Argumentation as Dialectical (Blair - Johnson)

8/2/2019 Argumentation as Dialectical (Blair - Johnson)

1/16

Argumentation as DialecticalJ. ANTHONY BLAIR AND RALPH H. JOHNSONDepartmentofPhilosophyUniversityof Windsor401 Sunset AvenueWindsor, OntarioCanadaN9B 3P4

ABSTRACT: The authors describe the rationale behind a recent development in theteaching of logic in North America and sketch some features of a dialectical theory oflogical criticism.

The authors argue that problems with validity and truth as the standards of logicallycompelling argument force abandonment of 20th century formal, deductive logic as anadequate theory of argument criticism. The latter's proper domain is implication, notargument. A new, dialectical theory of argument criticism is need.

The authors claim premises should meet a dialectical requirement of "acceptability" (inplace of the soundness requirement of "truth") - one which invokes the conception of acommunity of model interlocutors to generate standards of appraisal. Premise-conclusionconnections should meet dialectical requirements of relevance and sufficiency.KEY WORDS: Acceptability, argument, argumentation, community, dialectical, modelinterlocutor, premise-adequacy, soundness ideal, sufficiency, theory of argument criticism.

INTRODUCTION

What we are about to recount is in part the story of the intellectualjourney we found ourselves taking in our efforts to mediate between ourown training in logic and our attempts as teachers to equip our studentswith a logic that would become permanently stored in their intellectualsurvival kits. By casting our account in personal rather than objectiveterms, we indicate we are not trying to write history, but we have reason tobelieve - from our reading, and from conversations with colleagues in thewee hours after conferences and colloquia - that our experience has beenshared by many others, even if their stories will contain different chaptersor they order their chapters in a somewhat different sequence.

The account we give here is retrospective. We have come to see inhindsight how the understanding of argumentation as dialectical in naturewas a centripetal force which held together the debris created by thecollision of two vectors - the logic we were taught and the logic we foundourselves wanting to teach. Thus, although our account will place the term'dialectical' in a starring role, it is only in the last few years that we havebeen explicitly guided by the conception that argumentation is dialectical.'Argumentation 1 (1987)41-56. 1987 byD. ReidelPublishingCompany.


2/16

J. ANTHONY BLAIR AND RALPH H. JOHNSONPART I. ARGUMENTATION AS ESSENTIALLY DIALECTICAL

We shall begin with soundness, the ideal which a typical undergraduateintroductory logic course would have presented to its students in the 50sand 60s. That ideal belongs to what we would now call the theory ofcriticism (in contradistinction to the theory of argument) insofar as itanswers the question: "What is it for an argument to be a logically goodone, a cogent one?" The answer given by the soundness ideal is that it isan argument whose premises are true and whose inferences are valid.

Very shortly we shall be listing some challenges to this ideal. In order toavoid begging important questions we prefer to say, neutrally, that in acogent argument the premises and the premise-conclusion connectionmust both be "adequate." In these terms, truth is one candidate forpremise adequacy, and validity is one candidate for connection adequacy.

Our experience with this model - in the classroom, in our reading andin our thinking - has revealed a series of problems. One set of challengeshas been to the validity requirement, which holds that a connectionbetween premises and conclusion in an argument is adequate if, and onlyif, the conclusion follows from the premises, where this relationship of"following from" is usually read off as "follows necessarily from". Chal-lenges to the validity requirement came from several directions.

The first is implicit rather than explicit: it is the existence of probablereasoning. Probable reasoning has been recognized since Aristotle's day(cf., the Topics), but has been systematically studied only since the 19thcentury. The conclusion of a probabilistic argument follows from thepremises with some degree of probability rather than necessarily. Thischaracterizes much statistical reasoning, and also most scientific theorizingin which the conclusion usually outruns the evidence; or in which the fitbetween the premise set and the conclusion is never as tight as it is inmathematics, never so tight that some other conclusion is categoricallyruled out.

The realm of inductive argumentation is closely related to the realm ofscientific research, so it is not surprising that with the development of thenatural sciences, it had to be recognized that not all reasoning must fit theideal of soundness. More recently, survey research methodology showsthat the inference from any well-designed sample to the population carriessome risk of error: of going from truth to falsity. There is a range ofinference adequacy which cannot be denied. Inductive inferences varyfrom weak to strong; there is no all-or-nothing critique such as "valid-or-invalid" available.

The very existence of inductive reasoning, then, is an implicit challengeto the validity requirement: either we must say that no inductive reasoningcan ever be good (which seems preposterous), or else its goodness mustbe of a different sort, not represented in the ideal of soundness. 2

Granted that the theory which supports inductive inference is far less

42


3/16

ARGUMENTATION AS DIALECTICALwell-established, and that the concept of induction is problematic in a waythat the idea of deduction is not - still it seems clear that there is aspecies of reasoning whose goodness is not captured by the notion ofvalidity. Hence the validity requirement cannot be adequate.

This implicit challenge eventually yields to another explicit one. If weadmit that deductive reasoning is one species (properly judged by thestandard of validity) and that inductive reasoning (as typified by reasoningfrom a sample to population) is another, then how will we classify moralreasoning? Or that sort of reasoning in which the conclusion follows,ceteris paribus, or on balance, or in some other qualified way whichsuggests a more tenuous relationship between premises and conclusionthan would be the case with either deductive or inductive reasoning? Thelast ten years have witnessed a number of explicit challenges of this sort,perhaps the first of which was Wellman's in Challenge andResponse.3

The second challenge is not to the adequacy of the validity requirementbut to its applicability. First, there is the problem of translating fromordinary language into symbolic language.4 Second, ordinary languageargumentation involves a great deal of material that, from the point ofview of logical evaluation, must be regarded as clutter. So before theargument can be cast in symbolic form, decisions must be made aboutwhat belongs to the argument and what does not. Third, most argumenta-tion in ordinary language is incomplete or inchoate; hence in order toapply the theory, some decision must be made to fill in the parts that havebeen left out. The problems in reconstructing inchoate arguments withoutbegging important questions are immense, as the recent extensive litera-ture on "missing premises" attests.'

A third challenge comes from an altogether different direction.Although we have an adequate theory of validity, and hence can say ofcertain forms of reasoning with assurance: "That is valid." The situation isdifferent in respect to invalidity. To see this, imagine we have representedthe form of the argument, A, in a certain way, F; and suppose that F is aninvalid form; can we conclude that A is invalid? No, because for all weknow there is another valid form, F*, also represented by A. Thisasymmetry between judgements of validity and invalidity creates problemsfor those who defend the validity requirement; for it would mean that anargument's failing to meet that requirement could not be taken as a signthat the argument was defective.6

For these reasons, then, the validity requirement is problematic.There is also a problem with the truth requirement, and this problem

does not (unfortunately) disappear just because one recognizes the exist-ence of other forms of reasoning. If one assumes that the appropriaterequirement for the premise of an argument is truth, then it will followthat logic can say nothing at all about premise-adequacy. To ask whetherthe premise is true will take us outside of logic, after all. Hence it happensthat logicians in the 20th century have had almost nothing to say about the

43


4/16

J. ANTHONY BLAIR AND RALPH H. JOHNSONmatter of premise-adequacy. In Part III we will try to show that this is amistaken stricture.

There is a further problem with the truth requirement: As one movesaway from science and towards a different sphere of reasoning - thepractical sphere of human decision-making, the areas of morals, ethics,politics and everyday human affairs - that doctrine does begin to seemquestionable. This is not because the notion of truth is inapplicable inhuman affairs but rather because as one reviews the nature and function ofargumentation in this arena, it becomes clear that the premises need notbe true in order for the argument to be a good one.

We have thus far been speaking largely in terms of our own experienceas we now reconstruct it. That our experience was not atypical and was infact shared by others is indicated by the following developments.The appearance of Stephen Toulmin's The Uses of Argument in 1958 isone of the first explicit challenges to the ideal of soundness and, indeed, toFDL.7 For example, Toulmin questions the applicability of the Euclideanmodel to other domains and himself develops what he calls a jurispruden-tial model.

More recently, the kind of challenge directed by Toulmin has found avoice in emergence of the informal logic movement of the past fifteenyears. This movement has its origins in a pedagogical revolution having todo with the teaching of logic and dissatisfaction with the traditionalapproaches. Here is the way Howard Kahane - one of the importantinfluences on informal logic - describes his experience:

In class a few years back, while I was going over the (to me) fascinating intricacies ofthe predicate logic quantifier rules, a student asked in disgust how anything he'd learnedall semester long had any bearing whatever on President Johnson's decision to escalateagain in Vietnam. I mumbled something about bad logic on Johnson's part, and thenstated that Introductionto Logic was not that kind of course. His reply was to ask whatcourses did take up such matters, and I had to admit that so far as I knew none did. 8

From our viewpoint what Kahane was looking for is a theory of criticismwhich will help to illuminate the kinds of argument that actually occur inhis students' lives. As we begin to look at arguments in their naturallanguage settings, we have to look at their purpose and their function -not simply, as FDL would have it, at their structure.

The moral we draw from all these considerations is that something iswrong. But what? This question proved to have refractory power, i.e., itcalled into question the very conception of argumentation which was atwork in FDL and which criticism of the ideal of soundness and of theFDL tradition helped to shake. It seemed that if there were problems inthe theory of criticism then likely there were some in the theory ofargument. Slowly, it became clear that FDL had in mind one importantsubset of arguments, but the realm of argumentation was much broader

44


5/16

ARGUMENTATION AS DIALECTICALthan had been represented by FDL. What we have been describing here -admittedly in brief strokes - is the gradual erosion of a certain idealwhich we believe is accompanied by the emergence of the conception ofargument as dialectial. The time has come to explain how we use the term.

When we say that argument is dialectical,we use 'dialectical' in a waythat borrows heavily from the Aristotelian concept found in the Topics.Below we list the features which our concept of dialectical argumentationtakes from Aristotle's account. These features seem to illuminate argu-ments in a way that leads to a useful theory of argument criticism.(1) An argument understood as product - a se t of propositions withcertain characteristics - cannot be properly understood except against thebackground of the process which produced it - the process of argumenta-tion. The appropriate analogy is a move in a chess game or a play in afootball game (either kind of "football"), neither of which can be properlyunderstood out of its context. In Aristotelian dialectic, an interlocutor'scontribution has to be seen against the background of the question alreadyasked and the answers already given. In understanding argumentation, thisfeature points in the direction of background beliefs shared, or debated,by the community of informed people for whom the key propositions ofthe argument arouse interest and attention.(2) The process of argumentation presupposes a minimum of two roles:the role of questioner of a proposition (questions may be motivated by avariety of propositional attitudes: puzzlement, doubt, skepticism, rejection,devil's advocacy, ... ) and the role of answerer of those questions (theanswerer may accept the proposition in the question, or may merely holdthat the questions do not throw it into doubt or refute it). One person mayoccupy, successively, both roles; two or more persons may occupy differ-ent roles at different points in the discussion. This feature of Aristotle'smodel emphasizes the importance of understanding arguments directed byone person to another. It also has implications for the standards whicharguments should satisfy, for example, reasons offered should be relevantto doubts entertained.

(3) The process of argumentation is initiated, as the above featureimplies, by a question or doubt - some challenge - to a proposition. Thechallenge may be mooted as a possibility, or it may actually have beenposed. This factor helps us to develop guidelines for the interpretation oftexts - for example, in order to decide when a text contains an argumentor to interpret which parts of the text are argumentative and which arenot, one considers objections or doubts that have been levelled against thepoint of view being advanced, or that there is reason to expect theproponent of that point of view might anticipate. We are also guided bythis factor in assessing argumentation. For example, if a proponent of apoint of view appeals in its defense to a proposition that is also in doubt,then we insist that the supporting proposition must in turn be provided

45


6/16

J. ANTHONY BLAIR AND RALPH H. JOHNSON

with independent support, and the absence of such support is a flaw in theproponent's case.(4) Argumentation is purposive activity. Each participant has it as hisor her goal to change or reinforce the propositional attitude of the inter-locutor or of himself or herself. Thus the questioner aims at showing theanswerer that the proposition should be challenged (at best) or rejected (a tworst), or at discovering that it can withstand challenges. The answerer,conversely, seeks to show the questioner that the proposition can with-stand the challenges, or to discover that it should not be accepted (at leastin the absence of further support) or that it should be rejected. In takingthese features seriously we are led to realize that single arguments arenormally parts of a larger process and need to be interpreted andevaluated in that context. To say that argumentation is dialectical, then, isto identify it as a human practice, an exchange between two or moreindividuals in which the process of interaction shapes the product.

We have claimed that argumentation is dialectical and we haveexplained what we take that to mean. In the next three sections, we wouldlike to show how this insight can be made to yield dividends. In Part II, weshall show how this insight leads to a distinction between argumentationand inference which logicians have thus far either missed or failed to takeseriously. In Part III, we will show how the insight leads to alternatives tothe truth requirement in the matter of premise adequacy. In Part IV, weshall show how the insight helps develop alternatives to validity in thematter of connection adequacy.

PART II. THE DIFFERENCE BETWEEN ARGUMENTATION AND INFERENCE

In the previous section, we said that as a theory of criticism, FDL did notwash. Indeed, as we listed the challenges to the ideal of soundness, it washard to repress the thought that at least some of those problems stemmedfrom an underdeveloped conception of argument. The suggestion whichwe wish to make in this section - namely - that argumentation must bedistinguished from inference/implication9 will allow us to highlight theorigin of the problems with the ideal of soundness and also to answer thequestion: "If FDL is not concerned with argumentation in the dialecticalsense, then what is FDL about?"

Let us start by looking at the conception of argument in FDL. Ask alogician belonging to the FDL tradition for a definition of argument and(provided he doesn't misunderstand you to be speaking of some elementof a propositional function) he will likely give the following account: "Anargument is a set of statements some of which (the premises) are offeredin support of another (the conclusion)."Given what was said in the previous section about argumentation asdialectical, two problems with this definition stand forth. First, notice that

46


7/16

ARGUMENTATION AS DIALECTICAL

this account is structural and that it gives no real indication of thedialectical nature of argument. Second, notice that this definition seesargument as product rather than as process.

In addition there are problems with the examples of argument given byadherents of FDL. Consider the following example taken from Copi'sSymbolic Logic:l

[1] If Argentina joins the alliance then either Brazil or Chileboycotts it. If Ecuador joins the alliance then either Chile orPeru boycotts it. Chile does not boycott it. Therefore if neitherBrazil nor Peru boycotts it then neither Argentina nor Ecuadorjoins the alliance.

Here is an example from Lambert and Ulrich's The Nature ofArgument:1[2] Boston is a city and Boston is in the United States. Therefore

Boston is in the United States. l 2Perhaps the most fanciful example of an argument we have found is thisone, also from Lambert and Ulrich:

[3] The sky is blue, grass is green, therefore tigers are carnivorous.It is exceedingly difficult for anyone who takes the view that argumenta-tion is dialectial to imagine that [3] constitutes an argument in any sense ofthe term. Our point is not the trivial one that [3] is an exceedingly badargument; it is simply that [3], taken by itself, is not an argument. Hence,we are led to wonder what conception of argument can have led theauthors to suggest it.

Perhaps we can get clearer on this if we go back to:[2] Boston is a city and Boston is in the United States. Therefore

Boston is in the United States.Once again someone used to the idea of argumentation as dialecticalwould have trouble situating this performance in a dialectical setting. If thesetting is the United States, it becomes hard to see what the point of thisperformance would be. Is the writer attempting to persuade himself ofsomething he already knows but can't quite bring himself to believe? Wecannot imagine anyone producing such an argument in an effort topersuade an audience of the conclusion, for the premises contain theconclusion in a strikingly obvious fashion.In summary, if one considers both the definition of argument and theexamples it works with, it seems clear that FDL is not about argumenta-tion in the dialectical sense. But if FDL is not about argumentation, thenwhat is its subject matter?

Our answer is (it will strike some as obvious) that FDL is the logic ofimplication.Thus we have no problem seeing [2] as an instantiation of theinference rule: "from 'p and q' infer 'p'." This rule is surely correct. But

47


8/16

J. ANTHONY BLAIR AND RALPH H. JOHNSON

implication and argumentation are not the same, witness the followingpoints.

First, if argument is dialectical, the same cannot be said of inference. Itis or can be "monolectical" - discourse whose nature and significancedoes not depend upon an exchange between two or more interlocutors.Thus, I may infer from the presence of the smoke that there is a camp-fire.I need not report this inference to anyone, and it need not be validated byanyone else in order to achieve its goal. It is conceivable that the inferencecan remain "in the mind" and never be expressed and yet still be useful:i.e., I decide that I should proceed to go in that direction in the hope offinding food.

Second, their purposes are different as well: the purpose of argumenta-tion is rational persuasion (whether of the self or the other); i.e, thearguer's aim is to get the person to accept as true some proposition whichhe or she does not currently accept. But that which we infer need not becontroversial: thus when I reason in this way, "I left my wallet at home thismorning and the five dollars was in it so I can't have lost the five dollars inmy office," there need be nothing at all controversial or dubitable aboutwhat I conclude. The purpose of inference here is to discover what I donot know.

Third, implication and argument differ structurally. An inference canmove along one track; but an argument in the complete sense can onlydevelop against the background of heterogeneity of point of view and ofother arguments. This complexity will manifest itself in the structure of theargument.Why is the distinction important? The answer is that once implication isdistinguished from argument, it becomes clear that the appropriatestandards for the tw o will not necessarily be the same. Hence we mayanswer the question with which we began: "What is the role of FDL?" Ouranswer is that FDL certainly provides a marvelous, systematically reliabledoctrine about one species of inference/implication (namely, deductive);and similarly inductive logic provides a reliable account of the adequacyof one other form of implication. But FDL has little light to cast on theappropriate standards of argument.

PART III. PREMISE-ADEQUACY: ACCEPTABILITY

What all the varieties of argumentation seem to have in common is thedialectical feature that someone takes the role of proponent of a proposi-tion and someone takes the role of resisting (the critic or doubter:Aristotle's questioner). We shall call the occupant of the latter role the"audience" of the argument. The audience may intercede actively withobjections and criticisms: viz., quarrels, disputes, legal trials, formal

48


9/16

ARGUMENTATION AS DIALECTICALdebates, or sales pitches to reluctant customers. Or the audience may bemore or less insulated from the proponent, and its resistance may be asmuch imagined by the proponent as real. Such an audience might be aclass of students, or a service club luncheon group, or people attending apolitical rally, or the readers of a published article or book.

If he is to succeed in gaining the adherence of the audience to theproposition by using arguments, the proponent must begin with premisesthat the audience is willing to grant, to concede, to admit, to allow, tobelieve, to endorse, to agree with, to subscribe to, to approve of , toconsider - in a word, to "accept." This opening point seems so obvious asto be utterly uncontroversial. However, the traditional (i.e., the tradition inwhich we were educated: 20th century Anglophone) logician's response tothis first instalment has been that it is crucial to distinguish between whatthe audience might be willing to accept and what it ought to accept. Whatit should accept is nothing but true propositions. If an argument is allowedas "good" (questions of connection-adequacy aside) provided only that theproponent succeeds in getting the audience to accept the offered premises,and regardless of their truth, then such a standard of argument "goodness"is "rhetorical" (using that term pejoratively) or sophistic: the goal ispersuasion, winning agreement, not truth. Hence, the suggestion thatacceptability should be regarded as the correct premise-adequacy standardfor logically good arguments meets with parental disapproval in thehouseholds we grew up in.There is a second objection to allowing the mere fact that the audienceaccepts the proponent's premises to count as grounds for regarding thepremises as adequate and so the arguments in that respect logically goodones. This is the objection that such a standard implies relativism: forwhether or not an argument will count as logically good will be relative tothe epistemic standards of the audience, or perhaps of the proponent andaudience combined. If (epistemological) relativism is false or bad, then thisimplication is fatal to the proposal that premise acceptability be theoperative standard. Without declaring on the larger question of the meritsof epistemological relativism, we ourselves find to be unpalatable certainspecific relativistic implications of the position that any propositionwhatever that is accepted by an audience in an argumentative exchangeshould be regarded as thereby worthy of acceptance. For instance, wehold that some forms of fallacious reasoning consist of erroneousacceptance of premises (e.g., premises that are question-begging, orpremises that it is reasonable to regard as false). Hence we have to beconcerned about the critical implications of a rejection of relativism forour thesis that acceptability is the premise-adequacy requirement.

Moreover, it seems hard to understand why anyone would engage inargumentation unless he thought some objective standards of argumentadequacy were being respected. Otherwise, the social practice would be

49


10/16

J. ANTHONY BLAIR. AND RALPH H. JOHNSON

unintelligible and it certainly would be puzzling why it ever becamewidespread. Additionally, argumentation used to inquire would at the veryleast have to assume the standard of consistency and some additionalstandard to decide which of tw o inconsistent beliefs to hold. Suchconsiderations lead us to think that a condition of argumentation'sintelligibility is a presumption of independent standards of argumentcogency.

Before proceeding further, we must make sure the issue under scrutinyis not misunderstood. To do so we must make our terminology clear. Weare talking about the acceptability of the premises of what we term a"simple" argument - a set of statements adduced to support a claimwhere each supporting statement, alone or together with others, issupposed directly to support the claim, none of which is itself supportedby any argumentation. Given this picture, we can distinguish betweenpremises of simple arguments, which are unsupported, and premises ofmore complex arguments, which might themselves be supported by argu-ments. By "supported/unsupported" here we are talking about support/lack of support by further argument. We emphatically do not mean thatthe non-supported premises of simple arguments are unjustified in thesense that there is no justification of warrant for believing them. In fact,the question of the conditions of acceptability is precisely the question ofthe conditions under which one is justified in believing an unsupportedpremise of a simple argument. Hence, by hypothesis, these premisescannot be supported by arguments.

The resolution of the problem of achieving "objectivity" and avoidingtheoretically pernicious relativism lies in taking more of the medicinewhich gave rise to the problem in the first place. What is needed is tofollow through on the insight that argumentation is dialectical, and extendit to the whole surrounding community in which arguments occur. If.argumentation is to be a rational activity, as we believe it can and shouldbe, those occupying the dialectical roles in it - the roles of proponent andaudience - must take themselves to be addressing not merely theindividual "other" in the opposite role, but a larger community of others.This would be the community of interlocutors who hold well-informedbeliefs about the subject under discussion.

We wish to advance the hypothesis that in the paradigmatic case ofargumentation, those occupying the two dialectical roles conceive them-selves as trying to satisfy the demands of a community of interlocutorscharacterized by features which establish certain standards of objectivityas.a goal in the argumentative interchange.

The community of model interlocutors collectively will exhibit certaintraits of reasonableness which might be thought of as necessary conditionsof making a reliable objective judgement. The following list of traits isintended as suggestive rather than definitive. 13

50


11/16

ARGUMENTATION AS DIALECTICAL

(a) They are "knowledgeable." They are in possession of the body of"knowledge" relating to the proposition(s) in question regarded as well-established at the time of the argument. 4 They understand and can usethe methodology currently employed to check knowledge claims. More-over, they are aware of the limits of their knowledge, so that they knowwhen they do not know something that is relevant to the point(s) at issue.(b) They are reflective. They have a well-established disposition toquestion, challenge, probe, and wonder. Their "nonsense" detectors arefinely-tuned. They know when not to take things at face value, and theyare persistent in their questioning - as a good investigative reporterwould be.

(c) They are open. No individual we have met is without blind-spots,fixed ideas, biases, but in model interlocutors these soft spots are at aminimum, and in the community of model interlocutors as a whole, theymore or less cancel each other out. The community is not homogeneous:dissidents are tolerated, even encouraged; minority reports are given acareful hearing. Another aspect of this openness is that the modelcommunity is willing to change its opinions should the accumulation ofnew evidence make it only reasonable to do so. Its fondness for its ownbeliefs is counterbalanced by a strong aversion to believing what is shownto be false or dubious.

(d) They are dialectically astute. The members of the community ofmodel interlocutors are at home engaging in argumentative discussions.They are alert to possible problems of relevance, to the need for enoughevidence of the right kinds, and to the possibilities of counter-argumentsand conflicting evidence. They understand that argumentation is anintricate, many-levelled interchange of pro and con considerations, not aone-shot demonstration that settles the question once and for all.

We regard these four traits as working together. No doubt there will beother traits that a conception of argumentation as a rational dialecticalactivity will suggest, but these four begin to particularize the character ofmembers of the community of model interlocutors. The notion of such acommunity can be fleshed out further by suggestions about how it is tacitlyassumed in the paradigm of argumentation. We would mention fivefeatures of its operation.(1) For each assertion or proposition used in an argument there will bea particular group of model interlocutors - those who know somethingabout it and who have an interest in it. The interest may be practical orintellectual, but there must be some motivation to care about the accept-ability of the proposition. Arguers seeking to realize this ideal will see it asincumbent on them to know or learn the standards appropriate topropositions of that type, and try to satisfy them.

(2) The membership of the community of model interlocutors will varyfrom proposition to proposition. For some propositions the community

51


12/16

J. ANTHONY BLAIR AND RALPH H. JOHNSONwill consist of model experts (for example, when there is a well-establishedfield of "knowledge" and reliable credentials for expertise in it). Forothers, the community will be comprised of those imagined to possessexceptional wisdom or probity (for example, on questions of nationalimportance or in precedent-setting legal cases, we want our politicians tobe statesmen and our judges to be great jurists). For many other sorts ofproposition, on the other hand, the ideal standards of "knowledgeability,"reflectiveness, openness and dialectical astuteness will be in principleattainable by an y adult of normal intelligence, so that the community ofmodel interlocutors will consist of model "ordinary people," namely thosein possession of high critical standards (for example, when the proposi-tions have to do with the questions of daily life routinely facing everycitizen). This guideline directs arguers trying to meet standards to find outwhat claims are taken for granted in the "field" in question - what arematters of "common knowledge" - and what claims are controversial andso may not be accepted without defense.

(3) Clearly the community of model interlocutors will embody highstandards, but it is important that they be attainable by most membersmuch of the time. Argumentation in the real world, used about issues thathave to be settled by human beings within the limits of time and practi-cality, cannot require standards of omniscience and perfection. Weimagine the "model" interlocutors of the relevant communities to be fleshand blood people; though they are outstanding exemplars, they arenonetheless only "role models" - not gods. As applied to actual argumen-tation, for instance, this feature would imply that a proponent cannot beexpected to know everything or meet superhuman standards of reliabilityor credibility.(4) We emphasize the collective nature of the norm we are proposingby speaking of a "community" and not of an ideal individual. The point isthat the ideal of argumentation conceives of a range of critical questionsand a variety of critical points of view as needing responses. In practicethis entails that arguers must seriously consider challenges from differentperspectives.(5) Finally, we place the community of model interlocutors in history.Their "knowledge" is the "knowledge" of their time; their assumptions arethe assumptions of their particular period; the questions and challengeswhich they levy against the beliefs and assumptions of their age are thequestions and challenges of the leading critics and iconoclasts of thathistorical moment. If there are eternal truths, then our community willsurely acknowledge them; if, instead, there is nothing better than the "mostadvanced" theory or beliefs of the day, then those will be the best thatmembers of our community can espouse or hold. From the point of viewof real life argumentation, as contrasted with some sort of unattainableideal, the best that one can demand and hope for by way of standards are

52


13/16

ARGUMENTATION AS DIALECTICALjust the standards of the best available minds. Membership in that elitegroup is bestowed by those who already belong to it. In our view there isno escape from this self-reference, and in principle no harm in it.

In general, then, we are proposing that a premise in an argument isacceptable without defense just in case a person following the methodsand embodying the traits of the pertinent community of ideal interlocutorswould fail to raise a question or doubt about it. The premise is unaccept-able without defense, and should not be accepted until cogent argumenta-tion supporting it is provided, just in case someone following the methodsand embodying the traits of the pertinent community of ideal interlocutorswould raise a question or challenge to it.

This general theory of acceptability does not decide when a particularpremise of a particular argument is acceptable or not. We suggest that itdoes have practical implications in two ways, nevertheless. First, itprovides a rationale for disputes about the acceptability of a particularpremise. Second, it provides a rationale for developing more specific andmore immediately practical guidelines for the acceptability of premises.

We imagine that some of our readers will want to dismiss our concep-tion of a model interlocutor as an a prioristic,philosopher's theory, out oftouch with the actual practice of argumentation. Such an objection is anempirical claim. While we certainly have been led to some of its featuresby a prioristic considerations - such as imagining the conditions of thepossibility of argumentation's being a rational practical activity - we havealso been motivated by about fifteen year's work in the trenches, carefullyexamining and trying to understand and assess the argumentative texts ofnon-philosophers - indeed, of non-specialists of any kind - of theproverbial "man (and woman) in the street," arguing with passion ordetachment about issues that mattered enough to them to put pen to paperand to try to work out a coherent train of thought on a topic. Ourtheoretical predispositions may have blinded us to relevant features ofsuch texts. However, if so , that is a charge that must be made anddefended on specific points; it cannot be intelligibly levelled in principle. Ifthe schematic depiction of our hypothesis is specific enough to serve as atarget for such criticisms, we have succeeded in our aim of proposing analternative conception of premise-adequacy to the truth requirement ofthe soundness ideal.

PART IV . CONNECTION ADEQUACYWe have been discussing an hypothesis in terms of which we suggest thepremise-adequacy requirement of cogent arguments be conceived. Ifargumentation is properly to be understood as dialectical, then it is to beexpected that a dialectical analysis should apply to the connection-

53


14/16

J. ANTHONY BLAIR AND RALPH H. JOHNSONadequacy requirement of cogent arguments. Our view is that this require-ment has two separate aspects: a requirement of relevance (the premisesof an argument ought to be relevant to the conclusion) and a requirementof sufficiency (the premises ought to provide sufficient support for theconclusion).

We do not have a detailed dialectical account of relevance andsufficiency to offer, but we do believe this is a promising direction forresearch. To illustrate its possibilities, we will discuss one aspect of therequirement of sufficiency.

The aspect of sufficiency we have selected deals with what might betermed the "dialectical obligations" of the arguer. We have in mind thesituation in which a proponent arguing for a claim encounters objectionsfrom his audience. If those objections are not met in one way or another,the arguer's task has not been completed. We envisage possible objectionsof three kinds, corresponding to the three adequacy requirements of acogent argument: an undefended premise may be attacked as unaccept-able; and the premise-conclusion connection may be attacked either ongrounds that one or more of the premises is irrelevant, or on grounds thatas they stand the premises fail to supply sufficient support for theconclusion. When presented with an objection of any one of these threekinds, particularly an objection itself supported with reasons, theproponent of the claim in question owes some response. At the very leasthe must argue that the objection in question is irrelevant or ill-considered.An argument that fails to do so is in violation of the sufficiencyrequirement.

When, as frequently happens, a proponent offers arguments to ananticipated audience, and not an audience which is present and which canintercede directly, there is still an obligation to defend the claim, and alsothe premises of arguments offered in its support, against possible objec-tions. An argument which fails to include such defenses is just asincomplete in the second respect as one which ignores stated objectionsby active interlocutors.

This is a strong position to take, since it entails that a great manyarguments of the sort given in logic textbooks, whether those of the FDLtradition or the "new wave" textsls are incomplete arguments; for rarelydo such arguments attempt to deal with known or possible objections thatbear on the issue.

How then do we justify the judgement that an argument must becomplete in this sense? How is it to be decided what points ought to beargued for, and where the onus for defense ends, when there is noactually-interacting audience?

It is our suggestion that the concept of the community of modelinterlocutors can serve in answering these questions. First, an argument isconsidered to be incomplete when it does not engage the common, known

54


15/16

ARGUMENTATION AS DIALECTICALobjections to its conclusion and to the premises of arguments supportingit, simply because we conceive of the audience of responsible argument tobe the community of model interlocutors, and these objections will havebeen raised by members of that community. Since the purpose ofargumentation is to persuade these people to accept the claim, and it isknown that they will have these objections, the nature of the exerciserequires that those objections be addressed if there is to be any hope ofthe argument's achieving its purpose.

Second, we decide what claims need support by appealing to whatclaims are in fact problematic in the relevant community of model inter-locutors. Moreover we do not consider that a proponent has to answerevery conceivable objection, no matter how misguided or foolish, and thatis because such objections would simply not arise in the community ofmodel interlocutors as we envisage it. Reflective, dialectically astutepeople do not waste time on misguided or foolish matters. So it seems thatas a matter of fact some such concept as this community is in fact inoperation when we make judgments about whether a proponent has triedto answer all the relevant objections - that is, judgements about premise-sufficiency in the second sense.So much, in any case, for our sketch of how the understanding ofargumentation as dialectical in general, and our hypothesis about thecommunity of model interlocutors in particular, might apply to part of thesufficiency requirement for connection adequacy. We hope it is evocativeenough to suggest the merits of a dialectical approach.

CONCLUSION

The "theory" we have sketched has a perhaps confusingly dual character.To some extent it is empirical. That is, it has been inspired by observationsof actual argumentation and argument criticism, and if there is noevidence for our hypotheses in actual argumentation, the hypotheses arerefuted. At the same time, the argument is partly "transcendental" (inKant's sense), or a priori: we have been trying to characterize theconditions of the possibility of rational argumentation.

In this paper it has been our intention to report, on what we believe is atypical development of thought for Anglophone philosophers in NorthAmerica who identify their teaching of logic with the development ofinformal logic over the past fifteen years. This process has in part includedan increasing dissatisfaction with the ideal of soundness as a tool for thecritique of arguments and in part included a growing appreciation of thefact that argumentation is dialectical and of the implications of this fact forthe critique of arguments. These tw o processes proceeded paripassu, andhave increasingly come to influence each other. We have tried to identify

55


16/16

56 J. ANTHONY BLAIR AND RALPH H. JOHNSONthe main objections to the ideal of soundness, and to indicate the ways inwhich appreciation of the dialectical character of argumentation hasinfluenced the theory of logical criticism in our own work in informallogic.

NOTES

''Dialectical' is a much-used term, with many senses; anyone employing it owes hisreaders an explication of the particular meaning he assigns to it. We will discharge thisobligation in due course.2 There is a third alternative, taken by some, of reinterpreting validity as a range-coveringconcept so that it admits of degrees. See, for example, Stephen N. Thomas, PracticalReasoning in NaturalLanguage,3rd edition, Prentice-Hall, Englewood Cliffs, NJ, 1986.3 Carl Wellman, Challenge and Response, Justification n Ethics, Southern Illinois Univer-sity Press, Carbondale and Edwardsville, IL; Feffer & Simons, London and Amsterdam,1971.4 See Yehosha Bar-Hillel, 'Formal Logic and Natural Languages,' in Foundations ofLanguage 5 (1969), p. 15; and Michael Scriven, 'Philosophy of Education: LearningTheory and Teaching Machines,' The JournalofPhilsophy 62 (1970), 896-908.5 See the report of these in Ralph H. Johnson and J. Anthony Blair, 'Informal Logic: ThePast Five Years, 1978-83,'American PhilosophicalQuarterly22 (1985), 181-196.6 See Gerald Massey, 'Are There Any Good Arguments That Bad Arguments Are Bad?'Philosophy in Context 4 (1975), 61-77; and 'The Fallacy Behind Fallacies,' MidwestStudies in Philosophy6 (1981), 489-500.7 Stephen Edelston Toulmin, The Uses of Argument, Cambridge Univerity Press, Cam-bridge, 1958. "FDL" abbreviates "formal deductive logic."8 Howard Kahane, Logic and Contemporary Rhetoric: The Use of Reason in EverydayLife, 1st Edition, Wadsworth, Belmont, CA, 1971, p. vii.9 For our purposes here, we shall not draw a rigid line between implication - a relation-ship between statements - and inference - a mental act in which an implication isembodied or reported.10 Irving M. Copi, Symbolic Logic, 2nd Edition, Macmillan, New York, 1965, p. 50.1 Karel Lambert and William Ulrich, The Nature of Argument, Macmillan, New York,1980, p. 31.12 The NatureofArgument, p. 19.13 Readers will notice the similarity between this concept and Pereleman and Olbrecht-Tyteca's "universal audience."14 'Knowledge' and its cognates should be taken to stand in scare quotes throughout. Wewant thereby to side-step the problems relating to the definition and possibility of knowl-edge. In contexts outside of technical philosophy a person may be described as "knowl-edgeable" even when the content of her "knowledge" might more precisely be specified as"informed opinion." Furthermore, of course we are aware of the well-known philosophicalcharacterization of knowledge as justified true belief, which requires, if skepticism is false,that it be possible to speak of statements as being "true." But the objection to requiringtruth as the criterion of premise-adequacy does not imply that one may not speak of "true"statements.15 See our, 'The Recent Development of Informal Logic,' in Blair and Johnson (eds.)Informal Logic, The Proceedings of the First International Symposium on Informal Logic,Edgepress, Inverness, CA, 1980.

Argumentation as Dialectical (Blair - Johnson)

Documents

Transcript of Argumentation as Dialectical (Blair - Johnson)