Ireland) not. Examples discussed in the following sections will explain why this does not make them intrinsically fallacious, deceitful or uninformed, irrational and useless. Nonetheless, it may make them vulnerable in the sense that they are increasingly capable of being abused by deceitful speakers in speeches for audiences who are unable to distinguish the argumentative nuances involved in them. An example of an enthymeme involving both a fallible sign and a likelihood, generally deemed merely apparent, shows that these types of argument may involve logical necessity that is comparable with an argument from necessary sign. Logical necessity becomes apparent when the argument is constructed analogously to a first-figure syllogism. The example, taken from the discussion of enthymemes from An.pr. B 27, might have also occurred in the Rhetoric, since it is similar to others that the Rhetoric contains, even though it is not mentioned there: 25 (a) Pittakos is generous, for those who are ambitious, are also generous, and Pittakos is ambitious (An.pr. 70 a 26 f). 26 Obviously, the fact that Pittakos is ambitious is a fallible sign for the likely fact that he is generous. The relation between sign (Pittakos (possibly) being generous) is that of a particular to a particular, mediated by a reputable opinion represented in the maior ( “ Being generous applies [for the most part] to [all] those who are ambitious ” ). The enthymeme can be constructed analogously to (but not identically with) modus barbara. If this is the case, then one can also construct logically correct syllogisms using fallible signs. Hence, enthymemes from fallible signs do not generate logically incorrect inferences per se. However, they are as fallible or as likely to be valid as the reputable opinions which are used in the maior, provided that the observations are true on which the rest of the enthymeme is based. If the minor is undisputed, which in this case is a matter of observational evidence or veridical experience, then it still remains possible to refute the argument by attacking the maior in the same way as likelihoods are attacked. In order to produce a real refutation, one has to show that the likelihood mentioned is not likely at all. Making an objection by mentioning a particular case (or a few cases) where the likelihood does not apply ( “ Alkibiades is also ambitious but not generous ” ) generates a merely apparent refutation, for it only shows that the argument is not inevitably true (cf. Rhet. 1402 b 20 - 26). It would amount to a paralogism (paralogismos) if the argument were taken as a genuine refutation. 25 The fact that some enthymemes in the Rhetoric can be constructed analogously to syllogisms of An.pr. does not entail that Aristotle is, in fact, using the Analytics and its technical terminology in the Rhetoric. There are five references to the Analytics in the Rhetoric, but the entire technical apparatus of the Analytics (major term, minor term, middle term, universal/ particular affirmative/ negative quantifiers, figures and moods of syllogisms, types of conversion, ekthesis or reduction to first-figure syllogisms, let alone modal/ modally mixed syllogisms) is not mentioned at all. The Analytics, in this paper, are used as a diagnostic tool only rather than as an indication of how Aristotle analysed arguments himself at the time when he wrote or corrected the Rhetoric. 26 (a ’ ) “ If being generous (A) applies to [all]those who are ambitious (B), and if being ambitious (B) applies to Pittakos (C), then, necessarily, being generous (A) applies to Pittakos (C). ” Of course, modus barbara uses three universal affirmative propositions, not propositions with proper names. The latter are characteristically used in rhetorical communication, especially in practical contexts where those involved need to deal with particulars and need to come to a judgment about them. How Deceptive are Rhetorical Arguments from Signs or Likelihoods? 153 The example mentioned here and its refutation show that logical necessity is neither a necessary nor a sufficient condition for distinguishing, on the one hand, all seemingly apparent enthymemes from fallible signs from, on the other, real enthymemes from signs. Furthermore, logical necessity in a rhetorical argument does not automatically make it reasonable; nor do apparent enthymemes from fallible signs always represent unreasonable arguments. Logical necessity hardly provides an unambiguous dividing line between good and sham rhetoric. If it were indeed the case and might also mark a clear-cut division between rhetoric proper and sham rhetoric, Aristotle could be accused of presenting a theoretical framework for assessing features of arguments in concrete speeches which was barely capable of functioning as a good enough guide for saving the phenomena of reasonable discourse. That framework would excommunicate far too many everyday arguments as unreasonable or fallacious, many of which are in fact reasonably acceptable and practically required in their particular circumstances. 27 5. Do different classes of topoi form a dividing line between good and bad rhetorical arguments, good and sham rhetoric? When discussing general topoi for enthymemes, Aristotle appears to present us with two different lists, one for allegedly generating real enthymemes (Rhet. B 23) and one for apparent enthymemes (Rhet. B 24), before discussing their modes of refutation (Rhet. B 25). Hence, enthymemes based on topoi from list B 23 seem “ real ” or valid, whereas enthymemes based on topoi from list B 24 seem to be only apparent and fallacious. They reveal a close similarity to the list of tropoi in SE. The lists present two sets of argumentative rules, conventions of thought or rules of the pragmatics of verbal interaction which Aristotle himself seems to have left as a basis for enlargement and further research rather than asserting them to be a coherent and complete set. No claim to completeness, logical consistency or mutual independence of these topoi is made. There is no sign of an inherent system or systematic classification 28 such as the four predicables allow for most dialectical topoi. Instead, the lists are collections of examples taken from the Greek literary, forensic, oratorical, sophistical or historical traditions. 29 Nonetheless, they provide the practising rhetorician with a useful guide for looking for premises that may be used for a conclusion that he intends to draw (their inventive 27 Many commentators in contemporary Rhetoric and Argumentation Theory, considering the pragmatics of argumentation rather than mere logical form, are convinced that the analysis of real arguments concerning human action in law, politics, medicine or everyday life in general, shows that deductive arguments, let alone syllogisms in the strict sense, are far from omnipresent. Moreover, real-life arguments can hardly be suitably reconstructed in terms of purely logical form without losing sight of most of those of their convincing features that secure reasonableness. (See for instance Toulmin 1969, Perelman 1971, Walton 1990, 1996, 2010; van Eemeren 1993, 2002; Edmondson et al. 2009, Edmondson 2012, Hess-Lüttich & Warnke 2017.) 28 However, the list of topoi in B 24 corresponds roughly to the classification of tropoi for sophistical arguments in SE, but neither does it make the same claim of systematic completeness nor does it mention all tropoi of SE. The fundamental subdivision between linguistic and non-linguistic fallacies of SE is also adopted, but handled more loosely. Moreover, predicables involved, here, do not cover all argument schemes in the Topics (see Kakkuri-Knuuttila 2012, 55 - 81). 29 See Rapp 2002, 751 - 776 for authors of examples for topoi in B 23 and Rapp 2013, 138 for a table of authors of examples for topoi in B 24. 154 Markus H. Wörner (Galway/ Ireland) function). They also provide a hermeneutic tool for members of a sufficiently experienced audience to detect what means have been used to try to convince them and to show them where they might also find starting-points for objections, counter-syllogisms or refutations. The fact that both speaker and hearer make use of these rules or conventions, which form a crucial part of their - tacitly presupposed - joint culture of communicative know-how in verbal interaction, allows them to perceive what they say or hear as more or less reasonably argued. Without these meta-communicative rules or conventions of argumentation, the connections between states of affairs implied or presupposed in whatever was said may strike them as incoherent and unreasonable. However, compared with the collection of topoi in the Topics, their number in the Rhetoric is surprisingly small. Only a minority (topoi 7 and 9) of the 28 topoi for real enthymemes in B 23 can be termed candidates for deductions insofar as they generate enthymemes involving logical necessity. Another small group appears to come close to dialectical topoi whose syllogismoi may involve logical necessity (1, 2, 4, 12); 30 yet the examples provided by Aristotle show that this is not necessarily the case for every corresponding enthymeme. Another topos (10) may be used for genuine as well as apparent enthymemes including (rhetorical) induction, whereas one topos (25) may straightforwardly generate error (pseudos), as Aristotle says himself. In spite of this, the topos is not a member of the B 24 list. Furthermore, topos 15, recommended as a sophistical strategy in SE (172 b 36 - 173 a 6), represents “ the most sophisticated of the forms of argument that contradict reputable opinions (paradoxa) ” (Rhet. 1399 a 30 - 34). It may find its suitable place in SE, since creating paradoxa is one of the two main tasks of a sophist, but mostly not that of a rhetorician (cf. SE 172 b 10 f). Finally, some topoi on this list are not topoi for enthymematic deductions; instead, they represent strategic rules for the pragmatics of verbal interaction (topoi 6, 22 for instance). The majority of topoi mentioned in B 23 are not based on laws of logic at all, nor do they require them. Instead, they rest on collective experience, shared by speaker and hearer, which is sedimented in their opinions about how states of affairs are connected for the most part or with reasonable likelihood. They also rest on conceptual connections between correlated terms (good/ bad, giving/ receiving and so on), modifications of key-words (such as just or beneficial) or various senses of a word (such as rightly). Hence it is not clear why all members of the set of topoi of B 23 should generate real enthymemes. Some of them may, some of them may not. As far as list B 24 is concerned, one topos of the nine mentioned may or may not be used as a reasonable and useful apparent enthymeme, in spite of the fact that it is called a-syllogistic (topos 4), while topoi 2 and 3 are lines of argument explicitly declared fallacious (paralogistikoi); some may be candidates for apparent and also deceptive arguments because they direct the arguers to present wrong reasons for what they claim, although the 30 However, not all the topoi in the Topics yield deductive arguments/ inferences, but some yield probable inferences. Here are some examples: “ Things the generations of which are good/ bad, are themselves good/ bad. ” “ The generations of good/ bad things are good/ bad. ” (See Top. II 9, 114 b 18 - 22; III 2, 117 b 4 - 9; III 6, 119 b 8 - 13; IV 4, 124 a 20 - 30; VII 1, 152 a 1 - 4.) “ If a predicate belongs to a subject, then it also belongs to another subject to which it is more likely to belong. ” “ If a predicate does not belong to a subject, then it does not belong to another subject to which it is less likely to belong. ” (See Top. II 10, 115 a 7 - 11; III 6 119 b 17 - 30; IV 6, 127 b 26 - 27; V 8, 137 b 15 - 27; VI 7, 146 a 4 - 11.) Compare Green-Pedersen 1984, 27; Kakkuri-Knuuttila 2005, 31 - 74, 47 - 48. How Deceptive are Rhetorical Arguments from Signs or Likelihoods? 155 reasons given are nonetheless related to underlying facts (topoi 3, 5, 9 for instance). Finally, topos 1 merely creates the impression of an inference where there is none. In this respect, it is like an eristic syllogismos which only seems to deduce but does not really deduce at all. 31 This - admittedly brief - analysis of B 23/ 24 32 may suffice to show that the seemingly clear-cut division between lists of topoi believed to generate real enthymemes and those generating apparent enthymemes cannot be upheld as mutually exclusive, since there are ambiguities of categorization, possibly even misclassifications at work in both lists. Some topoi in B 23, insofar as they lead to error or paradoxical conclusions, may equally appear in B 24, and at least one topos in B 24, insofar as it leads in all likelihood to truth or verisimilitude in practice, may be placed in B 23. Furthermore, both groups contain at least some topoi which, given the appropriate circumstances, may contribute reasonably to an ongoing critically examinative discourse, for instance in the assembly or at court. 33 Consequently, both criteria of demarcation mentioned so far, logical validity and distinct and allegedly mutually exclusive topoi, remain unsatisfactory for this purpose; they do not form consistent demarcation lines between real and apparent enthymemes. Equally problematic remains the distinction between apparent enthymemes and eristic syllogismoi. Both may start from opinions that appear to be but are not reputable. This may be fairly obvious in eristic communication, to those who are used to it; it may not be so obvious, sometimes not even to the speaker, in the assembly or at court. Also, both may syllogize merely apparently but nsot really, starting from opinions while the opinions themselves may either be reputable or merely apparently reputable. Yet, “ none of the aforementioned endoxa have their phantasia completely on their surface, as the startingpoints of eristic arguments happen to do. For the nature of the false in them usually becomes quickly apparent to those who are able to detect nuances as well ” (Top. 100 b 26 - 101 a 1; tr. King). 34 The appearance of reputable opinions used in speeches in the assembly or at court may not be as apparent “ on the surface ” of what is said as they are in eristic communication. A topos which shows that some apparent enthymemes and eristic syllogismoi are closely related to each other but turn out not to be the same on closer inspection is “ the confusion of the simpliciter with that which is not simpliciter but secundum quid “ (Rhet. 1402 a 3 - 28; S. E.166 b 37 - 167 a 20). The Rhetoric compares its treatment in eristic with that in rhetoric: In dialectic it may be argued that what-is-not is, on the ground that what-is-not is what-is-not [. . .]; so also in rhetoric an apparent enthymeme may be based on the confusion of some probability secundum quid (ti eikos) with probability simpliciter (hapl ō s eikos). Now no particular probability is universally probable: as Agathon says, ‘ One might perchance say this was probable - that things improbable oft will hap to men ’ . For what is improbable does happen, and therefore it is probable that improbable things will happen. (Rhet. 1402 a 3 - 28; tr. Roberts) 31 See Top. 101 a 1 - 4, where it is said that a deductive eristic syllogismos based on apparently endoxical premises may still be called a syllogismos. An eristic syllogismos which forms a merely apparent inference may retain the name of eristic syllogismos but is not a syllogismos. 32 For a more detailed analysis see Rapp 2013, 130 - 155. 33 An example for the latter is topos 1 in B 24 if one compares it with the actual text of Isocrates ’ epideictic praise of Evagoras, which it is meant to represent (Evagoras 65 - 69). Aristotle ’ s rather brief summary of it as ‘ some he saved - others he avenged - the Greeks he freed ’ (Rhet. 1401 a 10 f) hardly does justice to the elaborate structure of the speech and its force of conviction. 34 For an explanation of this quotation carrying the force of conviction see King 2013, 185 - 199. 156 Markus H. Wörner (Galway/ Ireland) In both cases apparent contradictions may puzzle, confuse or even mislead an audience in areas of discourse such as ontology (in dialectics) or likely events in human affairs (in rhetoric) if its members are unable to ‘ detect nuances ’ and, hence, are unable to make distinctions introducing additional information. In the case of eristics, the argument “ whatever is not is - because whatever-is-not is whatever-is-not ” plays with different meanings of is, the is of identity and the is of existence. Its mere appearance lies on the surface. Nobody is likely to take it to be an argument meant seriously from the start. The argument cannot be paraphrased or amended except by adding a clause that specifies the meaning of is in each case (See Rhet. 1402 a 15 f ) without contradiction. The rhetorical argument that “ what is improbable is probable ” analogously plays with two different meanings of probable: That mothers love their children is probable generally (this is probable simpliciter); it is improbable simpliciter that they do not love their children. It cannot be probable and improbable simpliciter at the same time and in the same respect. However, a mother may not love her child in a particular case. This is probable secundum quid, but not simpliciter. It also cannot be probable and improbable secundum quid at the same time and in the same respect. The (poetical) original statement cannot be paraphrased without contradiction except by adding a further qualification capable of distinguishing between both senses of probable used in the same argument. However, the pragmatic function of such a puzzling statement, when it refers to a tragedy in which a mother murders her own children, will provide the audience with a memorable riddle concerning human life. Reformulations or paraphrases of arguments are frequently evoked via meta-communication ( “ What do you mean by what you said? ” ). As a rule, someone using an eristic syllogismos will be incapable of providing or unwilling to offer further paraphrases that yield a non-contradictory and non-ambiguous explanation of what he said. Otherwise, he would lose his case, but winning at all costs is his communicative purpose, extrinsic to any force of conviction of the argument itself. He needs to keep up the appearance of being wise because he is paid to do so. A rhetorical apparent enthymeme, in contrast, may be modified or paraphrased in such a way that an argument can be produced which potentially makes sense; if not, the speaker - who may not be trained sufficiently to detect nuances himself and, hence, may be unaware of them in what he says when using topoi for apparent enthymemes - will have to admit to the audience that he has confused things or has made a mistake. If he is sufficiently experienced and still uses apparent enthymemes which are not informed, reasonable and useful but simply are confusing for an audience which he knows to be untrained in detecting nuances, then he is acting like an eristic speaker. However, this does not so much depend on the chosen topos and the argument itself in their contribution to the intrinsic communicative aim of the enthymeme, namely to be a potentially convincing proof or refutation, but on the speaker ’ s communicative purpose - external to the argument itself - to deceive or to win at all costs. At the same time, a particular topos may specifically lend itself to this purpose. Consequently, the difference between the topoi used for inventing apparent enthymemes and eristic syllogismoi appears to be more a matter of the communicative purpose of an argument depending on the speaker ’ s deliberate choice or intentions than a matter of its correspondence to a particular topos. How Deceptive are Rhetorical Arguments from Signs or Likelihoods? 157 6. Are grounds for deception intrinsic to apparent enthymemes? It has been suggested that good and bad rhetorical syllogismoi and good and sham rhetoric can be distinguished by the absence or presence of a ground of deception (Täuschungsgrund) in the argument itself. Proponents of this position claim that this is “ a clear criterion for the distinction between real and merely apparent enthymemes ” (Rapp 2002, 778 f) and thus solves a longstanding divergence of opinion between interpreters. 35 We can take as a crucial test of this position the comparison between example (b) - see below - used in Rhet. B 24 for the topos “ from sign ” (topos 5) 36 for a merely apparent and thus so-called “ deceptive ” enthymeme and a structurally similar enthymeme (c) from a fallible sign that is regarded as “ genuine ” - in the sense of “ good and useful but not deceptive ” , given the appropriate circumstances: 37 (b) Dionysios is a thief, since he is a vicious man (pon ē ros) (Rhet. 1401 b 12 f ) 38 and (c) This woman is pregnant, since she is pale (An.pr. 70 a 19 - 24 39 ). 40 Both examples can be constructed analogously to second-figure syllogisms which - due to the relation of their terms - are always refutable and do not generate formally correct syllogisms. 41 Aristotle adds to the first example what might analogously be added to the 35 Rapp states: Ein Test für die Unterscheidung von ‘ echten ’ Enthymemen und nur scheinbaren Enthymemen ist deshalb der: Wer dazu bewegt werden konnte, ein nur scheinbares Enthymem anzuerkennen, der wird, (wenn er sich rational verhält) diese Zustimmung zurückziehen, sobald er über den Täuschungsgrund aufgeklärt wurde. Wer dagegen ein nicht-notwendiges (echtes) Zeichen-Enthymem anerkannt hat, der wird das Argument auch dann noch akzeptieren, wenn er darüber aufgeklärt wird, dass die Konklusion nicht notwendig zustande kam, denn im Unterschied zum Rezipienten des nur vermeintlichen Enthymems wurde seine Zustimmung nicht durch Täuschung über den Beweisgrund erschlichen (ibid., 779; italics are mine, M.W.). 36 I am using the numbering of topoi in Aristotle, Rhetorica, tr. W. Rhys Roberts, Oxford: Clarendon Press 1971. 37 See Rapp, ibid., 783: Dagegen bedarf das nicht-zwingende Zeichen, wenn es zur Bildung echter Enthymeme gebraucht wird, keineswegs dieses Täuschungseffektes, denn dass z. B. eine Frau blass ist, kann unter Umständen (italics are mine, M.W.) ein gutes Zeichen dafür sein, dass sie schwanger ist, wofür der Hörer keineswegs zur Annahme des falschen Satzes gebracht werden muss, dass alle blassen Frauen schwanger sind. Consequently, good and useful apparent enthymemes and deceptive ones are to be distinguished. In this case, an apparent enthymeme is not per se deceitful. On the other hand, a ground of deception allegedly inheres in apparent enthymemes. Hence, pace Rapp, good and useful apparent enthymemes now appear to be both deceptive and genuine. This apparent contradiction obviously highlights the problem discussed here. 38 (b ’ ) If being a thief (B) applies to a vicious man (A), and being a thief (B) applies to Dionysios (C), then being a vicious man (A) applies to Dionysios (C). 39 The corresponding example in the Rhetoric is: “ The fact that he breathes fast is a sign that he has fever ” (Rhet. 1357 b 18 ff ). 40 (c ’ ) If being pale (B) applies to women who are pregnant (A), and being pale (B) applies to this woman (C), then being pregnant (A) applies to this woman (C). 41 Valid second-figure syllogisms predicate the middle term of the major term and the minor term and generate only negative conclusions. However, Aristotle only mentions enthymemes with affirmative and - provided that the subject term of the conclusion is not a singular term or a proper name - indefinite-universal conclusions ( “ Those who [. . .] are [. . .] ” instead of “ All [. . .] are not [. . .] ” or “ Some [. . .] are not [. . .] ” ) as generated from signs. 158 Markus H. Wörner (Galway/ Ireland) second: “ There is, of course, no valid proof here ” (asyllogiston gar de touto) (Rhet. 1401 b 13 f ). Aristotle ’ s immediately added objection to (b) is: “ Not every vicious man is a thief, though every thief is a vicious man. ” One might be inclined to object in a similar way to (c): “ Not every pale woman is pregnant, though every pregnant woman is pale. ” For the most part, ordinary people do not object in this logical way in cases where the application of such an argument to an individual case is further specified by its circumstances. Given the specific circumstances, they may find it more or less plausible. However, although it is true to say that every thief is bad, because being a thief is intrinsically bad, one cannot maintain that it is true that every pregnant woman is pale, because pregnant women are intrinsically pale. It may nonetheless be true for the most part or more often than is rarely the case - which is the opposite of “ for the most part ” (cf. Top. 112 b 10 f) - that pregnant women are pale. The formulation of the major premise of many enthymemes of a similar kind, when made explicit, will not use a universal quantifier ( “ All ” ), but indefinite expressions such as “ women, who [. . .] ” ; “ those who [. . .] ” , “ things that [. . .], “ one does [. . .], ” leaving it undecided whether or not what is stated is claimed to be true universally. 42 Frequently, the content of enthymemes is neither unclear nor precise (m ē te asaphes m ē te akribes) in this respect. This vagueness allows enthymemes to be reasonable, in spite of the fact that it may leave them open to misinterpretation, when they would be rendered unreasonable if understood as universally valid. An audience untrained in handling this vagueness in an adequate way may indeed form a paralogism, but not because it is forced to accept a universal quantifier and, hence, a false implicit premise. However, a sign enthymeme taking the form of a second-figure syllogism is one where a successive presentation or conjunction of such signs may make the claim involved gradually stronger and more convincing for an audience. But is there a specific deception at work in case (b), about Dionysius the putative thief, as it is claimed, and is a specific deception necessary in the formation of merely apparent enthymemes? In case (b), the hearer follows the schema mentioned in SE: “ [. . .] because, when one thing is the case, another thing is necessarily the case, people believe that when the latter is the case, the former is also necessarily the case ” (SE 167 b 2 f). Consequently, the hearer is said by some commentators to succumb to this mix-up (Rapp, 783). He is allegedly made to take a non-necessary sign as a necessary sign and, hence, to take something as proven which is not proven. The speaker supposedly obtains the agreement of the recipient by fraudulently ( “ erschleichen ” ) insinuating a deception ( “ Täuschung ” ) concerning the grounds for proof ( “ Beweisgrund ” ). Case (c) however, seemingly an example of a genuine enthymeme from a fallible, non-necessary sign, is not ascribed this effect of deception ( “ Täuschungseffekt ” ). “ Depending on the circumstances ” , “ the paleness of a woman may well be a good sign of her pregnancy. The hearer is in no way forced to accept the false statement that all pale women are pregnant ” (ibid.; italics are mine, M.W). Yet, equally depending on the circumstances, the fact that somebody is well known as a pon ē ros for committing criminal offences, particularly if they involve deeds of taking away, 42 The same vagueness also characterises a great variety of special topoi that play a similar role. How Deceptive are Rhetorical Arguments from Signs or Likelihoods? 159 not returning or selling alien property in the past, or if he was found present when somebody was robbed, the well-known criminal character of a man standing accused in court may very well be counted as a reasonable and useful sign that he is likely a thief in the particular case at issue. It is highly unlikely, I think, that an experienced judge at court who is sufficiently used to detecting nuances, or a member of a jury, are forced by a fraudulently insinuated ground for deception also to accept the false statement that all criminals are thieves when they hear the accusation and its alleged reason, namely that Dionysios is a thief, because he is a criminal. They almost certainly take argument (b) as an incentive for continuing a potentially fruitful chain of proand contraarguments of accusation and defence which offers further evidence, signs or likelihoods, confirming or denying what they hear as an argument on the part of an accuser. If the hearers take the initial argument at face value, they may be suspected of not being sufficiently experienced in the discursive conventions of forensic practice, including the practice of critically examining arguments before coming to a final judgment. Emphasis on circumstances could be put in a similar way with relation to the verisimilitude of arguments such as (c). Neither case (b) nor case (c) involves a ground for deception per se. Circumstances may have the major impact on the kind or degree of likelihood of the connection between something believed to be a sign and what it is a sign of. 43 If these circumstances are ignored, the hearer may indeed form a paralogism due to an ellipsis of the where, when and how in the case. He may draw a false inference (paralogizetai) leading to a false judgment - which is not necessarily the same as being unfairly deceived or deceiving oneself (apat ā n). Moreover, he may indeed be deceived if, unknown to himself, the relation between sign and what it is a sign of is mostly not or extremely rarely the case, where the sign is otherwise meant to be a ratio essendi or cognoscendi of the thing signified. These are instances when we would hardly call the sign in question a sign at all. It is more like a vague hint. If - on the other hand - this is known to the speaker, who is also aware that the audience is ignorant of this fact, it may indeed amount to a deception. It may even be equivalent to a lie. Again, the circumstances of the utterance of an argument will decide whether or not apparent enthymemes like (b) or (c) are incentives to look for further evidence, cause errors due to paralogisms or whether or not they involve deception. Hence, the “ clear criterion ” (Rapp) of a ground of deception intrinsic to the argument is not strictly applicable in cases such as (b) and (c). Consequently, it does not function as a clear criterion which might allow us to distinguish between genuine (real) and apparent rhetorical syllogisms. One may arrive at the same conclusion by comparing the first example (d) of the use of the topos from sign taken from Rhet. B 24, but generating an apparent enthymeme (seen by those with a deductionist view of rhetorical syllogismoi as involving a ground for deception) with a structurally similar example (e) of an apparent enthymeme allegedly not involving deception: 43 Talking of “ degrees ” does not mean that something similar to statistical probability is meant here. Ordinary language uses subtle qualitative distinctions in this area, not quantified/ statistical methods of counting. 160 Markus H. Wörner (Galway/ Ireland) (d) Lovers are useful to their country, since the love of Harmodios and Aristogeiton caused the downfall of the tyrant Hipparchos (Rhet. 1401 b 10 ff ) 44 and (e) The wise are just, since Socrates is wise and just (Rhet. 1357 b 12 f; cf. An.pr. 70 a 16 - 20). 45 Cases (d) and (e) are arguments from sign where the sign stands to the signified in the relation of particular to universal. Both cases are capable of being constructed analogously to third-figure syllogisms as far as the position of their terms is concerned. Both middle terms provide a ratio cognoscendi, not a ratio essendi, for the truth of the conclusion. Both are refutable and a-syllogistic. 46 In Athens of the time of Aristotle, these two arguments use as their signs an outstanding, publicly well-known and paradigmatic case of a wise and just person (e) and of lovers useful to the democratic city state (d). It seems that, taking circumstances into account, both are informed, reasonable and useful apparent arguments from sign. Whatever the ground for deception is to be considered in case (d) or why it is absent in (e) - if it does not depend on particular circumstances that are left unmentioned - , remains unclear. Hence we should conclude that apparent enthymemes may be logically defective, but they are not eo ipso deceitful. Instead, they may be informed, reasonable and useful arguments given the appropriate circumstances. 7. A divisive criterion “ extrinsic to proofs ” Aristotle mentions a criterion “ extrinsic to proofs ” (ex ō t ō n apodeixe ō n) capable of explaining the difference between rhetorical and sophistic/ eristic arguers and also explicating the potential deceitfulness of their arguments. It refers to their communicative purposes rather than the intrinsic aim of their arguments themselves. Although both the sophist and the rhetorician may make use of appearances in argumentation, the principal mark distinguishing them is their deliberate choice (prohairesis) (Rhet.1355 b 17 f ). Deliberate choice is obviously a factor outside proofs and closely related to the moral and intellectual character of the speaker, which nonetheless affects the choice of proofs, their form, content and their phonetic presentation. A sophist earns his living from his merely apparent wisdom rather than from real wisdom, which he does not possess (SE 165 a 21 ff ). He wants to win arguments because this will increase his public reputation or income (SE 171 b 27). He may have learnt his art from dialectic, which is concerned with any kind of syllogism, genuine or apparent. However, whereas an apparent syllogism in dialectic may occur for the sake of being tested or for the sake of argumentative training, yet not with the intention to deceive, the sophist as an eristic arguer may produce a deceptive and frequently unfair argument (apat ē tikos kai adikos) (SE 171 b 9ff; 20 f). Moreover, he trades on people ’ s confusion and inability to distinguish real states of affairs from merely apparent ones, or whatever is truly 44 (d ’ ) If being useful to their country (A) applies to Harmodios and Aristogeiton (B), and being lovers [who caused the downfall of the tyrant Hipparchus] (C) applies to Harmodios and Aristogeiton (B), then being useful to their country (A) applies to lovers [who caused the downfall of the tyrant Hipparchus] (C). 45 (e ’ ) If being just (A) applies to Socrates (B), and if being wise (C) applies to Socrates (B), then being just(A) applies to the wise (C). 46 In valid third-figure syllogisms, major and minor terms are predicated of the middle term and only particular affirmative or negative conclusions are generated. How Deceptive are Rhetorical Arguments from Signs or Likelihoods? 161 valid from the only apparently valid. Finally, whatever he says himself or questions is, for the most part, not of a genuinely serious kind such that anybody ’ s direction of life depends on its truth or validity; he may be playing a kind of ‘ academic ’ discursive game, by refuting a proposed thesis in order simply to win an argument about a given problem. However, it is generally assumed that a rhetorician advising or dissuading in the assembly, defending or accusing at court or praising or blaming at a state festive occasion or a funeral, will hardly present a speech in which he does not really mean what he says; unless, that is, he has a specific intention or mix of intentions or is speaking in a social role making “ private use of reason ” rather than speaking as himself as an individual person making “ public use of reason ” (Kant 1963, 3 - 10). For the most part, occasions such as these will not involve him in discourse where the truth-value of an argument is less important than the techniques of its refutability. Yet, he may very well be involved in a kind of discourse where it is part of the very reason of the existence of the political institution (e. g. fundamental institutions of the Polis such as the court (dikast ē rion), the council (arch ē ), the assembly (ekkl ē sia) or the Areopagus) to put arguments in utramque partem to the test before coming to a final judgment or decision (krisis). This is because it may not be clear from the start which argument is according to the truth and does not hide how things are or which argument has the highest likelihood of revealing what the truth is. In these contexts, “ judges ” , Aristotle comments, “ ought to decide by considering not merely what must be true but also what is likely to be true: this is, indeed, the meaning of ‘ giving a verdict in accordance with one ’ s honest opinion ’“ (Rhet. 1402 b 31 ff ). No-one in these practical contexts can afford simply to ignore apparent enthymemes - provided that they are not unfair, intentionally deceptive or nonsensical - because too much may depend on finding the appropriate answer to a practical problem such as waging war, exiling somebody, sentencing someone to death or raising taxes. Apparent enthymemes must not be left ignored since they may involve elements of truth which will necessitate further investigation and critique, or they may lead to a conjunction of arguments adding cumulatively to a kind of evidence which is ultimately convincing. For the most part, critically examinative procedures are inherently present in such debates since people ’ s lives in a polis and their well-being may depend on them. Nonetheless, a rhetorician may or may not have a prohairesis similar to a sophist that is designed to lead to error or to deception. He may be particularly skilled in arguing for or against opposite sides of a question (t ’ anantia syllogizesthai) (Rhet. 1355 a 33 ff ). This allows him to attain a complex view of what the facts are, which is not likely if he only knows how to argue for one side only. It also allows him to refute unfair or inappropriate arguments more easily, because he is able to anticipate them. However, he should not, in actual practice, use this practical and diagnostic skill and turn it into a mechanism for inventing and using arguments for the wrong side (Rhet. 1355 a 29 ff ). “ The true and the better are more readily argued for and better to believe in by nature anyway ” (Rhet. 1355 a 36 ff ) and, most of all, “ we must not make people believe what is wrong ” (Rhet. 1355 a 31). Lastly, the demands of trustworthiness (axiopistia) do not permit this; they are presupposed in grounds for conviction “ outside the proofs ” such as the ethos of the speaker ( ē thos tou legontos) which includes prudent judgment concerning human affairs (phron ē sis), virtue (aret ē ) and benevolence towards the audience (eunoia). It may ultimately cause the rhetorician to 162 Markus H. Wörner (Galway/ Ireland) refrain from choosing to be deceptive. The same does not appear to apply to the eristic arguer: There are three things which inspire confidence in the orator ’ s own character - the three, namely, that induce us to believe a thing apart from any proof of it: good sense, good moral character, and good will. False statements and bad advice are due to one of the following three causes. Men either form a false opinion through want of good sense; or they form a true opinion, but because of their moral badness do not say what they really think; or finally, they are both sensible and upright, but not well disposed to their hearers, and may fail in consequence to recommend what they know to be the best course. These are the only possible cases (Rhet. 1378 a 6 - 14; tr. Roberts). 8. Varieties of informed, reasonable and useful apparent rhetorical syllogismoi and the place of enthymematic deductions: a proposal Leaving criteria aside which are “ extrinsic to proofs ” because they cut across real and apparent enthymemes as well as eristic syllogisms, we may finally seek a criterion of reasonableness for arguments which is connected with the endoxical status of different kinds of opinions used in argumentation; 47 we may look also for the different kinds of signs and their modal or epistemic relations to what they signify. This leads me to a proposal which is not discussed explicitly byAristotle himself but is in my view ad mentem Aristotelis, capable of identifying differences between enthymemes, and able to identify informed, reasonable and useful apparent enthymemes as well as merely apparent ones which are most likely to lead to error. The resulting system of classification reveals further complexities inherent in the area of whatever appears (phainomenon). Roughly speaking, enthymemes with the strongest force of conviction and which appear to be most reasonable involve at least one necessary or truly endoxical premise representing what is always, universally or mostly the case. 48 Obviously, there also exist weaker candidates. Some of them may come close to merely apparent and less reasonable or useful enthymemes, lending themselves to deception the lower their rank of acceptability is in a sequence of decreasing likelihood and reasonableness. This depends on the types of sign or likelihood they involve. The sequence may also indicate at what stage enthymemes may tend to become phony. Mentioning different modalities, Rhetoric B 25 points in this direction 47 Aristotle hints at two extremes of rhetorical arguments and opinions, leaving open a potential plethora of inbetween cases (neither proved nor generally admitted or believed) which may nonetheless occur quite frequently: It is possible to form syllogisms and draw conclusions from the result of previous syllogisms; or, on the other hand, from premises which have not been thus proved, and at the same time are so little accepted that they call for proof. Reasonings of the former kind will necessarily be hard to follow owing to their length, for we assume an audience of untrained thinkers; those of the latter kind will fail to win assent, because they are based on premises that are not generally admitted or believed (Rhet. 1357 a 7 - 13). 48 Cf. An.pr. 70 a 3ff: “ What men know to happen or not to happen, to be or not to be, for the most part thus and thus, is a probability, e. g. envious men hate, those who are loved show affection. ” Nonetheless, whatever is the case for the most part, i. e. in most cases or at most times, is not identical with whatever is likely to be the case. If one were to say, for instance, that “ For the most part, people ’ s hair gets grey when they age; Paul ’ s hair is getting grey; therefore he is ageing, ” one can say that “ It is likely that he is ageing since his hair is getting grey ” ; one cannot say that “ for the most part he is ageing since his hair is getting grey. ” How Deceptive are Rhetorical Arguments from Signs or Likelihoods? 163 without further elaborating on it. 49 Whereas enthymemes based on necessary signs cannot be refuted unless at least one of the premises is false, enthymemes based on likelihoods can be refuted by showing that whatever is claimed to be likely, is either not likely at all or that something else - possibly its contrary - is more likely (eikos m ā llon/ eikos in the stronger sense), either in the sense of in more cases or at more times or in both senses simultaneously (pleonakis/ plei ō ). In the latter case, the objection is most convincing (cf. Rhet. 1402 b 31 - 03a1). The more likely it is the case that the premises of the enthymeme are valid, particularly if they are valid for the most part, because this is the strong sense of likelihood (cf. Rhet. 1357 a 34; An.pr. 70 a 2 - 6), the more difficult it will be to refute the argument, provided it is formally correct. The less likely (eikos in the weaker sense) the premises are, the easier it will be to refute the argument. If they state something that is rarely, i. e. mostly not the case, then they may still be true in particular cases or at particular times, but also remain most easily refutable. Enthymemes from merely apparent likelihoods which are not even rarely likely, naturally tend to generate errors. 50 However, this may depend heavily on what people believe (oiontai/ dokountes) to be or not to be a likely state of affairs. Given the context of public rather than private communication, it is not whatever a particular individual believes to be likely which is of primary concern. Whatever is taken to be likely in public life is frequently not what seems likely to an individual, but what seems likely to people of a certain type. 51 In what follows, I shall concentrate on arguments from signs and also leave aside considerations of changes of an argument ’ s circumstances that affect audiences ’ or speakers ’ beliefs concerning likely signs. A typology similar to the one proposed, here, can be constructed analogously for arguments from likelihoods. Ceteris paribus, enthymemes from fallible, non-necessary signs, based on a more or less likely relation between the sign and what it signifies are most easily refuted if the sign is mostly not (rarely) the sign of what it is supposed to signify ( “ Paul has a sore throat; therefore he has a heart disease ” ). They may also be refuted if the sign can equally signify opposites ( “ Paul has a raised pulse rate; therefore he is afraid of failing the exam; ” “ Paul has a raised pulse rate; therefore he is excited by the thought of finishing the exam successfully ” ). Conversely, enthymemes from non-necessary signs are hardest to refute if sign and signified are related in most cases or most of the time. In this case, they come closest to what is perceived as a necessary sign ( “ Paul has a permanently stuffed nose; therefore he has sinusitis ” ). Conversely, enthymemes from signs whose alleged signs have no relation to 49 Nonetheless, this should not necessitate an analysis in terms of Aristotle ’ s theory of (mixed) modal syllogisms since this theory, as developed in An.pr., is absent here. 50 “ [. . .]about things that could not have been, and cannot now or in the future be, other than they are, nobody who takes them to be of this nature wastes his time in deliberation ” (Rhet. 1357 a 5 ff ). 51 “ In the same way the theory of rhetoric is concerned not with what seems probable to a given individual like Socrates or Hippias, but what seems probable to men of a given type; and this is true of dialectic also ” (Rhet. 1356 a 33 ff ). However, although this cannot be ignored, it does not seem sufficient to argue from whatever appears likely to individuals or to individuals of a certain kind if one has to use likely states of affairs as the basis of potentially convincing arguments; instead, whatever is likely to be the case independent of individuals ’ or particular kinds of individuals ’ beliefs is frequently regarded as crucial. The Rhetoric does not recommend other ways of discerning likelihoods independently of reputable opinions except by emphasizing the importance of a thorough collection and investigation of facts (ta hyparchonta) before speaking (cf. Rhet. 1396 a 4 - 96 b 18). 164 Markus H. Wörner (Galway/ Ireland) what they are meant to signify generate errors as long as they are believed to be signs ( “ Paul is bald; therefore he is clever ” ). The underlying stages of reasonableness of arguments from sign may be said to follow the following scheme: 52 1. Tekm ē rion-enthymemes use signs of the “ necessary ” type, but none that are “‘ necessarilynot ’ signs ” . If speakers use signs that are “ necessarily-not ” signs as if they were genuine signs, they mostly generate something deemed unreasonable ( “ Paul is resting; therefore he is dead ” ). Tekm ē rion-arguments, provided their premises are valid, form real enthymemes; “ necessarily-not ” signs generate sham arguments if they claim to be signs. Signs which are neither necessarily nor impossibly signs of what they signify are “ non-necessary ” . These form the basis of most enthymemes from fallible signs which are the more reasonable the more their relation to whatever they signify approaches the “ necessary ” type: Signs signify [necessarily] or [neither necessarily nor necessarily not] or [necessarily not] what they signify. 2. Enthymemes between the opposites of “ necessarily ” and “ necessarily not ” signs, using the “ neither-necessarily-nor-necessarily-not ” types of sign are the “ non-necessary ” signs of S ē meion-enthymemes. “ Non-necessary ” Signs signify [mostly] or [neither mostly nor mostly not] or [mostly not] what they signify. Signs belonging to the “ mostly ” type usually generate strong or quite reasonable arguments; they are closest to tekm ē rion-enthymemes. Both fulfil the definition of an enthymeme in the Rhetoric. Their opposites, those falling into the “ mostly-not ” group, may be called “ hardly ” reasonable. They are merely apparent enthymemes, coming closest to not being enthymemes at all. 3. Enthymemes between the opposites of “ mostly ” and “ mostly not ” signs using “ neithermostly-nor-mostly-not ” signs are like chemical compounds; they may be regarded as systematically unsaturated. They need to be conjoined by further arguments, adding supporting evidence and concurring on the point they are trying to make: 53 “ Neither-mostly-nor-mostly-not ” Signs signify [neither mostly nor equally likely/ likely not] or [equally likely or likely not] or [neither equally likely/ likely not nor mostly not] what they signify. 52 A more differentiated version of this scheme is suggested in Wörner 1990, 83 - 94. 53 In Aristotle ’ s Meteorologica II 2 - 3, a comparable method of looking for a convergence of sign arguments as a basis for scientific work is used in the discussion of reasons why the sea is salty (Bourgey 1975, 175 - 182). How Deceptive are Rhetorical Arguments from Signs or Likelihoods? 165 This group includes signs that may signify something of which they are equally likely to be or not to be the signs; they indicate a relation of possibilitas ad utrumlibet between sign and signified. In their case, the argument concerned may either be called “ still reasonable ” or “ not quite reasonable ” . Enthymemes using signs of this type may be called apparent enthymemes but still be more or less reasonable and useful depending on circumstances. However, they are in urgent need of support from additional arguments in order to carry a justified force of conviction. Due to the fact that this typology is not simply based on matters of fact independent of people ’ s judgments, but also on people ’ s beliefs concerning the relation between signs and what they signify, the borderlines between the types of signs mentioned are frequently fuzzy and fluid. However, fuzziness of borders does not mean that there are no clear cases to be identified. Ordinary discourse mostly relies on these as standards, particularly when human affairs are under scrutiny, such that they may turn out to be otherwise (endechomena kai all ō s echein) and where undisputed additional evidence is impossible or hard to attain. None of these types can be called eristic per se since they do not include inherent verbal trickery, unfairness or an intention to deceive, in spite of the fact that some of them may be a-syllogistic and also lend themselves more easily to deception than others. Yet all of them may be used in what counts as more or less informed, more or less reasonable and useful arguments. The fact that some of them are a-syllogistic or no longer fulfil the “ for-the-mostpart ” condition of the definition of the enthymeme in the Rhetoric does not make them fallacious or deceitful. They do not render deceitful the rhetoric which employs them. Dismissing such arguments as sham rhetoric tends to misinterpret rhetorical reasonable practice - a practice on which everyday life depends. However, rhetorical education and a rhetorical culture is needed, be it private or public, not only to speak well about contingent matters such as human action, but also to be alert enough to detect nuances in what is said. Without it, a whole generation may be deceived by demagogues. Gadamer´s understanding of rhetorical communication may indeed be superior to Kant´s. Bibliography Allan, James 2001: Inference from Signs. Ancient Debates about the Nature of Evidence, Oxford: Oxford University Press 2001 Aristotle 384 - 322 BC: “ Prior Analytics ” , in: Jonathan Barnes (ed.): The Complete Works of Aristotle - The Revised Oxford Translation (Vol.1), Princeton: Princeton University Press (1984): 39 - 113 Aristotle 384 - 322 BC, “ Topics ” , in: Jonathan Barnes (ed.): The Complete Works of Aristotle - The Revised Oxford Translation (Vol.1), Princeton: Princeton University Press (1984): 167 - 277 Aristotle 384 - 322 BC: “ Sophistical Refutations ” , in: Jonathan Barnes (ed.): The Complete Works of Aristotle - The Revised Oxford Translation (Vol.1), Princeton: Princeton University Press (1984): 278 - 314 Aristotle 367 - 322 BC: Rhetorica, tr. William Rhys Roberts, Oxford: Clarendon Press (1971) Aristotle 384 - 322 BC: “ Meteorology ” , in: Jonathan Barnes (ed.): The Complete Works of Aristotle - The Revised Oxford Translation (Vol.1), Princeton: Princeton University Press (1984): 555 - 625 Aristotle 384 - 322 BC: “ Politics ” , in: Jonathan Barnes (ed.): The Complete Works of Aristotle - The Revised Oxford Translation (Vol.2), Princeton: Princeton University Press (1984): 1986 - 2129 166 Markus H. Wörner (Galway/ Ireland) Ballweg, Ottmar and Thomas-Michael Seibert (eds.) 1982: Rhetorische Rechtstheorie, Freiburg/ Br.: Alber Barnes, Jonathan, Malcolm Schofield and Richard Sorabji (eds.) 1975: Articles on Aristotle 1. Science, London: Duckworth Barnes, Jonathan (ed.) 1984: The Complete Works of Aristotle - The Revised Oxford Translation (Vols. 1&2), Princeton: Princeton University Press Bolton, Robert 2012: “ The Aristotelian Elenchus ” , in: J. L. Fink, The Development of Dialectic from Plato to Aristotle, Cambridge University Press: Cambridge (2012), 270 - 295. Bourgey, Louis 1975: “ Observation and Experiment in Analogical Explanation ” , in Jonathan Barnes et al. (eds.), Articles on Aristotle 1. Science, London: Duckworth (1975), 175 - 182 Burnyeat, Myles 1996: “ Enthymeme: Aristotle on the Rationality of Rhetoric ” , in: Amelie Oksenberg Rorty (ed.): Essays on Aristotle ’ s Rhetoric, Berkeley: University of California Press (1996): 88 - 115 Edmondson, Ricca, Markus Wörner and Jane Pearce 2009, “ When Wisdom is called for in Clinical Reasoning ” , in: William Stempsey (ed.), Theoretical Medicine and Bioethics 30 (2009): 231 - 247 Edmondson, Ricca and Karlheinz Hülser (eds.) 2012: Politics of Practical Reasoning - Integrating Action, Discourse and Argument. Lanham: Lexington Books Eemeren, Frans van et al. 1993: Reconstructing Argumentative Discourse, Tuscaloosa and London: The University of Alabama Press Eemeren, Frans van and Peter Houtlosser (eds.) 2002: Dialectic and Rhetoric - The Warp and Woof of Argumentation Analysis, Dordrecht: Kluwer Academic Publishers Fink, Jakob 2012: The Development of Dialectic from Plato to Aristotle, Cambridge University Press: Cambridge Gadamer, Hans-Georg 1971: “ Rhetorik, Hermeneutik und Ideologiekritik. Metakritische Bemerkungen zu ,Wahrheit und Methode ‘“ , in: Hermeneutik und Ideologiekritik. Mit Beiträgen von K.- O. Apel, J. Habermas et al., Frankfurt/ M.: Suhrkamp (1971): 57 - 82 Garver, Eugene 1994: Aristotle ’ s Rhetoric. An Art of Character. Chicago: University of Chicago Press Green-Pedersen, Niels 1984: The Tradition of the Topics in the Middle Ages: The Commentaries on Aristotle ’ s and Boethius ’ ‘ Topics ’ , München/ Wien: Philosophia Verlag Hempel, C. G. , Aspects of Scientific Explanation, New York: The Free Press 1965, 394ff Hermeneutik und Ideologiekritik. Mit Beiträgen von K.-O. Apel, J. Habermas et al., Frankfurt/ M.: Suhrkamp 1971 Hess-Lüttich, Ernest and Ingo Warnke (eds.) 2017: Diskurs - Semiotisch: Aspekte Multiformaler Diskurskodierung (Diskursmuster - Discourse Patterns), Berlin: De Gruyter [forthcoming] Isokrates 370 - 365 BC (? ): “ Evagoras ” , in: Isocrates - Volume III - with an English Translation by Larue van Hook, Cambridge/ Mass.: Harvard University Press (1945): 2 - 51 Kakkuri-Knuuttila, Marja-Liisa 2005: “ The Relevance of Dialectical Skills to Philosophical Inquiry in Aristotle ” , in: Rhizai: A Journal for Ancient Philosophy and Science Vol. II No. 1 (2005): 31 - 74 Kakkuri-Knuuttila, Marja-Liisa and Miira Tuominen 2012: “ Aristotle on the Role of the Predicables in Dialectical Disputations ” , in: Oxford Studies in Ancient Philosophy, vol. XLIII (2012): 55 - 81 Kant, Immanuel 1784: “ What is Enlightenment? ” In: L. White Beck (ed. and trans.): On History, New York: Macmillan (1963): 3 - 10 Kant, Immanuel 1790: Critique of the Power of Reason - The Cambridge Edition, ed. by P. Guyer et al., Cambridge: Cambridge University Press (2001) Kassel, Rudolph 1976: Aristotelis ars rhetorica. Berlin: de Gruyter King, C. G. 2013: “ False endoxa and fallacious argumentation ” , in: Rapp/ Hasper 2013, 185 - 199 Koutras, Dimitrios (ed.) 2003: On Aristotle ’ s Poetics and the Art of Rhetoric, Athens: The Lyceum Press Manetti, Giovanni 1993: Theories of the Sign in Classical Antiquity, Bloomington and Indianapolis: Indiana University Press How Deceptive are Rhetorical Arguments from Signs or Likelihoods? 167 Oksenberg Rorty, Amélie (ed.) 1992: Essays on Aristotle ’ s Poetics, Princeton: Princeton University Press 1992 Oksenberg Rorty, Amélie (ed.) 1996: Essays on Aristotle ’ s Rhetoric, Berkeley: University of California Press Perelman, Chaim and Lucie Olbrechts-Tyteca 1971: The New Rhetoric: A Treatise on Argumentation, Notre Dame and London: University of Notre Dame Press Rapp, Christof 2002: Aristoteles, ‘ Rhetorik ’ (Vols. 1 & 2), Berlin: Akademie Verlag Rapp, Christof 2013: “ Fallacious Arguments in Aristotle ’ s Rhetoric II. 24 ” , in: Christof Rapp and Pieter Hasper (eds.), Logical Analysis and History of Philosophy - Special Issue 15: Fallacious Arguments in Ancient Philosophy, Münster: Mentis (2013): 122 - 158 Rapp, Christof and Pieter Hasper (eds., 2013), Logical Analysis and History of Philosophy - Special Issue 15: Fallacious Arguments in Ancient Philosophy, Münster: Mentis Forster, E. S. (tr.) 1971: De Rhetorica ad Alexandrum, in: Rhetorica, tr. William Rhys Roberts, Oxford: Clarendon Press (1971) Ryan, Eugene 1984: Aristotle ’ s Theory of Rhetorical Argumentation. Montreal: Les Editions Bellarmine Schreiber, Scott 2003: Aristotle on False Reasoning, Albany: State University New York Press Sprute, Jürgen 1982: Die Enthymemtheorie der aristotelischen Rhetorik, Göttingen: Vandenhoeck und Ruprecht Stroud, Scott 2014: Kant and the Promise of Rhetoric, Pennsylvania: Penn State University Press Toulmin, Stephen 1969: The Uses of Argument, Cambridge/ Mass.: Cambridge University Press Weidemann, Hermann 1989: Aristotle on Inferences from Signs (Rhetoric 2, 1357 b 1 - 25), in: Phronesis Vol. XXXIV/ 3 (1989), 343 - 351 Walton, Douglas 1990: Practical Reasoning, Savage/ Md.: Rowman and Littlefield Walton, Douglas 1996: Argumentation Schemes for Presumptive Reasoning. Mahwah/ N. J.: Erlbaum Walton, Douglas et al. 2010: Argumentation Schemes, Cambridge/ Mass.: Cambridge University Press White Beck, Lewis (ed. and trans.) 1963: On History, New York: Macmillan Wörner, Markus 1982: Enthymeme - ein Rückgriff auf Aristoteles in systematischer Absicht, In: Ottmar Ballweg and Thomas-Michael Seibert (eds., 1982), Rhetorische Rechtstheorie, Freiburg/ Br.: Alber, 73 - 98 Wörner, Markus 1990: Das Ethische in der Rhetorik des Aristoteles, Freiburg/ München: Alber 1990 Wörner, Markus 2003: The Body of Proof and the Soul of Tragedy, in: Dimitrios Koutras (ed.), On Aristotle ’ s Poetics and the Art of Rhetoric, Athens: The Lyceum Press (2003): 41 - 69 168 Markus H. Wörner (Galway/ Ireland)