Foundations of the Theory of Argumentation

INFORMATION TO USERS

This reproduction was made from a copy of a document sent to us for microfilming. While the most advanced technology has been used to photograph and reproduce this document, the quality of the reproduction is heavily dependent upon the quality o f the material submitted.

The following explanation of techniques is provided to help clarify markings or notations which may appear on this reproduction.

1. The sign or “target” for pages apparently lacking from the document photographed is “Missing Page(s)” . I f it was possible to obtain the missing page(s) or section, they are spliced into the film along with adjacent pages. This may have necessitated cutting through an image and duplicating adjacent pages to assure complete continuity.

2. When an image on the film is obliterated with a round black mark, it is an indication of either blurred copy because of movement during exposure, duplicate copy, or copyrighted materials that should not have been filmed. For blurred pages, a good image of the page can be found in the adjacent frame. I f copyrighted materials were deleted, a target note will appear listing the pages in the adjacent frame.

3. When a map, drawing or chart, etc., is part of the material being photographed, a definite method of “sectioning” the material has been followed. I t is customary to begin filming at the upper left hand comer of a large sheet and to continue from left to right in equal sections with small overlaps. I f necessary, sectioning is continued again-beginning below the first row and continuing on until complete.

4. For illustrations that cannot be satisfactorily reproduced by xerographic means, photographic prints can be purchased at additional cost and inserted into your xerographic copy. These prints are available upon request from the Dissertations Customer Services Department.

5. Some pages in any document may have indistinct print. In all cases the best available copy has been filmed.

UniversityMicrofilms

International300 N. Zeeb Road Ann Arbor, Ml 48106

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.


8302878

Harpine, William Douglas

FOUNDATIONS OF THE THEORY OF ARGUMENTATION

University of Illinois at Urbana-Champaign PH.D. 1982

University Microfilms

International 300 N. Zeeb Road, Ann Arbor, M I 48106



FOUNDATIONS OF THE THEORY OF ARGUMENTATION

BYWILLIAM DOUGLAS HARPINE

B.A., College of William and Mary, 1973 M.A., Northern Illinois University, 1974

THESISSubmitted in partial fulfillment of the requirements

for the degree of Doctor of Philosophy in Speech Communication in the Graduate College of the

University of Illinois at Urbana-Champaign, 1982

Urbana, Illinois


UNIVERSITY OF ILLINOIS AT URBANA-CHAMPAIGN

THE GRADUATE COLLEGE

SEPTEMBER 1982

WE HEREBY RECOMMEND THAT THE THESIS BY

W ILLIA M DOUGLAS HARPINE

F.N TTTT.FTI FOUNDATIONS OF THE THEORY OF ARGUMENTATION

BE ACCEPTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR

THE DEGREE D F DOCTOR OF PHILOSOPHY________________________________________

Dire dfor of Thesis Research

------Head of Department

1

t Required for doctor's degree but not for master's.


For Elaine and David


ivACKNOWLEDGMENTS

Daniel J. O'Keefe and Kenneth Andersen read earlier drafts of this thesis thoroughly and attentively; their comments and suggestions improved many portions of the work. William P. Alston, who first pointed out to me the importance of John Pollock's work, is also responsible for leading me and a few other souls through a grueling but effective semester of training in philosophical thinking.

My adviser, Joseph W. Wenzel, made more contributions to the success of this project than I could possibly list. He showed gentle insistence on the highest standards, astute analysis of the issues that I was investigating, and great patience with an unusually obstinate student. Without his assistance this research could never have been successfully completed.

My wife Elaine not only typed most of the manuscript, but also supported me with love and encouraged me throughout the past five years.


V

TABLE OF CONTENTS

Chapter Page

I. THE STUDY OF ARGUMENTATION............. 1Purpose of the Study............... 2Arguments, Propositions, and Logic . . . . 5Rhetoric and the Field of Argument . . . . 9Theory Building and the Humanities . . . . 26Procedure........................ 32Conclusion........................ 37Notes ............................... 38

II. TRADITIONAL MODELS OF ARGUMENTATION:DEDUCTIVE LOGIC AS TRUTH-FUNCTIONAL ........ 43

Meaning.......................... 47Traditional Logic, Reductionism,and Truth-Conditions ................... 57Rhetoric and Truth-Functional Logic . . . 65Are There Alternate Criteria ofEvidence?........................ 72Conclusion........................ 77N o t e s ............................ 78

III. TRADITIONAL LOGIC: ENTHYMEME ANDINDUCTION............................ 84

The Enthymeme and Probability..... 84Induction and Probability ............. 99


viConclusion............................. 110N o t e s ................................ 112

IV. PROCEDURES AND FIELDS: THE DIALECTICALVIEW...................................... 116

Can Dialectics Substitute for Logic? . . . 116Fields of Argument..................... 129Conclusion............... 143N o t e s ................................ 146

V. A THEORY OF CONCEPTUAL INFERENCE............ 150Inference............................. 154Grounds and Sources................... 171Conclusion............................. 195N o t e s ................................ 197

CONCLUSION.................................... 201BIBLIOGRAPHY .................................. 208V I T A .......................................... 216


CHAPTER ITHE STUDY OF ARGUMENTATION

In The Screwtape Letters, the elder devil Screwtape writes to his nephew Wormwood on the finer points of seducing humans into sin:

My Dear Wormwood,I note what you say about guiding your

patient's reading and taking care that he sees a good deal of his materialist friend. But are you not being a trifle naif? It sounds as if you supposed that argument was the way to keep him out of the Enemy^s clutches. That might have been so if he had lived a few centuries earlier. At that time the humans still knew pretty well when a thing was proved and when it was not; and if it was proved they really believed it. They still connected thinking with doing and were prepared to alter their chain of life as the result of a chain of reasoning.But what with the weekly press and other such weapons we have largely altered that. Your man has been accustomed, ever since he was a boy, to have a dozen incompatible philosophies dancing about together inside his head. He doesn't think of doctrines as primarily "true" or "false", but as "academic" or "practical", "outworn" or contemporary", "conventional" or "ruthless". . . . Don't waste time trying to make him think that materialism is true! Make him think that it is strong, or stark,or cour- ageous— that it is the philosophy of the future.That's the sort of thing he cares about,

Screwtape's observations on the modern mind notwithstanding,argurents can be evaluated by standards not derived frommere effectiveness. Rather, arguments can be evaluated

1


on logical grounds. A critic can use a set of standards to evaluate the soundness of arguments. This thesis discusses an approach to the field of argumentation, which may be defined as the study of reason-giving discourse in situations where a person or group of persons wishes to influence smother. This field may properly study both the interaction of arguers and the content of their communication.

Purpose of the StudyThis thesis takes up a group of related questions on

the foundations that must be laid during the construction of a theory of argumentation. In particular, this study asks, what are the logical foundations for a theory of argumentation? Answering this question involves addressing a series of problems. The first is, is there anything inadequate about the syllogism, symbolic logic or nonformal deductive logic? Do these traditional logics meet the needs of the critic who assesses arguments that are abstracted from discourse intended to influence someone's beliefs, attitudes, or actions?

If indeed traditional logic is inadequate, and I believe that it is, what is wrong with it? Why does it fall short of adequacy? On what assumptions can a better and more adequate theory be worked out? For example, should we accept Toulmin's opinion that arguments may be


evaluated chiefly by the standards of particular fields?If so, what is the most useful way to characterize a field? How can an improved logic encompass criticism of the soundness of arguments that might concern matters of fact and circumstance, are often value-laden, and which might often have implications for action?

What kinds of standards should arguments meet? What kinds of criteria are invoked to justify the claim that an argument meets a given standard? Most logicians allow only criterion is truth-functional validity. This, for reasons that I will take up in the next chapter, is too limited: it permits little or not place for substantialinferences. The need for additional standards and criteria

Ois hardly controversial. The issue is, what kinds of standards and criteria should be employed?

A number of epistomologists deny that deduction is the only correct means of inference in the justification of non-basic propositions. That is, they deny that deductive inference is the only model of inference appropriate to the justification of knowledge and belief. Others propose that rules of evidence or criteria of justification must underlie substantial inference. These two approaches seem to differ in little but name. A principal task of this thesis will be to develop or adopt a way of understanding


standards that justificatory inference should meet and thecriteria that reveal whether an argument has met a given

4standard.In short, the main problem is, can a logic be devised

which is adequate to the needs of a rhetorical critic concerned with the soundness of arguments? If this last question can be answered, it will provide an understanding of the logical foundations for a theory of argument. To answer these questions, I will maintain that the logical study of argumentation has assumed an invalid theory, namely, that the meaning of conceptual relations can only be given by the analysis of the conditions under which a proposition is true or false. This error has in some cases led writers to make use of traditional logics that are woefully inadequate for rhetorical criticism. In other cases, it seems to have encouraged a movement against making use of logic in the theory of argumentation: some writers,recognizing the inadequacy of traditional logic, attempt to divorce argumentation from logic because they see no clear alternative. This thesis outlines one.

A theory of argumentation should assume a theory of propositional meaning based on the conditions under which a proposition is justified. The field of argumentation, in part, studies the standards and criteria by which a critic


can evaluate the soundness of inference in the arguments of rhetoric. In other words, argumentation subjects the content of rhetoric to evaluative criteria. Argumentation is thus an area of common interest to logic and rhetoric. A critic can evaluate arguments on rational grounds; a part of the theory of argumentation is the logical apparatus which guides such criticism.

Arguments, Propositions, and Logic Within the logical perspective, an argument is con

ceived of as an object of analysis. This conception does not imply that an argument is an existent thing; an argument is necessarily the interpretation of a critic as much as it is the product of an arguer’s speech. An argument, for purposes of logical evaluation, is an abstraction. An interpreter abstracts it from the on-going flow of communication. Likewise, the interpreter judges the meaning that the argument has and comes to understand (or fails to understand) the concepts expressed in the argument and the significance of those concepts. The interpreter can demand of the argument that it be of a certain quality, and can apply an analytical method to the argument to determine if the argument is of the quality he desires. It is upon this analytical method that this thesis focuses.5


O'Keefe distinguishes between two senses of the word "argument." One, which O'Keefe calls "argument^," is "a kind of utterance or a sort of communicative act." An "argument2 " is a kind of interaction. This is a distinction between making an argument and haying an argument.6 The logical point of view takes up the study of something that is more like argument-̂ . It is, however, concerned with more than just the utterance or communicative act. More precisely, logical criticism abstracts the content of utterances from the discourse in which they occur, and lays them out for evaluation. Thus, the logical prespective is concerned with the content of the speech act, as well as with its performance. With O'Keefe's view broadened in this way, the study can be understood as concerned with the study of arguments as speech acts and as propositions. The propositional content can be abstracted from its occurrence in an argument2 and from the speech act performed in uttering it.^ The proposition(s) indicate the content of the argument, its sense, and can be expressed as one or more declarative sentences.

As Toulmin points out, one person often makes an assertion to another. His interlocuter may then ask him to justify his assertion, to give arguments supporting it.These arguments, once made, are then subject to evaluation


according to the standards and criteria of logic. The theory of argumentation includes a concern for evaluating the drawing of conclusions from evidence. The criteria for evaluation in turn imply standards that rhetors should meet in the creation of discourse.

Wenzel points out that the term "argument" is used in three particularly important senses:

Of the several senses in which scholars use the term "argument" and its relations, three are of immediate importance: argument as process,argument as procedure, and argument as product.. . . When used by specialists, each sense of the term refers implicitly to a distinct perspective taken in the examination of arguers and their behaviors, and the perspectives are roughly aligned with the disciplines that have historically been concerned with argument.Thus, the three senses correlate respectively with the perspectives of rhetoric, dialectic, and logic.8

The rhetorical perspective applies the standard of effectiveness; the dialectical perspective applies the standard of candidness, and the logical perspective employs the standard of soundness. Notwithstanding the relationships among these three standards, important distinctions can be drawn between the standard of the logical perspective and those of the other two perspectives. The logical critic undertakes to determine the quality of "an argument abstracted from rhetorical process or dialectical proced-

9ure." The logician's task is to lay out the theoretical


understanding which guides the process of evaluating the soundness of arguments.

The rhetorical and dialectical perspectives are both concerned with the communication contexts in which argumentation occurs. The rhetor seeks to influence the thoughts and actions of other people, and the rhetor's behavior is governed by the tacit social rules that regulate interaction. In dialectic, the social rules are made explicit; indeed, "The conscious articulation of rules is a defining characteristic of the dialectical perspective on argument."^ Thus, while the logical theorist abstracts propositional arguments from discourse, the rhetorical or dialectical theorist is necessarily immersed in the study of the processes of communication.

The logical perspective has, until recently, dominated the study of argumentation. However, as the inadequacy of traditional logic has become more sharply recognized, scholars have come to emphasize the dialectical and rhetorical perspectives. At the same time, some scholars have attempted to cope with the inadequacy of traditional logic by claiming that standards of soundness are closely related to standards of effectiveness or candidness. This thesis endeavors to show that a more adequate understanding of logic might serve the purposes of logical criticism,


9eliminating the need to subordinate logic to rhetoric or dialectic.

Rhetoric and the Field of Argument The theory of argumentation, seen in the context of

rhetorical theory, should embody a conception of logic adequate for analyzing rhetorical arguments. Although other approaches have gained ascendency, the traditional approach, grounded in Aristotle's understanding that logical studies are an integral part of rhetorical studies, can have a breadth and richness of application greater than what is usually acknowledged. Developing this position will require a broader than usual view of logic and of the relationship between logic and rhetoric.

An Approach to Rhetoric and LogicThis study seeks the underlying principles of the log

ical view of argument. The principles of inference should apply to the study of rhetoric. Throughout this thesis, I will consider logic to be the study of the standards and criteria for evaluation of soundness in rational discourse. Fisher defines logic as "a systematic set of procedures" to be used in the analysis and evaluation of the inferences embodied in discourse.^ Most logicians, of course, have a narrower view of logic. For example, Quine notes the trend


10to limit the term "logic" to "the science of necessary inference."^2 Such a view clearly rules out most, if not all, of the arguments found in rhetorical discourse.

Because of that narrow conception, rhetoricians frequently condemn logic as a barren study concerned only with the minute analysis of the formal structures of trivial arguments. Poster's opinion is typical: logic helps "butlittle in the practical difficulties of reasoning well and of exposing the more plausible kinds of unsound reasoning and verbal confusion;" seldom do real arguments fit the "simple pattern of the syllogism.Because of such criticism, some rhetoricians increasingly seek to explain argumentation without reference to logic. However, the usual, limited view is not the only view of logic. Logic can and should expand into a broader field of study which has much to contribute to our understanding of rhetoric. Perelman, lamenting the failure of logicians since Descartes to recognize the legitimacy of arguments of probability or likeness to truth, points out Aristotle's emphasis on dialectical reasoning as indispensable for practical decision or action.*4

There is little question that formal logic is inadequate for the purpose of analyzing arguments. However, few have considered the most fundamental reason why formal


11logic is inadequate to argumentation: the presence of theillness has been discovered and extensively documented, but its diagnosis has not been clearly explained. The reason for this is that the fundamental assumption behind formal logic, that the meaning of propositions is given by statingthe conditions under which they are true, is rarely noted

15and almost never questioned. This conception is the most common one given to justify logical theory, although logicians differ in the details by which they work out theories of meaning. Davidson's opinion is typical of such an approach: "The definition [of meaning] works by givingnecessary and sufficient conditions as a way of giving the meaning of a sentence."16

Because this fundamental reason for the inadequacy of formal logic has not often been considered, the full reason for the inadequacy of formal logic has not been uncovered. For this same reason, as I will argue in subsequent chapters, alternative conceptions of the logical perspective are incomplete. A sound logical theory of argument requires a sound theory of meaning.

Among those writers who have been of greatest influence in the study of rhetoric, Perelman and Olbrechts-Tyteca have come closest to explicitly stating this aspect of the problem:


12The very nature of deliberation and argumentation is opposed to necessity and self-evidence, since no one deliberates where the solution is necessary or argues against what is self-evident. The domain of argumentation is that of the credible, the plausible, the probable, to the degree that the latter eludes the certainty of calculations.17Most philosophers treat logic as formal, truth-

functional inference: if the premises are true, theconclusion is true. But how often do we have suchcertainty in the realms of rhetoric? This thesis arguesfor a greatly expanded conception of logical standardsand criteria. Such a broadened logic is more adequate tothe study of argumentation.18

Rhetoric, Argumentation, and RationalityArgumentation does not study all aspects of rhetoric.

The logical perspective on argumentation subjects the content of the discourse to rational standards of evaluation. As Bryant notes, "Rhetoric . . . has always run along with logic and with whatever modes and methods of sound and productive thinking mankind has conceived and formulated.w1^As Screwtape realized, sound argumentation is likely to be productive of sound conclusions and wise decisions. It seems self-evident that we should prefer sound arguments to unsound ones, for the same reason that we should prefer wise decisions to foolish ones. The criticism of rhetorical acts may, in part, evaluate the quality of arguments;


13

encouraging sound arguments encourages wiser rhetoric. We may follow the example of Socrates, who criticized Gorgias for failing to argue soundly and for, instead, swaying

20audiences with whatever effective material he could muster.The outcome of argumentation should be founded on the

soundness of arguments: we should expect the resolutionof a process of argumentation to reflect the force of the better argument. Recognizing this purpose, the logical perspective seeks to provide the tools for determining which argument is better.

Breadth of applicability. First of all, the standards and criteria must apply to substantial arguments; that is, to a broader class of arguments than the tautologies of formal logic. Rhetorical exigencies do not normally admit of certain solutions; nonetheless the argumentation to which they give rise may be as adequate and sound as a critic can rightfully expect. A theory of argumentation must encompass the logic of arguments that are substantial, uncertain, difficult, or ambiguous. Second, the rhetorical exigence inherently yields problems which are not merely factual, but valuational. Decisions must be made, gains and costs compared, human loss and fulfillment recognized.The theory of argumentation must study the argumentation of values.


14The theory of argument helps to explicate types of

standards that arguments may meet. An argument must meet a criterion, or set of criteria, in order to come up to meet a given standard. For example, a traditional theory of logic may propose the standard of absolute certainty. An argument might meet this standard by meeting the criterion of analytic necessity. In this application, the term "certainty" applies to a quality of the argument, that its conclusion is unquestionably true. The term does not refer to a merely subjective feeling of certainty. Whether any argument in fact could meet such a standard is debatable, but no important conclusion of this thesis requires such an assumption. It may be that each standard carries with it a single criterion. Or it may be that an argument can meet a standard by meeting one of multiple criteria. Meeting a criterion justifies the claim that an argument comes up to some standard.

The truth of many matters important to rhetors is uncertain or hidden. A statesman predicting the best path toward peace, an economist interpreting the employment situation and, in fact, most rhetors, lack access to many certainties. In rhetoric, we must often make what we can out of the best evidence available, however uncertain that may be.


15When engaging in rhetoric, people cannot often restrict

themselves to discourse that meets the rarefied standards of Plato's Phaedrus. Nor is it always sensible to demand that arguments be completely justified. As Bitzer noted in his discussion of the exigence, discourse is often called for at a particular time. By the time one acquires all the information necessary for complete justification, even if such justification is possible at all, it almost certainly would be too late to act. This thesis is only peripherally concerned with knowledge, in the strict sense that opposes knowledge to mere true belief. Rather, the thesis is concerned with the justifications offered for claims.

‘.%us arises the problem of multiple criteria. Argumentation cannot rely on the single, intuitively compelling, but inadequate model of deductive logic. Instead, theories of argumentation universally recognize that there are multiple criteria for the soundness of an argument, not just the single criterion of formal validity. Is Toulmin right that there are countless criteria, that each field of argument develops its own criteria independently of all other fields?^! Or are there categories, transcending fields, by which these many criteria may be understood? One would hope so, as the alternative borders on criterial anarchy, but an adequate account of such categories remains to be given.


16Importance of contexts. The theory of argument must

comprehend the necessity of interpreting arguments. The primary concern, from a logical perspective, is to comprehend the.propositional content of arguments. This thesis discusses various approaches by which arguments may be critically evaluated to see if they are sound. These methods should recognize that arguments are abstracted from rhetorical discourse. The critic therefore must begin by abstracting the argument, laying it out, and interpreting it in light of the rhetorical situation. As Kneupper explains:

A variety of situational variables do affect the meaning of an argument. These variables are not diagrammed. But it is precisely such situational variables which rhetorical critics attempt to reconstruct in order better to approximate the meaning of the argument. Rhetorical critics utilize such reconstructions to enable them to identify meta-messages, which suggest that the explicit form of the argument is inter- pretively transformed by source, receiver, or both.

It is wise to be wary of too facile or literal interpretation of the language of a speech text. But alternative meanings are discovered through using nonverbal and situational cues to transform the language explicit in the text.22

We cannot, for example, expect to comprehend an argument'ssignificance without some understanding of the speaker'sintentions or the reasons for which the speaker introducesthe argument into a discourse. Correctly interpreting an


17argument in light of its context is critical to its evaluation. The evaluation of arguments must, for example, recognize the purpose of the discourse, its audience, and the problem at issue.

Anderson and Mortensen argue that rhetorical contexts are not amenable to a field-invariant logic because rhetorical arguments do not generally express the movement from premise to conclusion with sufficient precision. The arguments of rhetoric are subject to contexts. Rhetorical arguments are not, they continue, expressed in terms of sufficient stability to permit the application of context- invariant connectives. Especially when argumentation deals with morals or evaluation, or uses figures of speech, "The meanings of the connective terms are not clearly specified by the linguistic context of the argument itself."Anderson and Mortensen fear that when a rhetorical critic attempts to apply a context-invariant logic to rhetorical discourse, he will soon encounter material so ambiguous with meanings so "unstable" that "a connective with objective meanings is not available.” Thus, argumentation cannot always be judged by the same method of analysis.23

It would be a mistake to view Anderson and Mortensen*s analysis as a claim that rhetorical arguments frequently embody, as an inherent feature, the fallacy of equivocation.


18A more reasonable interpretation is that argumentation must be evaluated with regard to the context in which discourse arises. This view is an extension of a viewpoint similar to Toulmin's. In other words, Toulmin recognizes the importance to logic of understanding the context in which arguments arise, while rhetoric can contribute to the analysis of the contexts and their significance.^

This approach should be distinguished from that which Willard criticizes in a recent paper. Willard characterizes the logical perspective "as rooted in the presupposi- tional framework of positivism and faculty psychology. . . . Understood in this way, an 'argument' is a serial predication— a thing with formal and substantive characteristicshaving objective existence apart from the perspective of

25social actors."It may be arguable that Willard is accurately criti

cizing some version of the logical perspective (although his subsequent characterization of Toulmin as a representative of a neo-positivist approach is troubling) but Willard's criticisms do not apply to my point of view. At no point is an argument to be treated as an objective entity on a par with rocks and trees. The importance of interpreting an argument in light of its context is explicitly recognized. No appeal is made to faculty psychology.


19The critic's procedure of abstracting an argument from discourse , making sense of it in light of the teachings of rhetorical theory, and evaluating it, does not require any of the peculiar assumptions implicit in what Willard loosely refers to as a positivist view of argument.

Logic and Conceptual RelationsThis ideal of deductive inference is one that our

everyday arguments can rarely, if ever, meet. The logician's usual procedure is to set a standard of argument founded on the conception of the deductively valid argument. If the arguments of law, politics, science, or morals fail to meet this standard, then these arguments cannot be accepted as valid. Such an approach is unproductive, especially since the arguments of rhetoric cannot normally meet the logician's standard. A system of standards and criteria should instead be capable of accounting for and criticizing the arguments of different fields. We must, in part, test our understanding of logic with reference to arguments as we find them. A system of strict deductive logic, if inadequate for our usual ways of arguing, must therefore be suspect.^®

The usual deductive view assumes that the meaning of propositions and concepts is given by stating their truth- conditions. Since logic studies conceptual relationships,


20logic studies the relationships of the truth-conditions of concepts and propositions. To give the meaning of a concept is to state the conditions under which the concept may be truthfully ascribed to an object, or to define a more abstract concept in terms of other concepts so defined. For example, to give the meaning of a simple noun like "horse," we state the conditions that must be fulfilled in order to call something a horse. The term "horse" is correctly ascribed only when the object to which reference is made is, in fact, a horse.^

The meaning of a proposition is, on this account, also given by stating its truth-conditions. That is, the meaning is given by stating the conditions under which the proposition is true, with the understanding that the proposition's meaning is a function of the meaning of its parts. Statements are typically analyzed as predications. For example, "Ned is a horse" is true if the property of being a horse is truthfully ascribed to Ned. The meaning of "Ned is a horse" is given by stating the conditions under which it is true. As Wittgenstein explains:

To understand a proposition means to know what is the case if it is true.

(One can understand it, therefore, without knowing whether it is true.)

It is understood by anyone who understands its constituents.28


21The truth-conditions of a singular proposition are

analyzed by producing a condition or set of conditions that, taken together, are sufficient to account for the truth of the proposition. Each of these conditions must individually be necessary for the truth of the proposition. For an example of how these procedures work we may consider a possible formulation of the truth-conditions for the proposition that "Sam knows that his son is alive." We give the truth-conditions by saying that this proposition is true if and only if:

(a) Sam believes his son is alive.(b) It is true that Sam's son is alive.(c) Sam is justified in believing that his son

is alive.29The meaning of a compound proposition is completely

determined by the truth-conditions of its constituents.This may be made clear by use of a device of Wittgenstein's whereby the truth-conditions of the various constituents of a compound proposition may be illustrated. Let us suppose that P and Q stand for any two propositions, and that expressions of the form "P or Q" mean "either P is true or Q is true or both." This expression may be defined by organizing all possibilities of truth and falsity into a table:


22 P or Q

(1) T T(2) T F(3) F T(4) F F

The proposition "P or Q" is true for lines (1), (2), and (3), and false for line (4). Thus, lines (1), (2), and (3) constitute the truth-conditions for "P or Q." If we were then to substitute English sentences for P and Q, we could define "P or Q" by stating the conditions that would have to be fulfilled in order to meet the conditions of lines (1), (2), or (3). If P is "Rose is a horse" and Q is "Ned is a donkey," then we would give the meaning of "P or Q" by stating the conditions to be fulfilled for these statements to be true. The expression "Rose is a horse or Ned is a donkey" is true, by this analysis, if either or both of its constituent propositions are true.

These principles enable the formal logician to divide propositions into tautologies, contradictions, and contingent (substantial) propositions. A tautology is a proposition that is true under all conditions. That is, all empirical conditions constitute the truth-conditions. For example, any expression of the form "P or not-P" is a tautology. A contradiction is false regardless of circumstances.


23For example, "P and not-P" is a contradiction, which asserts that both P and the negation of P are true. It is false under all conditions.

A contingent proposition is true under some circumstances and false under others. The meaning of a contingent proposition can be given by stating the conditions under which it is true. The example of a few pages back, "P or Q," is a contingent proposition.

The view that logic is based on the analysis of trutfr- conditions dominates the study of the rational assessment of arguments. The traditional view has a number of weaknesses, however, of which the most important is this.Rarely, in issues giving rise to rhetorical exigences, do we find it possible to determine, with satisfactory certainty, whether a given proposition is true or false. We must often rely on judgment and interpretation of a highly fallible order.

Pollock, who argues articulately for a revised approach to logic, discusses the criticized view as follows:

A logical connection must arise from the meanings of the concepts or statements involved in the knowledge claims. And (and here is the step which I shall deny) it has traditionally been supposed that the only way to analyze the meaning of a statement or a concept is to give its truth conditions~to say what conditions must


24be satisfied in order for the statement to betrue or for the concept to be correctlyascribed to an object.30These difficulties can best be met by a completely

different conception of logical relationships. On this view, the meaning of a concept is given by the conditions under which it may be justifiably ascribed,, or by defining it in terms of other concepts so defined. We do not give the meaning of the concept by referring to the conditions under which the term may be correctly ascribed, but by the conditions under which it may be justifiably ascribed.This approach is equally adequate for more difficult or abstract concepts; for example, we can define justice by stating the conditions under which an act may justifiably or reasonably be considered just.”̂

Futherroore, the meaning of a proposition is now to be given by stating its justification-conditions rather than its truth-conditions. We should not consider the meaning of a proposition to be given exhaustively by its truthtable. Instead, the meaning is given by the conditionsunder which it may be asserted justifiably. For example, we do not analyze the proposition that "The arms limitation treaty will reduce the likelihood of war" by stating the conditions under which it is true or false. Instead, the meaning of this proposition is given by stating the


25conditions under which it may be asserted with justification --the conditions under which we have good reason to accept it.

On the usual view of logic, conceptual relationships are analyzed by analyzing the conditions under which the truth of one proposition requires the truth of another proposition. For example, in the expression "if P then Q," we can infer the truth of Q if we know that P is true. If we know that Q is false, we know that P must also be false. But if we do not know whether P is true or not, we cannot infer anything about the truth or falsity of Q. Logical relationships. But if conceptual relationships are stipulated only by truth-conditions, then the resulting use of logic will be limited to circumstances in which we know truth or falsehood of propositions.

But if conceptual relationships are understood by justification, not by truth, this limitation does not hamper the rational assessment of arguments. It becomes possible to conceive of a system of argumentation capable of dealing with the doubtful, uncertain world of problems arising in rhetorical discourse. Certainty is no longer a precondition of rationality. We may, for example, be justified in believing Q because we are justified in believing

3 2P, even though we may not know the truth of P.


26Theory Building and the Humanities

Academicians often speak of theories of criticism, the theory of knowledge, and so forth. Less often do theorists express a notion of what theories are, how one might produce a theory, or how one should judge the worth of a theory. A theory of argumentation is an explanation of general concepts and principles of argumentation that comprehensively covers the subject of argumentation, is not restricted in time or place, and yields an understanding of argumentation.

All theory building assumes the value of generalized, systematic learning and scholarship. Many fine works of scholarship, including much rhetorical criticism and history, are concerned with the understanding of particular events.In theories, however, particular events are interpreted as instances of the concepts and principles of the theory. For example, Stelzner's interesting analysis of Roosevelt's War Message interprets and explains the speech, point by point, with little reference to general principles and concepts. Other works of rhetorical criticism are, on the other hand, founded in a theory. A widely read example is David Ling's essay on Kennedy's Chappaquidick speech, which is organized around Burke's dramatistic pentad. The more recent analysis of the rhetorical implications of a mass murder and suicide


27is a more subtle example of analyzing a particular event by

3 3Burke's theory. The last two works show concern for matters more general than the events they analyze.

The building of theories in the humanities serves three functions in the development of knowledge: layingout a systematic typology, providing explanation to yield a sense of understanding, and providing principles for the evaluation of a subject matter. These functions of theories are by no means independent; in fact a theory cannot be very useful if it fails to fulfill any of them.3^

The first of these functions, typology, has an important place in the theory of argumentation. Aristotle organized lines of argument according to the various topoi in which they cam be found. Perelman, likewise, classifies arguments into various categories: arguments of dissociation, of reciprocity, of waste, and so forth. Scholarship proceeds along the best path when based on a coherent conceptual scheme, which organizes the field of study and conveniently identifies the various phenomena and ideas under consideration. The classification sorts out those similarities and differences among the objects of study that are relevant to the theorist's interests. That is, a theory lays out a conceptual framework on which its principles are built.35


28Likewise, a theory of argumentation should make pos

sible an understanding of argumentation. The scholar should clarify the murky, make known the hidden, and organize the haphazard. Gaining a sense of understanding requires having a systematic account not only of the basic concepts of the field, but also of the relationships among those concepts. Explanation yields the principles of the theory; a theory is composed of concepts and principles that show the relationships among those concepts.

Non-scientific theories provide typology, explanation, and understanding. But most humanistic theories, certainly the best of them, perform a function unique to the humanities: they evaluate. This does not mean that the theoryimposes a simplistic, didactic morality on various events; these evaluations are often complex and subtle. The literary critic evaluates works of literature: these worksare good or bad, uplifting or ignoble; they reflect the true spirit of human nature or falsify it. The traditional logician judges arguments to be valid or invalid. The theologian judges theories of man and God as moral or immoral, true or false, acceptable or heretical. Both the concepts and principles of the theory may have evaluative connotations. Toulmin*s view of logic recognizes the importance of logic as an evaluative or critical study "concerned with the soundness of the claims we make— with


29the solidity of the grounds we produce to support them—or, to change the metaphor, with the sort of case we pre-

■ 36sent in defence of our claims."Effective theoretical evaluation is not merely a

matter of labeling the various objects of a study; certain groundwork must be laid. First, the theorist must have met the earlier functions of theories: typology, explanation, and tinder standing. Futhermore, the explanation must be general in order to be theoretical. An ad hoc explanation of some one-time sequence of events does not constitute theoretical evaluation.

The application of evaluative terms— good, bad, evil, sound, valid, just, and so forth— must be based in a theoretical typology: a system of categories must be available for the evaluation. A critic could apply particular events relating to a speech to the definitions in the category system or typology of the theory. For example, a traditional theory of argumentation might include a typology of arguments that are "formally sound," "rhetorically effective," and "true." It would also provide criteria by which these concepts can be applied to particular arguments. For example, an argument expressed as a syllogism with properly distributed terms might be evaluated as "formally sound." But the concept of "formally sound" is part of a typology of evaluative terms, and gains its significance as


30a member of the typology. Theoretical evaluation is not '•■ad hoc, but proceeds according to the defining principles laid put in the theorist's typology.

A theory is subject to evaluation both on the basis of criteria relevant to the internal structure of the theory and on the basis of criteria relevant to the success of the theory's explanation and evaluation of its subject matter.

The three principal internal criteria are consistency, completeness, and clarity. A theory is inconsistent if it contains propositions that are contradictory. Although this may seem obvious, the complexity of most theories presented in the humanities makes inconsistency not only possible but likely. Even Kant is frequently accused of presenting different and inconsistent versions of his categorical moral imperative. A theory should also be complete in its explanation; it should be capable of a thorough account of its subject matter. In the area of argumentation, Cicero is often criticized for failing to provide an account of deliberative oratory comparable to his theory of forensic argumentation. The third criterion of a theory's internal adequacy is its clarity and precision. Lack of clarity often shows up in charges that the theory is inconsistent. In other cases, lack of precision and clarity might involve inadequate definitions, the use of ambiguous


concepts, or an apparent shift of ground during the explication of a theory.

The external criteria are important but less tangible. The theory's esqplanation must be accurate and its evaluations fair. Our best determination of the accuracy of a theory is likely to be a comparison between the theory's principles and our considered judgments of particular cases the theory's principles should generalize our considered judgments. Perhaps an occasional minor clash between a theory and our judgments can be forgiven. Repeated violations call for the revision or rejection of the theory.For example, demonstrative logicians frequently rule many arguments unsound on the grounds that the arguments do not meet the rigid standards of demonstrative logic. But if the principles of demonstrative logic rule out large numbers of carefully considered arguments raised by intelligent persons, it may become necessary to revise the principles of logic. For example, as Toulmin notes, the arguments for our knowledge of a world outside our own minds do not meet the standards of demonstration. But this does not, contrary to the empiricists, establish that we have no knowledge of the world. What it establishes is that thedeductive ideal does not yield the only sound standard of

37argument. A theory of argumentation should not radically


32conflict with the standards or conclusions of particular, respected fields. Therefore, the inadequacy of a theory can be brought out by presenting examples contrary to the theory's explanation. For such criticism of a theory to be useful, it is desirable to diagnose why a counterexample can arise: where is the theory's defect? We want to knowhow a theory is defective in order to devise a better one. Furthermore, a good test of a humanistic theory is to apply it to significant problems of its field of study. It is one thing to theorize in the abstract; a theory should also apply clearly to specific problems. Rawls, for example, tests his theory of justice by applying it to the problemsof equal liberty, distribution of goods, duty, and civil

3 8disobedience. Similarly, the best test of a theory of rhetorical criticism is its successful use in the evaluation of discourse.

ProcedureThis study applies the methods of theory building to

specific problems in the field of argumentation.

MaterialsThe study examines the writings of certain authors

who address one or more of the questions raised in this study. The study is largely restricted to contemporary


33writers on the assumption that most of the better ideas of earlier figures have infiltrated the works of today's thinkers. I take up the work of early thinkers only to obtain an important insight or two that have escaped more recent figures. The majority of the work examined falls into one of these two categories:

1. The field of speech communication has produced a substantial volume of work in the field of argumentation, much of which takes the logical perspective or something close to it. Although I do not review all of this material, much of which is pedagogical, I examine works that make a serious effort to address the theoretical questions of this study. The volume of this material is greatly reduced by culling out substantially redundant work, hence I examine only a few particularly important textbooks.

2. The analytic school or epistemology has quite recently shown interest in examining the modes of inference by which we arrive at conclusions based on evidence. Although only a few writers, including Toulmin,Butchvarov, Price, Chisholm, and Pollock,


34have shown concern with problems of argumentation, the quality of their contribution and its significance is remarkable.

MethodEarlier portions of this study examine the major

approaches to the study of argumentation. Chapters two and three take up variations of the traditional approaches to the study of argumentation and logic. This approach, loosely based on Aristotelian logic, considers argumentation to be predominantly syllogistic, inductive, or enthy- mematic. Chapters four and five examine the modem views. These latter theories contribute to a contemporary theory of argumentation.

These discussions examine possible answers to the questions posed at the beginning of the chapter. As each of these answers has a particular advocate or group of advocates, laying out each answer involves discussing the stated or implicit assumptions, goal3, and methods of the different authors. The major interest is to understand how the goals, assumptions, and methods lead the author to a particular answer. The ultimate result is the development of a typology of standards and criteria. I choose a typology of standards because argumentation emphasizes the evaluation of discourse. The set of logical standards provides


35the logical foundation of the evaluation of arguments.Basing the theory's typology on the various standards of evaluation permits the most convenient and adequate system for evaluating arguments.

After being laid out, each proposed answer is evaluated in terms of two criteria, based on the earlier discussion of theory building:

1. Do a theory's assumptions and methods involve fallacies that invalidate them?Clearly inadequate reasoning often invades the work of the finest scholars, and some promising approaches can be dismissed when a serious error comes to light. The explanation may be inconsistent, or it may be confused or ambiguous. Theories are supported by arguments, and these arguments vary in quality.

2. Do the theory's assumptions lead to an explanation of argument that is inadequate?A theory may provide an incomplete account incapable of explaining all important kinds of arguments. Or, the writer's assumptions might lead to a false or unclear theory of argument. A critic who wished to attack


36Toulmin might produce examples of arguments which function differently from what his model would lead us to expect.

The third step is to explicate my own answer to each question. Having examined the strengths and weaknesses of the different answers to a given question, I argue that traditional logic is inadequate because it assumes that logical relationships must be truth-functional. The enthymeme and induction are adequate parts of a logical system, but can usefully be incorporated into the framework of a larger theory built on rigorous foundations. This will permit more systematic and precise conclusions. A comprehensive theory must be adequate to account for the multiple criteria of marketplace argumentation. These criteria can be understood in terms of (1) the types of grounds to which rhetors can appeal to justify claims and (2) a conception of logical inference derived from a more adequate theory of meaning.

This third step requires the explication of a system of standards and criteria, together with a justification for the type of system offered. This yields my answer to the principal question, "What are the logical foundations of argumentation?" Beginning in this chapter, with a general attitude toward the relation of logic and argumentation,


37

the study concludes when the answers are brought together to form an organized framework. Thus the conclusion of the study both summarizes and synthesizes the answers to the three questions, proposing the directions to be followed in producing a theory of argumentation from a logical point of view.

ConclusionThe logical perspective assumes that the fundamental

reason for making arguments is to resolve problems through communication that, at least in part, is rational. We want to find out what is true and good, or most likely to be true and good. Rhetoric is not the mere art of persuasion, nor is it the source of all our knowledge, but the arguments used in rhetoric express our knowledge, our beliefs, and the reasons for them. These arguments are subject to critical evaluation.

This study investigates the relationship between logic and rhetoric by developing the premises from which a logically based theory of argument can be derived. The thesis argues against the traditional view of deductive logic and for an improved system of greater applicability.


38Notes

Chapter I

1. C. S. Lewis, The Screwtape Letters (New York:The Macmillan Company, 1944}, p. li.

2. Daniel J. 0*Keefe, "Two Concepts of Argument,” Journal of the American Forensic Association 13 (1977): 129-32.

3. See, e.g., William Trufant Foster, Argumentation and Debating (Boston: Houghton Mifflin Company, Riverside Press, 1908), p. 143.

4. Rules of evidence are discussed in Roderick Chisholm, Theory of Knowledge, 2nd ed. (Englewood Cliffs, N.J.: Prentice-Hall, Inc., 1977), pp. 1-6; H. H. Price,Belief (London: George Allen & Unwin Ltd, 1969), pp. 92-129. On criteria, see Ludwig Wittgenstein, Preliminary Studies for the "Philosophical Investigations," Generally Known as the Blue and Brown Books (New York: Harper &Row, Publishers, Harper Torchbooks, 1965), pp. 24-25;Rogers Albritton, "On Wittgenstein's Use of the Term 'Criterion,'" Journal of Philosophy 56 (1959):845-57; Panayot Butchvarov, The Concept of Knowledge (Evanston, 111.: Northwestern University Press, 1970), pp. 274-307; Stephen Edelston Toulmin, The Uses of Argument (Cambridge: At the University Press, 1958), pp. 232-53.

5. Joseph W. Wenzel, "Perspectives on Argument," in Proceedings of the Slimmer Conference on Argumentation, ed. Jack Rhodes and Sara Newell (Falls Church, Va.: Speech Communication Association, 1980), pp. 112-133 discusses the logical perspective and analysis. See also GeorgeW. Ziegelmueller and Charles A. Dause, Argumentation: Inquiry and Advocacy (Englewood Cliffs. N.J.: Prentiee-HaIT7toEf7T^75T7^h. 1.

6. O'Keefe, "Two Concepts of Argument," p. 121.7. John R. Searle, Speech Acts: An Essay in the

Philosophy of Language (Cambridge: At the UniversityPress, 1969), pp. 29-33.

8. Wenzel, "Perspectives on Argument," p. 126.


399. Ibid., p. 121.10. Ibid.11. Walter R. Fisher, "Toward a Logic of Good

Reasons," Quarterly Journal of Speech 64 (1978) :376- 84. For similar views, see Donald C. Bryant, "Aspects of the Rhetorical Tradition~I: The IntellectualFoundation," Quarterly Journal of Speech 36 (1950):169-76 and Ray Lynn Anderson and C. David Mortenson,"Logic and Marketplace Argumentation," Quarterly Journal of Speech 53 (1967):143-51.

12. Willard Van Orman Quine, Elementary Logic, rev. ed. (New York: Harper & Row, Publishers, HarperTorchbooks, 1965), p. 1.

13. Foster, Argumentation and Debating, p. 88.14. Chaim Perelraan, Le champ de 11 argumentation

Brussells: Presses Universitaires de Bruxelles, 1970),p. 7.

15. John L. Pollock, Knowledge and Justification (Princeton, N.J.: Princeton University Press, 1974),ch. 1.

16. Donald Davidson, "Truth and Meaning,"Synthese 17 (1967):310. The essays in Gareth Evans and John McDowell, eds., Truth and Meaning: Essays in Semantics(Oxford: Clarendon Press, 1976) are relevant on this issue.

17. Ch. Perelman and L. Olbrechts-Tyteca, The New Rhetoric: A Treatise on Argumentation, trans. JohnWilkinson and Purcell Weaver (Notre Dame, Ind.: Universityof Notre Dame Press, 1971), p. 1.

18. John L. Pollock, Knowledge and Justification, gives the most precise discussion of this approach to logic.

19. Bryant, "Aspects of the Rhetorical Tradition— I," p. 173.

20. My view on these topics reflects the influenceof Karl Wallace, "The Substance of Rhetoric: Good Reasons,"Quarterly Journal of Speech 49 (1963):239-40, and Donald C. Bryant, Rhetoric: Its Functions and Its Scope," QuarterlyJournal of Speech 39 (1953);401-24.


4021. Toulmin, Uses of Argument, e.g. pp. 16, 103-104,

112.22. Charles W. Kneupper, "On Argument and Diagrams,"

Journal of the American Forensic Association 14 (1978):185.23. Anderson and Mortensen, "Logic and Marketplace

Argumentation," pp. 145-51.24. On this issue, see also Thomas B. Farrell,

"Validity and Rationality: The Rhetorical Constituents ofArgumentative Form," Journal of the American Forensic Association 13 (1977) : 146; Edward Z. Rowell, ,bProlegomena to Argumentation," Quarterly Journal of Speech 18 (1932): 592.

25. Charles Arthur Willard, "A Reformulation of the Concept of Argument: The Constructivist/InteractionistFoundations of a Sociology of Argument," Journal of the American Forensic Association 14 (1978):l’2TI

26. Stephen Toulmin, Human Understanding (Princeton, N.J.: Princeton University Press, 1972), 1:68 explains the truth”functional method very clearly; Toulmin, Uses of Argument, ch. 4 illustrates the application of the truth- functional method for epistemological problems and shows its inadequacy to the analysis of such problems. G. E. Moore, "A Defence of Common Sense" and "Proof of an External World,” both in Philosophical Papers (London: George Allen & Unwin Ltd, 195$),‘pp. 32-59 and 127-50 respectively, makes pretty much the same points.

27. The version of the deductive view that I am outlining is quite generally accepted among logicians; see e.g. Ludwig Wittgenstein, Tractatus Logico-Philosophicus, trans. D. F. Pears and B. fc*. McGuinness (London:Routledge and Kegan Paul, 1971); Rudolph Carnap, Introduction to Symbolic Logic and Its Applications, trans.William H. Meyer and John Wilkinson (New York: DoverPublications, 1958); and these three books by Willard Van Orman Quine: Mathematical Logic, rev. ed. (Cambridge,Mass.: Harvard University Press, 1955), Elementary Logic,and Philosophy of Logic (Englewood Cliffs, N.J.: Prentice-Hall, Inc., 1970). I have throughout avoided bringing in any doctrines peculiar to the positivist movement. Insofar as the issues immediately under discussion are concerned, there is virtually no controversy in these sources.


4128. Wittgenstein, Tractatus, par. 4.024.29. Definitions of this sort are evaluated in Edmund

L. Gettier, "Is Justified True Belief Knowledge?" Analysis 23 (1963):121-3. This is a good example but does not illustrate my definition of "knowledge."

30. Pollock, Knowledge and Justification, p. 8. Pollock notes that the criticized terminology derives from positivism, but that the type of argument to which he objects "is really the same sort of thing that Locke, Berkeley, and Hume, and for that matter Socrates, were doing."

31. Ibid., ch. 1.32. For a good discussion of certainty as the precon

dition of rationality, see Perelman and Olbrechts-Tyteca, The New Rhetoric, pp. 1-14.

33. David A. Ling, "A Pentadic Analysis of Senator Edward Kennedy's Address to the People of Massachusetts, July 25, 1969," Central States Speech Journal 21 (1970): 81-86; Hermann G. stelzner, b’War Message,' December 8, 1941: An Approach to Language," Speech Monographs 33: 419-37; Jeanne Y. Fisher, "A Burkean Analysis of the Rhetorical Dimensions of a Multiple Mass Murder and Suicide," Quarterly Journal of Speech 60 (1974):175-89.

34. This viewpoint on theory-building is adapted from Paul Davidson Reynolds, A Primer in Theory Construction (Indianapolis: The Bobbs-Merrill Company, 1971 withreference to Piotr Sztompka, System and Function: Towarda Theory of Society (New York: Harcourt Brace Jovanovich,Publishers, Academic Press, 1974) and Jerald Hage, Techniques and Problems of Theory Construction in Sociology (New York: John Wiley & Sons. Wilev IntersciencePublication, 1972).

35. Butchvarov, Concept of Knowledge, pp. 6-7.36. Toulmin, Uses of Argument, p. 7.37. Ibid., ch. 4; see also Chisholm on criteria,

Theory of Knowledge, ch. 7 and esp. pp. 123-32.


4238. John Rawls, A Theory of Justice (Cambridge,

Mass.: The Belknap Press of Harvard University Press,1971), ch. 4-6. Some of my ideas on the evaluation of theories draw on insights in this book, especially from pp. 3, 51-52.


CHAPTER IITRADITIONAL MODELS OF ARGUMENTATION;DEDUCTIVE LOGIC AS TRUTH-FUNCTIONAL

The traditional theories of argumentation, grounded in the logical theories of ancient and medieval times, apply traditional logic (or improvements on it) to the analysis of rhetorical arguments. An adequate theory of argumentation will have to combine into a coherent system an understanding of the criticism of deductive or conceptual arguments and inductive or exemplary arguments. The major problems are with deductive arguments. The discussion in this chapter concentrates on the often ignored fundamental assumptions underlying traditional logic, with attention to the applicability to rhetoric of a logic based on such assumptions. The inadequate theory of conceptual relationships (meaning) underlying formal deductive logic receives special attention.

The diagnosis of why traditional deductive logic falls short of adequacy is instructive for the development of new models. Although argumentation should reject exclusive reliance on formal logic, it must accept some model of conceptual inference. Analyzing the reason for the inadequacy of formal logic, its grounding in a truth-condition theory

43


44of meaning, permits the development of a superior model of conceptual inference.

Logic takes itself to be a study of conceptual relationships. The theory of conceptual relationships underlying formal logic assumes that "a knowledge of the truth- conditions of a sentence is identical with an understanding of its meaning."1 Traditional logic studies the relationships of truth and falsehood among terms and propositions.In a valid argument, the conclusion is necessarily true ifthe premises are true. Thus, traditional logic knows only

2the standard of certainty.Traditional logic is restricted to the manipulation of

propositions that are determinably true or false. Logical relations mean only a relationship among truth-values. This feature is most fundamentally responsible for the limitations of formal logic.

The literature on this subject consistently uses the arcane terminology inherited from logical positivism; this thesis uses a parallel terminology in order to ease the task of comparison with views expressed in previous literature. To the traditional logician, a proposition is a sentence that is either true or false. Thus, traditional logic is two-valued: a proposition must have either the valueof being true or the value of being false. Further, a


45proposition may be true in some circumstances and false in others. The circumstances tinder which a proposition is true are its truth-conditions. Truth-conditions, if they can be stated at all, can always be stated as a set of conditions that are both necessary and sufficient for the proposition to be true.

By such an analysis, we could take the proposition "Polly is a bird" to mean:

Polly has wings;Polly is an animal;Polly is warm-blooded;Polly is a biped;Polly has feathers.

Each of these must be true if "Polly is a bird" is true; together, the truth of all five propositions guarantees the truth of "Polly is a bird.5* This is a typical statement of truth-conditions for a simple case.

A tautology is a proposition true for all possible circumstances, since its truth-conditions are all possible conditions; a contingent proposition is true for some possible conditions but not others. The truth-conditions of a proposition are the circumstances under which the proposition has the truth-value of being true. Traditional logic is considered truth-functional because the truth of the conclusion is a function of the truth of the premises; the truth-value of the conclusion varies systematically with the truth-value of the premises.


46The syllogism, for example, is truth-functional. If

the premises are true, the conclusion is also true:All horses are animals;Rose is a horse;Therefore, Rose is an animal.

Underlying truth-functional inference is the view that a concept can be defined or analyzed only by stating the conditions under which it is truthfully applied. Logical arguments are valid because of the relationships of truth- functionally defined concepts. The above argument can thus be interpreted as giving the conceptual relationships among "Rose," "horse," and "animal." These conceptual relationships are such that if Rose is a horse she is also an animal. Relationships among concepts are defined, according to the formal logicians, truth-functionally.^

Aristotle emphasizes that it is the truth of the premises of a syllogism that necessarily yields the truth of the conclusion: "A syllogism is discourse in which, certain things being stated, something other than what is stated follows of necessity from their being so. I mean by the last phrase that they produce the consequence, and by this, that no further term is required from without in order to make the consequence necessary.Mills and Petrie, who defend the use of the syllogism in argumentation, agree that, "The validity of an argument guarantees only a conditional truth to the conclusion. If the premises


47are true, then the conclusion must be true; and the logical force of the 'must' is reflected in the logical truth of the corresponding conditional statement."5 Likewise the Port-Royal Logic, starting from the assumption that "Nothing is more to be esteemed than aptness in discerning the true from the false," states that "A syllogism is valid only if the conclusion is a necessary consequence of the premisses: If we assume the premisses to be true, thenthe conclusion must be true."®

MeaningThe critique of deductive logic, although ranging over

a large number of issues, repeatedly returns to the theme of denying the theory of meaning and conceptual relationships that underlies traditional logic. This section discusses the main themes of the traditional theory of meaning and briefly mentions the lines of suggested modification. A revised view of meaning provides the foundation for an improved understanding of logical relationships.

The Traditional TheoryThe widely accepted theory of meaning under discussion

is fully developed in the writings of the German mathematician Frege. It is no exaggeration to call Frege the founder of the modern philosophy of language; both the


48ideal language and ordinary language schools of thought owe a great debt to his work. Several of Frege's ideas have been influential on contemporary thought and provide a basis for Searle's more adequate amplification of the traditional theory. Frege's theory of meaning, or one related to it, necessarily underlies truth-functional logic.

The theory of meaning is usually considered to consist of an account of the referential function of language, the sense or propositional content of sentences, and the communicative function of speech acts. Searle argues that the speaker1s intention needs to be added to this account.

Frege's investigation into the nature of language was based on these three rules:

always to separate sharply the psychological from the logical, the subjective from the objective?

never to ask for the meaning of a word in isolation, but only in the context of a proposition;

never to lose sight of the distinction between concept and object.7Sense and Reference. The modem outcome of these

principles that has been of the greatest influence is the doctrine of sense and reference. Frege explains this distinction as follows:


49What is intended to be said by a=b seems

to be that the signs or names "a" and "b" designate the same thing, so that those signs themselves would be under discussion; a relation between them would be asserted. But this relation would hold between the names or signs only in so far as they named or designated something. . . .

It is natural now, to think of there being connected with a sign (name, combination of words, letter), besides that to which the sign refers, which may be called the reference of the sign, also what I should like to call the sense of the sign, wherein the mode of presentation is contained. . . . The reference of "evening star" would be the same as that of "morning star," but not.the sense.8

As Dummett explains, "The sense of an expression is . . . that part of its meaning which is relevant to the determination of the truth-value of sentences in which the ex-

Qpression occurs." One gives the sense of an expression by stating its truth-conditions. A proper name makes reference to some object. A word's sense can be explained by stating the conditions under which it can be correctly ascribed to an object. For example, the word "bird" can be correctly ascribed to all those objects and only those objects that are winged feathered bipeds. The word "bird" refers to these objects. Sense is the content of a sentence or expression.

Frege expands his analysis to include the analysis of declarative sentences. He argues that a sentence contains a thought, which is the sense expressed by the sentence.


50By "thought" he does not mean "the subjective performance of thinking but its objective content, which is capable of being the common property of several thinkers."*0 Frege holds the reference of a sentence to be its truth-value? a true sentence refers to the true and a false sentence refers to the false.** Few today agree that a sentence has a reference; it is Frege's discussion of the sense ofa sentence that has been most important in the history of, • 12 logic.

The sense of a sentence is usually considered to begiven by a statement of its truth-conditions. Each of theterms of the sentence is defined by giving the conditionsunder which it is truthfully applicable, and the sentence,understood as composed of constituent parts, is defined bygiving the conditions under which it is true. McDowellexplains that Frege's understanding of sense and reference"are captured by the following conception of a theory ofsense for a language:"

It assigns a suitable property to each simple sentence-constituent discerned in the language by an appropriate syntax, and states rules which determine suitable properties for complex expressions formed in each of the ways permitted by that syntax, given the relevant properties of their components; the property thus determined for a complete sentence is

- that of being true if and only if some specified condition holds. A theory of sense for a language, then, shows how to derive, for any indicative sentence, a theorem of the form


51"s is true if and only if where "s" is replaced by a suitable designation of tKe sentence and "e>" by a sentence. Briefly, a theory of sense, on this conception, is a theory of truth.13Force. A full understanding of the sentence’s meaning

requires, in addition to a theory of sense, an understanding of the force of the sentence, which requires an analysis of the speech act performed in the utterance of the proposition. We use sentences for varying purposes, such as to make commands, ask questions, or express wishes. McDowell gives his opinion on the interaction of sense and force:

A theory of force would do two things: first,license the identification of linguistic actions, given enough information about them, as performances of propositional acts of specified types (assertion, question, and so on); and, second, show how to recover, from a sufficiently full description of an utterance, which may be an utterance of an elliptical or non-indicative sentence, a suitable designation of a suitable indicative sentence. The idea is that a theory of sense and a theory of force, in combination, should enable one to move, from a sufficiently full description of a speaker's utterance, uninterpreted, to a description of his performance as a propositional act of a specified kind with a specified content, that is, a description on the pattern of "He is asserting that £," "He is asking whether £," and so dn.1^

In other words indicative sentences are taken as the paradigm case, with questions, commands, and so on to be understood as different speech acts, carrying different force, for conveying the same propositional content. Searle presents these as expressing the same propositional content:


521. Sam smokes habitually.2. Does Sam smoke habitually?3. Sam, smoke habitually! j.54. Would that Sam smoked habitually.Intention and a Full Theory of Meaning. Searle essen

tially accepts the traditional approach to the study of meaning but argues that it is incomplete. Adapting a suggestion of Grice's he suggests the addition of the speaker's intention to communicate the sentence's truth- conditions to the understanding of the meaning of the sentence.

With the addition of this notion, Searle offers a full explanation of meaning. He gives the following as necessary and sufficient conditions for understanding what it is to mean something, where £ is the speaker and H is the hearer:

£ utters sentence T and means it (i.e., means literally what he says) =

S utters T andTa) S intends (i-I) the utterance U of T to

produce in H knowledge (recognTtion7 awareness) that €he states of affairs specified by (certain of) the rules of T obtain.(Call this effect the illocutTonary effect,IE)

(b) ^“intends that i-I will be recognized invirtue of (by means of) H's knowledge ofcertain of) the piles governing (the elements of) T,16

This account of meaning says that to say something and meanit means that the speaker intends his utterance to conveyknowledge to the hearer that the truth-conditions of the proposition have been met, that the speaker intends to produce


53this recognition of knowledge by getting the hearer to recognize this intention, and that this effect will be produced under the conditions specified by the rules governing the speech act performed (questioning, asserting, or whatever).

The above analysis of meaning distinguishes the function of the speech act, which is in part to gain recognition of some state of affairs, from the propositional content of the sentence. Searle explains this distinction: In the total illocutionary act the content is the proposition; the function is the illocutionary force with which the proposition is presented.^ Throughout the thesis, discussion of meaning will be concerned with the meaning of propositions, unless the context clearly specifies otherwise.

A major purpose of the thesis is to argue that, in condition (a) of Searle's definition of meaning, it is the justification-conditions and not the truth-conditions that usually convey propositional meaning. Searle writes that: "The older philosophers are not wrong when they said: toknow the meaning of a proposition is to know under what conditions it is true or false. But their account was incomplete, for they did not discuss the different illocutionary acts in which a proposition could occur.


54I argue that it is necessary also to revise or abandon the notion that propositional meaning is given by truth- conditions. Such a revised view guides a more adequate approach to logic.

The above account of meaning is not, by any means, the\

property of any single philosophical school, or even asingle group of schools. These assumptions (in particular,the assumption that propositional meaning is given only bystating the truth-conditions) apply equally to all truth-functional logic, regardless of whether formal or informal;substantial or analytic; or quasi-mathematical; or syllo-

19gistict "Indeed, the emphasis on truth conditions in the theory of meaning stems almost entirely from the attempt to clarify the logical structure of sentences and to account for the logical relations between them."20 Logicians typically characterize deductive inferences as founded on the meanings of terms and sentences; they believe that logic is the study of necessary inference; only a theory of meaning using truth-conditions could possibly explain how meanings could yield necessary inference. This type of definition is fundamental to traditional logic: it explains why conceptual relations justify the inference that, in a valid argument, true premises yield a true conclusion.


55

A Revised TheoryPollock defines justification-conditions as fol

lows :Given a statement P, let us call the conditions under which one would be justified in believing- that-P the justification conditions of the statement P.̂ J-

He also argues for this principle of propositional meaning:Given any statement P, either the meaning of P is characterized by Qie justification conditions of P and P, or else the truth conditions of P can He stated in terms of statements whose meanings are so characterized.22

A key question for this study is whether this principle is preferable to the traditional view. But first, a few considerations clarify the new principle and show that it is an intuitively reasonable alternative to the usual view.

The justification-conditions of a proposition uniquely specify the proposition's meaning. However, stating its truth-conditions also uniquely specifies the proposition's meaning. In cases where truth-conditions can be stated, the justification-conditions are typically conditions in which one is justified in believing that the truth-conditions have been met. This new conception of meaning does not rule out the old theory, but broadens it.^

As Pollock explains:There cannot be two different concepts which are both the concept of the same kind of thing.


56And conversely, given any two different kinds of things, the concepts of these two different kinds of things are distinct (e.g., the concept of a blue thing is distinct from the concept of a red thing). Therefore, as the identity of a statement is uniquely determined by the kind of state of affairs which it states to obtain, it.is also uniquely determined by the

In particular, the concept of a kind of state of affairs is either definable or characterized by the conditions under which we would be justified in judging that the present state of affairs is or is not of that k i n d .25

However, although the old theory is not refuted, neitheris it very useful. Justification-conditions better explainthe ordinary usage and meanings of the things we say.For example, we might be able to state truth-conditionsfor some-proposition, such as "This is a bird." But, asPollock notes, an ornithologist creating such a definitionhas to know how to identify a bird, that is, how to usethe sentence "This is a bird" justifiably, before he couldbegin to investigate the nature of birds and determinetheir various unique c h a r a c t e r i s t i c s .26 to know the meaningof a concept or sentence is to know how to use itjustifiably.27

The superior power of justification-conditions to clarify the sense of a proposition is clearer on consideration of the often trivial nature of truth-conditional accounts of the meanings of various expressions. Truth- conditions often (but not always) sound rather like

of state of affairs which


57restatements of the proposition whose meaning is beingexplained. One can find instances of this kind in thewritings of great philosophers, for example:

It will be noted that the state of affairs referred to is the signification of the proposition: not its denotation. 'Mary is making pies'asserts the state of affairs, Mary making pies now, has a certain status; namely, that it is actual, that it is incorporated in the real world.*8

This account of the proposition's sense (signification) is no doubt correct, but is not likely to be of much explanatory utility. Trivial accounts of this sort are particularly common when a writer wishes to give the meaning of an expression for purposes of reductive analysis.

With this general understanding of the scene of battle, we are prepared for a full discussion of the two major weaknesses of the traditional theory: first, it createsthe epistemological dilemma of choosing between scepticism or an irrelevant reductionism; second, it cannot account at all for certain types of propositions and arguments that are of great concern in rhetoric.

Traditional Logic, Reductionism, and Truth-Conditions

Sceptical philosophers note that we claim knowledgeof, for example, the world about us, but that we have directepistemic access only to our own internal states,


58perceptions, feelings, and so forth. They then note that no such knowledge can analytically entail any knowledge about the world. In a very simple case, if it seems to me that I see red when I look at a wall, one might think that this proves that the wall is red. But this is not so, claims the sceptic, because there are many circumstances under which it could seem to me that I see red when nothing red is present: the wall may be lit bycolored lights; I may be dreaming; I may be a brain in a vat being fed false perceptions by an evil neurophysicist. Thus, a statement about my own perceptions cannot entail knowledge about the world. This argument establishes that something is wrong either with our purported empirical knowledge or with the conception of a reason presupposed in the argument. Therefore, let us examine the conception of reasons.

In Knowledge and Justification, Pollock assumes, along with most modem logicians, that logical relationships are concerned with the analysis of discourse, of language. However, he contends that philosophy has held to inappropriate standards of conceptual relationships because they have held to an inappropriate theory of conceptual relationships.29

Pollock further argues that we base different sorts of knowledge on various sources. We base our knowledge of the


59physical world on sense perception; if not for sense perception, we would have no knowledge of physical things.Likewise, our source of knowledge for the existence of and thoughts of other people's minds is the behavior of theirbodies; the source of our knowledge of our personal past

j 30is memory, and so on.We normally justify claims to knowledge by appealing

to one or more sources of knowledge. The basic problem ofepistemology is the necessity of producing a proof thatjustifications of the types that we can offer do in factjustify the claims that we make: "Thus, if we cannotjustify our customary grounds for knowledge claims, then

31we cannot take them as justifying our claims to knowledge.Two methods might justify the grounds for knowledge.

We might offer an inductive justification which establishesthe reliability of some source of knowledge, or we mightoffer a non-inductive logical connection between a claimand a source of knowledge.

This obscure point might be clearer if we look at somesyllogisms:

(1) All horses are animals;Rose is a horse;Therefore, Rose is an animal.

In this, as in all valid syllogisms, the conclusion's keyterms, "Rose" and "horse," are found in the premises. Now,let us consider an invalid syllogism:


60(2) All horses are animals;

Rose is a horse;Therefore, Rose is a loving creature.

This syllogism is invalid because a term, "loving creatures," appears in the conclusion that does not appear in the premises. To make a valid, truth-functional argument, we can formulate and confirm a generalization, that "All animals are loving creatures." This is the inductive generalization approach. Or, we can reduce the terms used in the conclusion to the terms used in the premises. That is, we can say that "is a loving creature" means the same as "is an animal." In this latter case, the claim that "All animals are loving creatures" is considered true a priori.

Regardless of whether we add an inductive generalization or a reductive definition to syllogism (2); valid syllogisms can now be constructed:

(3) All horses are animals;All animals are loving creatures;Therefore, all horses are loving creatures.All horses are loving creatures;Rose is a horse;Therefore, Rose is a loving creature.

The reductive method is not always absurd, as this example might make it seem. For example, consider this syllogism:

(4) Rover is a dog;All canines bark;Therefore, Rover barks.

This invalid syllogism can be salvaged merely by an argument


61showing that all dogs are by definition canine. Traditional epistemology has attempted both the inductive and reductive methods to bridge the gap between justifications and claims.

Inductive justifications will often, but not always, do the job. I might inductively check the reliability of my memory of a certain period of my past experience by referring to documents, photographs, and other artifacts dating from the same period. However, an inductive justification only works if there is some relevant source of knowledge other than the one being tested.

In certain important kinds of cases, other sources arenot available. For example, there is no knowledge of thephysical world that does not, at least indirectly, resultfrom sense perception. It is true that I know some thingsabout physical objects because I have been told about them,or remember them, or have read about them, but knowledgefrom such memories or oral or written testimony can all betraced back to sense perception. A similar problem existsfor knowledge about other minds; we only know about otherpeople's minds because of what is observed of other bodies.In cases of these sorts, no inductive generalization can be

3 2produced.The other alternative is a non-inductive logical connec

tion, which "must arise from the meanings of the concepts


6233or statements involved in the knowledge claims." Pollock

argues that the traditional view requires that such a logical connection be established with a reductive analysis:

Starting from the truth conditions of a statement, we could never establish a logical connection between that statement and the source of knowledge which is supposed to yield the statement unless those truth conditions were stated in terms of the same concepts as are used in describing the source. Thus, for example, we could never establish a logical connection between perception and statements about the material world unless we could state the truth conditions of the latter in terms of the concepts used in describing perception.34

In other words, a reductive analysis requires us to say that statements about the physical world must be reduced to statements about our perceptions of the world; statements about other minds must be reduced to statements about behavior, and so forth. For example, the statement "Mike is in pain" might mean "It seems to me that I hear soundsexactly like screaming," "It looks to me exactly as if

35there is a grimacing face," and so on. It is taken to be a logical truth that, when we talk about the physical world, we are merely talking about our own perceptions— our own mental states and experiences.3®

The underlying problem here is, in cases where induction does not work, how do we get from our reasons to a conclusion? How do the grounds put forth to justify a claim actually support that claim? The reductive procedure


63solves this problem by abolishing the difference between the grounds and the claim. In Toulmin's terminology, reductive analysis such as the phenomenalist philosophy of mind resolve the type-jump in which reasons of one logicaltype are offered as grounds for conclusions of another

37logical type. This procedure unquestionably provides alogical connection between our claims to knowledge and thejustifications that we can offer for those claims, but theexplanation is quite unsatisfactory.

With regard to our knowledge of the future, whichToulmin points out can only be grounded in our knowledge ofthe past, a reductive analysis must claim that a statementabout the future can be reduced to, i.e. is equivalent to,statements about the past. Toulmin examines this typical(imaginary) reductive explanation:

When we say that we have reason to believe that there "will be" a lunar eclipse tomorrow night, all that this statement really means is that corresponding regularities have been apparent in the past record of lunar eclipses, etc., etc.38

Toulmin responds to this position:In this way, though at the cost of some paradox, a step from retrospective "grounds" to predictive "conclusion" is reinterpreted as a simple logical transformation. The supporting grounds in this argument are explicitly about the past; the conclusion also is now said to be implicitly about the past. As a result, the step from the past records to the misleadingly future conclusion becomes a purely formal one--like


64

the arithmetical step from a collection of data about all the individual inhabitants to a conclusion about the "average Chicagoan."39

However, this is not what we want at all; when we make a statement about the future, it is because we are interested in the future; when we justify claims about the physical world, it is because we want to make claims about the physical world, not mere claims about sense perception.^0

These various problems are about our ability to provide reasons for a wide variety of claims. Unfortunately, the traditional analysis has no further alternative to offer except for a thoroughgoing scepticism: that we canoffer no justification whatever for many of the things that we would like to justify. Toulmin concludes an analysis of such problems:

If one follows Hume, one ends by allowing the Court of Reason to adjudicate only in cases where analytic arguments can properly be demanded: ethical and aesthetic arguments, predictive and causal conclusions, statements about other minds, about material objects, about our memories even, fall in turn before the philosophers' criticism, and we find the judicial function of the reason progressively more and more restricted.*1Therefore, what is needed is an account of logical

reasons that (1) recognizes that we ground conclusions of one type in reasons of another type, that we cannot always state our reasons in the same terms as we can state our conclusions, but that (2) there is,


65nonetheless, a logical connection between many such claims and the reasons that can be offered for them. If we retain the notion that we look to the meaning of propositions, to the conceptual relationships between reason and claim,we must look for a better way to conceive of the meanings

. . 4 2of propositions.

Rhetoric and Truth-Functional Logic The following discussion attempts to establish that it

is problematic or questionable, at best, to attempt to state truth-conditions for many types of propositions which a critic might abstract from a body of discourse. This provides grounds, partly independent of those just presented, for rejecting the traditional theory of meaning and its accompanying view of logic.

One of the most difficult problems in formulating a foundation for the theory of argumentation from a logical point of view is posed by the large variety of problems to which the rhetor must speak. The rhetorician must be prepared to deal immediately with a vast range of propositions; simple examples only provide hints as to how to discuss the more complex, ambiguous arguments found in rhetoric.43


66

Propositions of FactGreat volumes of discourse make use of propositions of

fact. For example, Stevenson's speech on the Cuban missile crisis advocated a proposition of fact, that Russian missiles had been based in Cuba, and supported this proposi-

44tron with numerous other propositions of fact. Furthermore, propositions of fact are often adduced to support propositions of value. For example, to prove that a car is a good car, one might argue that it uses fuel slowly, that it will not collapse in a minor accident, that it has a powerful engine, and so on.

Traditional theories of logic are better equipped to analyze propositions of fact than any other kind. Our usual concepts of truth and falsehood are most clearly relevant to propositions of fact. For instance, the question "Is it true that President Reagan is relaxing environmental restrictions on the burning of coal?" is clearly meaningful, and we would often like to be able to answer such questions. We certainly expect to encounter no special difficulty in explaining the circumstances under which this proposition is true. However, logical connections among such propositions can be problematic. Many arguments using propositions of fact claim their conclusions tentatively or with reservations; traditional logic cannot expect to account for such arguments.45


67Propositions of Value

A proposition of fact is true because of its accuracy in stating circumstances as they are. It is difficult to imagine how such a claim could be made of a proposition of value; certainly, such a claim cannot be easily maintained for a proposition of value that is legitimately controversial. Only in unusual cases could it be said that a proposition of value states circumstances as they are. An argument could be made that it is meaningful to call some value propositions true or false. However, at least some value propositions are neither true nor false, although they can be justified by discourse. I am not fully convinced that the concept of truth is useful for characterizing any value propositions. It is, however, possible that the concept of truth as the term is normally used is not crystal clear, and we may have to live with some ambiguity in the borderline cases. The analysis of value propositions by stating their truth-conditions is problematic and therefore truth-functional logic is inapplicable to them.

Amoral propositions of value indicate evaluations of whether something, someone, some event, or whatever is good or bad without any implication about its being right or wrong. For example, we might ask if some particular car is a good car. Or we might claim that our present economic


68policy is poor without claiming that the policy is unmoral* Additional arguments would be needed to show that an ineffective policy is also immoral.

It might make sense to ask such questions as "Is it true that this is a good car?" or "Is it true that this is a good policy?" If objective criteria are available for good cars, such a question might be answerable; truth- conditions might be given. Such criteria would obviously be relative to particular persons, circumstances, and uses, so that one might need to ask, in order to be precise,"Is it true that this is a good car for a college professor to drive to work?" In these cases, however, one notes a certain awkwardness in the use of the word "true." As we progress through the categories of propositions, this awkwardness will become more pronounced and the reasons for it easier to articulate. For now, it can simply be noted that people differ in their evaluative judgments, and that objective criteria for resolving such differences are not necessarily available. Stating the conditions that would make such propositions true will be more problematic yet.

All moral propositions of value are not subject to claims of truth and falsehood at all. In cases where there is no disagreement, it might appear prima facie that a claim of truth could be meaningfully upheld. If someone


69asked me (perhaps a child would ask such a question after reading of some horrible crime), "Is it true that genocide is wrong?" I would probably say yes.

Suppose, however, that we deal with a moral issue that is, because of the very nature of the problem, ambiguous and difficult. It would seem dogmatic in the extreme to assert that the proposition "Abortion is wrong" is true or that it is false, in the same way that one would confidently assert that Chicago is in Illinois. One can indeed advance reasons for a moral judgment, and the theorist of argumentation should be able to set forth criteria for evaluating moral arguments. However, such propositions are not true or false; instead, they are good or bad, justified or unjustified.

For the sake of argument, let us suppose that Weaver is right that there are fundamental moral truths.46 But if he is, a broad range of problems of application still remain. For instance, even if the moral rule that "murder is wrong" is true, there remain difficult questions of the sort that occupy law courts: "Was it wrong for Mr. A tokill Mr. B?" Questions of this sort can often be answered only with great difficulty, even if all necessary information and moral truths are available. This type of problem led Kant to some of the conclusions in his essay on judgment:


70If it [logic] sought to give general instructions how we are to subsume under these rules, that is, to distinguish whether something does or does not come under them, that could only be by means of another rule. This in turn, for the very reason that it is a rule, again demands guidance from judgment. And thus it appears that, though understanding is capable of being instructed, and of being equipped with rules, judgment is a peculiar talent which can be practised only, and cannot be taught.47

To claim that value propositions that apply moral rules are not definable by truth-conditions may be to belabor the obvious. In the case of propositions to which concepts of "true" and "false" do not apply, a truth-functional logic is questionable.

Propositions of PolicyIf stating truth-conditions for a proposition of value

is problematic, it is inconceivable for a proposition of policy. What could it possibly mean to say, "It is true that the United States should lower its import tariffs?"In such cases, no one who understands the English language could reasonably assert that the proposition given in answer is true or false. If Mary asked me if it was true that John had been an unfaithful husband, and if I knew enough about John, I could say yes or no. But if she asked me if it was true that she should divorce him, I could only say that it depends on what she wants. Such a question


71cannot be answered "true" or "false." However, justifications can still be given and policy questions like this can be decided rationally.

The significance of policy questions to rhetoric is beyond dispute. Countless speeches are devoted to attempts to convince an audience to agree to a policy decision. Policy decisions can be made at a personal level ("Should we go to the movies tonight?" or "Should we get married?"), or at any level appropriate to public decisions ("Should Virginia adopt the death penalty?" or "Should the United States adopt a program of national health insurance?").Many of the circumstances in which the need for persuasion arises are those in which a policy decision is being disputed. A logic that is not adequate to public policy decisions is not adequate to rhetoric. A proposition of policy might be the logical conclusion of a speaker's arguments, or one proposition of policy might be used to support a more comprehensive (or a subordinate) policy. Yet, the traditional theory is largely incapable of analyzing arguments using such propositions.

Traditional logic is of limited applicability to the problems of everyday arguing because it is truth-functional and founded in an inadequate theory of meaning. The syllogism gives conclusions with certainty or not at all.


72As Locke remarked, "He that will not eat til he has demonstration that it will nourish him; he that will not stir til he infallibly knows the business he goes about will succeed, will have little else to do but sit still and perish."48

Propositions of fact, value, and policy are all essential to rhetoric. Understanding these in terms of truth- conditions is controversial for propositions of fact, futile for propositions of value, and impossible for propositions of policy. Any logic founded on the assumption that propositions are to be analyzed only by their truth- conditions must be inadequate to the full scope of rhetoric.

Are There Alternate Criteria of Evidence?The syllogism reflects the importance of inference that

is, in some sense, conceptual. Any sample of reason-giving discourse reveals, for example, numerous cases of propositions inferred from other, general propositions. The occasional calls of Bacon's modem followers for a purely inductive theory of argumentation ignore the wide significance of deduction.

Basing his analysis on a few brief passages in Wittgenstein's Blue Book, Butchvarov asks what model of conceptual inference might meet needs such as those outlined in this chapter.48 The object of Butchvarov's search is


73criteria for evidence other than the demonstrative criterion.

A first-order criterion of evidence is a criterion employed inr for example, everyday inference: Sam's holding his cheek is a criterion for saying that he has a toothache; if test paper turns blue, this is a criterion for saying that an acid is present. A second-order criterion of evidence is a criterion for calling something evidence; something counts as evidence if it meets a second-order criterion. From the standpoint of the traditional logic, there is only one first-order criterion: demonstration. A non-inductive argument is sound if and only if it is a demonstration. Likewise, there is only one second-order criterion: something is evidence for a claimif and only if it entails that claim in a demonstration.

The arguments raised in this chapter establish that this is too limited a conception of logical reasons. What we want is to identify and characterize non-inductive logical reasons that are not demonstration. Marketplace arguments use multiple first-order criteria,50 Many of these arguments are neither sound truth-functional deductive arguments nor sound inductions, yet they are, presumably, often sound. By what conception of logical reasons are these sound arguments? Is there a way to conceive of logical


74reasons by which such marketplace arguments are logical reasons? 0rr since such arguments are (for the reasons discussed earlier in this chapter) not sound by the conventional standards of logic, do we need to conclude that they are sound only when judged by some extra-logical standard? Or worse, do we concede the sceptic's case against such arguments?

If there are to be first-order criteria for evidence adequate to justify rational belief, but falling short of necessary inference, two alternatives are available: tofind a general second-order criterion of evidence, of which demonstration is but a single instantiation, or to find one or more additional second-order criteria of evidence. Butchvarov's search focuses on multiple second-order criteria, of which demonstration is one. These multiple

51criteria must not equivocate on the meaning of "evidence."Butchvarov arrives at a hypothesis called the "thesis

of nonequivocating multiple criteria of evidence."His thesis claims that there can be many criteria of evidence; such a view is essential to any model of substantial arguments: otherwi.se we are limited to the restrictive andinadequate criterion of demonstration. Butchvarov assumes that criteria of evidence, at least to some extent, grow out of the canons of evidence found in various fields, and


75are not based on a single preconceived standard.52 His thesis would salvage such canons from the charge that they are not logical.

Butchvarov reviews various possible models of nondemonstrative inference, applying a reductio to the possibilities. Butchvarov considers various forms of deductive relationships: material implication, disjunction, and soforth. None of these is suitable, which should be of no surprise since all are derived from formal logic. The relationship between grounds and claim is not one of genus and species, which is the pattern of the categorical syllogism. The relationship is not synthetic a priori, a variety of inference that few philosophers have ever accepted. Nor can the relationship be one of logically sufficient conditions, for then it would again be demonstration.As Butchvarov notes, the relationship that we want is not formal or analytic. However, it must be a conceptual relationship because we are seeking a deductive pattern of inference. Butchvarov finds no conventional model of conceptual relationships that accomplishes the desired purpose.This leads him to conclude: "The required relation wouldbe noninductive, conceptual, nonformal, and nonanalytic, and even weaker than material implication. I suggest that the opacity of the view that there is such a relation is unsurpassed in the history of philosophy."53


76Butchvarov's conclusion is not surprising, but it is

disturbing. It is easy to dismiss formal logic from the theory of argumentation. We instead demand a model of substantial logic. What does such a model look like? The obvious answer is that it is a model that is non-formal, non-analytic, and non-necessary. But this merely tells us what the model is not, not what it is.^

A logical theory of argumentation is possible only if there are criteria of evidence other than demonstration.It is not enough to hypothesize that some scheme of inference must exist in order for us to come to certain kinds of conclusions. Instead, Butchvarov insists, we must have a reason for other forms of inference to be considered sound, and we must be able to explain, categorize, and account for such forms. It seems entirely reasonable that some othermodel must exist. But what is it?

Several positions advocated in the speech communication literature appear to offer possible solutions to this quandary. These views hold that an argument is sound if its

(1) Is pragmatically effective in moving an audience.

(2) Is accepted in a communication context that enforces sound procedures; e.g. a rational enterprise, a dialogue, or a field of argument.


77(3) Is sound according to evaluative cri

teria developed in a communication context that enforces sound procedures,

(4) Is probably true.55The final possibility is that good reasons are logical reasons, but that the traditional view of logical reasons

Cgneeds to be expanded. This view holds that logical reasons are not necessarily either deductive reasons (in the traditional sense) or inductive, but that there is a third type of logical reason. The third type emcompasses much or all of the class of what Toulmin calls substantial arguments. This view, developed in later chapters, builds a broader logic on the revised theory of meaning.

ConclusionTraditional deductive logic, long presented as the

paradigm of good reasoning, is simply not adequate to evaluate most rhetorical arguments. Many well-founded conclusions cannot be demonstrated. Further, the application of truth-functional logic to rhetoric is questionable: for many propositions of value and policy, it is paradoxical at best to ask for truth-conditions. Yet, traditional logic offers no model of conceptual inference that might fill this void.


78

Notes Chapter II

1. Rudolph Carnap, Introduction to Symbolic Logic and Its Applications, trans. William H. Meyer and John Wilkinson (New York: Dover Publications, 1958), p. 15.

2. ibid. See also, for example, Willard Van Orman Quine7"~fflementary Logic, rev. ed. (New York:Harper & Row, Publishers, Harper Torchbooks, 1965), ch. 2;S.P. Barker, Induction and Hypothesis: A Study of theLogic of Confirmation (Ithaca, N.Y.: Cornell UniversityPress, 1957), p. 3; Brian Ellis, "Epistemic Foundations of Logic," Journal of Philosophical Logic 5 (1976):202.

For purposes of this chapter only, the term "traditional logic" will refer to the generally accepted conceptions of deductive logic, including the syllogism, symbolic propositional logics, and related approaches. This convention will enable the discussion to avoid various circumlocutions. There is, in fact, no consistent usage for naming the type of logic under criticism in this chapter. Some writers use the expression "formal logic," which gives the false impression that form is the principal feature under attack. The term "deductive" is also not satisfactory; this thesis argues for a greatly revised version of deductive logic. "Syllogistic," "propositional," and "symbolic" logic are too narrow. Calling this logic "truth-functional" begs the question." The expression I have chosen is artificial, but neutral with regard to the issues.

3. More precisely, a concept's meaning can be stipulated by explaining how it affects the truth-value of the proposition'in which it is used. My examples presuppose simple propositions such as "X is an a" in which the concept "a" is ascribed to the object "X". “

t

4. Aristotle, Analytics Priora, trans. A.J. Jenkin- son, in W.D. Ross, ed., The works of Aristotle 1 (Oxford;At the Clarendon Press, DZ8)', 24blfl.


5. Glen E. Mills and Hugh G. Petrie, "The Role of Logic in Rhetoric," Quarterly Journal of Speech 54 (1968): 263.

6. Antoine Arnauld, The Art of Thinking: Port-Royal Louie, trans. James Kickoif and Patricia James (Indianapolis: The Bobbs-Merrill Company, Inc., 1964),pp. 7, 177.

7. Gottlob Prege, The Foundations of Arithmetic:A Logico-Mathematical Enquiry into the Concept of Number, trans. J. L. Austin (New York: Philosophical Library,1950), p. xe.

8. Gottlob Frege, "On Sense and Reference," trans.Max Black, in Peter Geach and Max Black, eds., Translations from the Philosophical Writings of Gottlob Frege (Oxford: Basil Blackwell, 1$60), pp. 56-57.

9. Michael Dummett, Frege: Philosophy of Language,2nd ed. (Cambridge, Mass.: Harvard University Press, 1981),p. 89.

10. Frege, "On Sense and Reference," pp. 61-62.11. Ibid., p. 63.12. Cf. C. I. Lewis, An Analysis of Knowledge and

Valuation (La Salle, 111.: 6pen Court, l9*7l), p. 53 £or aposition similar to Frege's. Fortunately, this matter hasno relevance for the thesis, which will not mention reference again.

13. John McDowell, "Truth Conditions, Bivalence and Verificationism," in Gareth Evans and John McDowell, eds., Truth and Meaning: Essays in Semantics (Oxford: Clarendontress, 19^6), p. 42.

14. Ibid., p. 44. See also Lewis, Knowledge and Valuation, p. 49.

15. John R. Searle, Speech Acts: An Essay in thePhilosophy of Language (Cambridge: At the UniversityPress, 1969), p. 22, 31-33.

16. Ibid., pp. 49-50.17. Ibid., p. 125.

with permission of the copyright owner. Further reproduction prohibited without permission.

8018. Ibid.19. See Stephen Edelston Toulmin, The Uses of

Argument (Cambridge: At the University Press, 1958), pp. 135-41 on these distinctions.

20. Roger Scruton, "Truth Conditions and Criteria," The Aristotelian Society, supp. 50 (1976):200; cf.Crispin Wright, Truth Conditions and Criteria," The Aristotelian Society, supp. 50 (1976):219.

21. John L. Pollock, Knowledge and Justification (Princeton, N.J.: Princeton University Press, 1974),p. 6.

22. Ibid., p. 18. The last phrase in this definition allows for possibilities in which one cannot justify a complex proposition, e.g., "It is raining but nobody knows that it is raining." Although the compound proposition has no justification-conditions--no one could be justified in believing it— it can be analyzed into two statements whose meaning can be characterized by their justification- conditions.

23. See Ibid., pp. 17 ff.; Ellis, "Epistemic Foundations of Logic;" Panayot Butchvarov, The Concept of Knowledge (Evanston, 111.: Northwestern University Press,1970), pp. 294 ff.

24. The incorrect theory holds that the only way to state propositional meaning is to state truth-conditions and that no other theory preserves traditional logic. It is interesting to note Ellis' proof that traditional logic can be salvaged with the new conception, "Epistemic Foundations of Logic."

25. Pollock, Knowledge and Justification, pp. 17-18.26. Ibid., p. 13.27. See e.g. Ibid., pp. 12-14.28. Lewis, Knowledge and Valuation, p. 51. Lewis

uses " signification,J to correspond with Frege's "sense" and "denotation" to correspond with reference.

29. Pollock, Knowledge and Justification, p. 8. The foregoing discussion follows chapter one of this work.


8130. A later chapter will discuss such sources of

knowledge more thoroughly, but this brief statement will do for the present.

31. Pollock, Knowledge and Justification, p. 7.32. Stephen Toulmin, Knowing and Acting: An

Invitation to Philosophy (New York: Macmillan PublishingCo., Inc., 19^6), pp. 106 ff.; Pollock, Knowledge and Justification, pp. 7-8. My presentation differs slightly £rom that of either of these sources, in that I have not argued for the more extreme claim that inductive generalizations of this sort are never possible. Such an argument is both questionable and unnecessary.

33. Pollock, Knowledge and Justification, p. 8.34. Ibid., p. 8.35. C. I. Lewis presents a fully developed version

of phenomenalism in Book II of Knowledge and Valuation, esp. p. 249. Lewis holds that the meaning of empirical statements is that they are predictions of possible sense experience.

36. In Toulmin's terminology, reductive analyses such as the phenomenalist philosophy of mind resolve the type-jump between backing and claim. Phenomenalism simply defines the type-jump out of existence. See Knowing and Acting, p. 131; Uses of Argument, pp. 229 ff. Note that a reduction turns an apparently substantial argument into an analytic argument.

37. See Toulmin, Knowing_and_Acting, p. 131; Uses of Argument, pp. 229, 30.

38. Toulmin, Knowing and Acting, p. 130.39. Ibid., p. 130.40. Ibid. A more technical, and conclusive, argument

against phenomenalism is that of Roderick M. Chisholm, "The Problem of Empiricism," Journal of Philosophy 45 (1948): 512-17.

41. Toulmin, Uses of Argument, p. 175.


8242. Simply to clarify a difference in terminology that

might look like a difference in substance, let us examine Toulmin's claim that many arguments of the sort under discussion are not logical because the conclusion says more that what is in the premises; see e.g. Knowing and Actingp. 132; Uses of Argument, p. 125. It is clear that Toulmin is restricting his use of the term "logical1' in these contexts to what I have called truth-functional logic, and that the conclusion goes beyond the premises because the conclusion is not stated in the same terms as the premises: that the argument does something more than assert identity of definition.

43. Ray Lynn Anderson and C. David Mortensen, "Logic and Marketplace Argumentation," Quarterly Journal of Speech 53 (1967):143-151. ------- -------------

44. Adlai E. Stevenson, "A Premeditated Attempt: TheBuilding of the Sites," in Wil A. Linkugel, R. R. Allen, and Richard L. Johannessen, eds., Contemporary American Speeches: A Sourcebook of Speech Forms and Principles,2nd ed. (Belmont, Ca.s Wadsworth Publishing Company,Inc., 1969), pp. 87-93.

45. Toulmin, Uses of Argument, p. 154.46. Richard M. Weaver, Ideas Have Consequences

(Chicago: The University of Chicago Press, 1944),pp. 130-131.

47. Immanuel Kant, Critique of Pure Reason, trans. Norman Kemp Smith, unabridged ed. (New York: St. Martin's Press, 1965), B172.

48. John Locke, An Essay Concerning Human Understanding, ed.-Alexander Campbell Fraser (New York:Dover Publications, 1959, 2: ch. 14, par. 1.

49. Butchvarov, Concept of Knowledge, part IV;Ludwig Wittgenstein, Preliminary Studies for the "Philosophical InvestigationsGenerally Known as the Blue and Brown Books (New York; Harper & Row, Publisners, Harper tforchbooks, 1965, p. 24; Rogers Albritton, "On Wittgenstein's Use of the Term 'Criterion,'" Journal of Philosophy 56 (1959):845-857.


8350. I have in mind here the criteria of sound argu

ments of the sorts seen in any argumentation textbook, such as those relating to expert and eyewitness testimony, historical precedent, etc.

51. Butchvarov, Concept of Knowledge, p. 273. The ensuing discussion follows pp. 273-94.

52. This is a principal thesis of Toulmin, Uses of Argument, e.g. pp. 30-35; Stephen Toulmin, Human Understanding (Princeton, N. J.: Princeton University Press, 1972), 1:41-52.

53. Butchvarov, Concept of Knowledge, p. 294.54. Butchvarov also discusses Toulmin's view,

although very briefly. See Ibid., pp. 279, 303-5.55. These are schematics of the positions to be

examined, not precise statements of them.56. Butchvarov discusses some of these possibili

ties, but his comments are too brief to be enlightening.


CHAPTER IIITRADITIONAL LOGIC:

ENTHYMEME AND INDUCTION

Rhetoricians in general have not favored the use of strict deductive logic, but rather the use of enthymeme and induction. These are so clearly relevant to rhetoric that a fairly cursory discussion in this chapter suffices. Rhetoricians adapt traditional deductive logic in their conceptions of the enthymeme. Some modem views of the enthymeme, however, fail to offer logical solutions to the logical problems posed in the previous chapter. Other, more logically-oriented views provide a better basis for deductive arguments in a refined theory of argumentation.In contrast, induction has generally weathered assaults on its value.

The Enthymeme and Probability A few advocates still endorse the use of formal logic

to evaluate arguments abstracted from rhetorical discourse.'*' The enthymeme, the heart of Aristotle's system of rhetoric, underlies most modem theories based on Aristotle's work. Reliance on formal logic is all the more ironic when we recognize that Aristotle gave the central place in rhetorical theory to the enthymeme. Rhetoric requires a logic

84


85that does not demand certainty in its standards, that applies to human values and decision-making, and that provides direction for the criticism of all of the types of arguments relevant to the resolution of problems through rhetoric.

This section investigates the adequacy of the enthy- roeme as a model for argumentation by considering arguments from probability and then evaluating the enthymeme's contribution to the theory of argumentation. This study does not enter the debate over what Aristotle really thought about the enthymeme but only examines those accounts of the enthymeme that might have something to offer to the logical perspective on argumentation.

Aristotle showed the logical potential of the enthymeme by distinguishing between valid and sham enthymemes. He makes this distinction on logical grounds: the sham

4enthymemes are faulty even though they may be effective.The enthymeme is frequently construed as a syllogism founded

Con probability. It is, furthermore, a syllogism intentionally adapted to rhetoric.

The literature on the enthymeme follows three different directions. The first is to apply logical standards other than that of absolute necessity; in other words, to apply multiple criteria of evidence. The second direction is to develop a theory of rhetorical argument. Roughly speaking.


86this view attempts to develop standards that are partly rhetorical, which make effectiveness one of the criteria of soundness. The third view distinguishes enthymemes from syllogisms only by their form of expression.

The Enthymeme as Probable ArgumentAccording to the first view, enthymemes arise out of

rhetorical situations involving conflict among persons.The enthymeme is founded on the contingent rather than the certain; it contrasts with demonstration because its premises are probable.

This view provides the framework for an answer to the problem of multiple criteria; an argument is sound if it is probable. But this rather vague answer, if it is to succeed, requires an account of the criteria by which a claim can be judged probable as well as an explanation of why these criteria are sound: the significance of "probability"must be explained in terms of the way a probability claim can be justified.

Despite Aristotle's qualification that the enthymeme can be based on infallible signs, the enthymeme is traditionally associated with arguments of probability. Mudd, for example, claims that the fundamental distinction between the syllogism of logic and the enthymeme is "the degree of certainty of the premises."6 The enthymeme, by


87this view, is a deductive inference founded on probable premises.

Aristotle holds that a probability in the enthymeme "is a generally approved proposition: what men know tohappen or not to happen, to be or not to be, for the most part thus and thus, is a probability, e.g. 'the envious hate', 'the beloved show affection.'"^ McBumey interprets a probability to be a principle or fact "already probable, whose application to particular phenomena accounts for their probability."® But the literature does not yield a clear analysis of what this means.

Probability as a substitute for certainty. Rhetorical theory's traditional concern with probability arises from a realization that rhetoric cannot hope to meet the standards of evidence that philosophers would like to apply to themselves. The arguments of rhetoric are not to be criticized

9by the standards of apodeictic proof. Thus, rhetoric relies on probable arguments. Rhetoricians might want to "energize truth," but in practice, few matters are susceptible to such a high standard. So, rhetoric settles on "as high a degree of probability as the combination of subject, audience, speaker, and occasion admits."^-® Rhetoric often settles on probable arguments, not because they are desirable, but because certain and necessary arguments are not


88available, and we have to have something. This something is tagged with the label "probability.”

Rhetoric, on this view, must settle for probability as a substitute: probability is second best; certainknowledge would be better if we could find it. But since we cannot find certainty, we take what we can get. Probability usually enters theories of argumentation in just that way: not in a theory of probability or a clear explanation of fundamental argumentative standards, but merely as a contrast with certainty. Ehninger, for example, claims that the politician, debater, lawyer, and philosopher are "cut off from the resources of empirical verification on the one hand, and of formal demonstration on the other;" in order to lay out their arguments as a case, they must assemble such arguments as make their claims "appear probable."1'1' Brentlinger likewise calls probability a "substitute for certitude" and holds that probability is a "characteristic of propositions, the value of which is merely

12conjectured." Similar contrasts fill the literature on the problem of probability in argumentation. The definition of probability is usually negative; many writers do not discuss what characteristics or conditions define "probability."


89Field-dependence. From such approaches we can infer,

perhaps, the understanding of probability as applicable to propositions that are not certainly proved to be true, but for which one has rational grounds. Presumably these grounds consist of the evidence of testimony, perception, statistics, memory, and so forth, described in most argumentation textbooks: in other words, the varying criteriaof marketplace argumentation.

All that one needs to ask of this traditional literature is for it to (1) provide a rigorous account of the sense in which these multiple criteria constitute evidence or logical reasons and (2) provide a typology and explanation of these criteria.

As Butchvarov noted, conventional theories of logic do not provide an account of how non-deductive arguments can constitute evidence. It is an essential task for the theorist of argumentation to explain how multiple criteria can justify a conclusion. Once this task is accomplished, we can build a theory of argument on such a new conception.

Enthymeme as Audience-Centered ArgumentIn order to understand an argument, it is important to

look beyond the content printed in the transcript of a speech to a comprehension of the argument's development in a rhetorical situation. This is a principal contribution


90of rhetorical theory to the task of evaluating arguments.

Lloyd Bitzer's article "Aristotle's Enthymeme Revisited" proposes that the enthymeme is a joint construction of speaker and audience, while Farrell proposes a concept of "rhetorical argument" that shows many similarities to the enthymeme. The audience-centered logic seeks to account for arguments on probable matters and values. This contemporary interpretation of the enthymeme arises from the rhetorical perspective. Farrell seems to view audience-centered logic as a substitute for conventional logic. This approach gives a valuable and instructive emphasis on the relationship between argumentation and the audience.

The construction of arguments from the beliefs, opinions, and values of the audience is, of course, nothing new. McBumey's interpretation of Aristotle mentions this idea:

I submit that in Aristotle's rhetorical system the enthymeme is the element or unit of all persuasive discourse. The admission of "emotionally loaded" terms and propositions is in fact one of the important characteristics of the enthymeme; the premises which compose an enthymeme are usually nothing more than the beliefs of the audience which are used as causes and signs to secure the acceptance of the propositions.13


91What is new in Bitzer's essay is that the audience becomes the defining construct of the enthymeme.

Bitzer's model. Bitzer's essay is, on the surface, an attempt to uncover a correct definition of the enthymeme. He notes from various passages in Aristotle's writings that enthymemes may be "certain and necessary" and therefore concludes that the probability of the premises is not an essential or defining quality of the enthymeme.^ Nor is the enthymeme an elided syllogism. Having dismissed the two usual explanations of the difference between the syllogism of demonstration and the enthymeme, Bitzer puts forth the novel thesis that the enthymeme and syllogism differ in "consequence of a difference in the functions or purposes of the arts. . . . Rhetoric must ask for premises --must begin with premises held by the audience— because persuasion cannot take place unless the audience views a conelusion as required by the premises it subscribes to.The rhetor may argue and draw conclusions from various premises. In many cases, the conclusions might be matters on which the audience has no prior belief, or to which the audience might even be hostile. However, the rhetor draws premises from the audience's beliefs. This leads to Bitzer's definition of the enthymeme as "a syllogism based on probabilities, signs, and examples, whose function is


92rhetorical persuasion. Its successful construction isaccomplished through the joint efforts of speaker and

Ifiaudience, and this is its essential character."Bitzer does not present the joint construction of

arguments as a criterion of the soundness of an argument; it is merely the essential characteristic of the enthymeme. Thus, the enthymeme can emerge only in rhetorical interaction between speaker and audience. Bitzer does not propose that his interpretation provides a foundation for evaluating arguments. His article is better seen as providing a basis for understanding and interpreting arguments.

Farrell’s model. The attempt to escape from the logical fog surrounding the application of logic to rhetoric has led one writer to propose that the only response to the irrelevance of truth-functional standards is to develop a notion of rhetorical validity distinct from logical soundness. The basic notion is that traditional induction and deduction do not account for the adequacy of most rhetorical arguments. However, instead of seeking a broader or different conception of logic, Farrell believes that an argument's soundness must be evaluated in light of the argument's effectiveness in moving an audience. Thus, he looks for an extra-logical solution to the problem of


93justifying multiple criteria of evidence. Although a logical view would examine arguments to determine if they make inferences that are adequately justified by the appropriate canons of sound inference, based on evidence of such quality as to meet the appropriate standards, Farrell's view looks to the impact of the argument as delivered to an audience.

Farrell complains that writers "from George Campbell to Stephen Toulmin have . . . considered validity to be a purely formalistic, or contentless" concept, best represented by law-like, even analytic relationships. Farrell points out that the arguments of rhetoric purport to be substantive and that their loci are "in probable relationships of speaker, audience, and position relevant to par-

17ticular situations of choice and avoidance." His theory posits three constituents of rhetorical argument:

- i. The complicity of the audience in argumentative development;

ii. The probable relationship between rhetorical argument and judgment.

iii. The normative force of knowledge presumed and created by rhetorical argument.18

These constituents represent Farrell's analysis of the precise nature of the weakness of formal logic. Farrell emphasizes that, since rhetoric seeks to accomplish its aims through "the mediation of an audience," the mediation of the audience is necessary in "furthering the steps of


94reasoning involved in rhetorical argument." The audience is not merely active in response to rhetorical acts: the

IQaudience helps to constitute the rhetorical act. Thus, to Farrell, the rhetor acts in cooperation with an audience; the reasoning embodied in discourse involves the beliefs and values of the audience. Bitzer quite properly recognizes the significance of communicative interaction in providing a framework for interpreting arguments. But Farrell proposes that communicative interaction provides a basis for evaluating the soundness of arguments.

Farrell states this position quite clearly when he writes:

What is suggested by my alternative conception is that the 'rationality' or 'soundness' of rhetorical argument is not purely internal to -- - the terms of the argument itself. On the contrary, not even the potential soundness of a rhetorical argument may be surmised until an audience has either acted, or has been rendered capable of acting.20

Farrell makes the logical concept of probability dependenton the rhetorical concept of audience action:

Rhetorical argument has traditionally centered on matters regarded as probable. There are at least two reasons for this apparently arcane classification. First, the matters usually defined as rhetorical are practical questions: matters of choice and avoidance, in which there are "live" options: one must choose, one wayor another. A second, and corollary reason for the "probability" found in rhetorical arguments is that the audience itself, as a


95conglomerate of human agents, retains the responsibility for choice and action. While the audience may be addressed rationally, or cajoled, or manipulated, the speaker is never "entitled" to one particular kind of judgment from an audience.21In this last passage, Farrell hints that a rhetorical

argument is probable because it is only probable that an audience will act on it.

In Farrell's favor, we certainly do, in practice, use acceptance as a sign of soundness. For example, I believe certain medical theories to be sound because my doctor accepts them. However, an argument is not sound or unsound on account of its effectiveness. My doctor's acceptance of a medical theory does not make it sound. Rather, he judges the theory to be sound because it meets the usual, well-worked-out criteria laid down by medical science. An unsound argument can be effective, and a sound argument ineffective. For example, a sound argument might fail to move an audience because it is poorly delivered or poorly worded. Effectiveness is at best a clue to an argument's soundness, but is not a criterion of soundness.

To put it another way, the analysis of a communication situation does not substitute for logical criteria. This may be made clear if we imagine the situation of an audience member trying to apply logical standards to a speech. It would obviously be circular for such a person to say


96(1) "I am acting, therefore the argument is sound, therefore I should accept it, therefore I should act," and it would be spineless to say (2) "Other audience members are acting, therefore the argument is sound, therefore I should act." The latter option merely applies the judgment of others who are presumably making the circular judgment (1). Rather, an audience member wants to say something like,"The argument is sound because it is of sufficient quality; therefore I should act," with quality being determined by reference to non-circular criteria.

A second weakness of this approach is that it requires a distinction between rhetorical and non-rhetorical argument. Rhetorical arguments, which deal with "practical" questions, are to be evaluated in part by their effectiveness; nonrhetorical arguments, one assumes, are not. Farrell's criteria are obviously irrelevant to arguments that do not seek action. One would prefer a theory free of this confusing dualism. For these reasons, Wenzel is right to suggest that Farrell fails to discriminate clearly and consistently between a logical perspective's concern for soundness and a rhetorical perspective's concern for effectiveness.22

Approaches such as Farrell's are motivated by a realization of the inadequacy of traditional deductive logic for


97many of the tasks for which it is called. Nonetheless, Farrell wishes to retain an understanding of how arguments can be critically evaluated: the notion of rhetoricalvalidity is introduced to fill the vacuum left by formal validity.

The important question is, do logical standards apply to rhetoric? If an improved, broader logic is available, it would be unnecessary to develop a notion of rhetorical validity seemingly independent of the standards applied to non-rhetorical discourse. One of the greatest harms caused by the inadequacy of formal logic is that theorists such as Farrell, who see no prospect that logic could be suitable for the study of arguments, seek to find non-logical standards to serve the same function. No objection can be raised against the use of non-logical standards for non- logical purposes, but only a logic can reasonably be expected to serve the purposes of logic.

The Enthymeme as an Elided SyllogismThe enthymeme is sometimes presented simply as a

syllogism with a missing premise. On this view, a premise may be omitted from the syllogism if it is well known to the audience and if the audience can be counted on to fill it in. For example, a speaker might argue that "The United States will go bankrupt, for it is headed toward a debt


98of more than 500 billion dollars." By filling in themissing premise a syllogism can be constructed:

Whatever nation has a national debt of over 500 billion dollars will go bankrupt;

The United States is such a nation;Therefore the United States will be bank

rupt. 23Viewed in this way, the enthymeme is no alternative to the syllogism, since it must be reconstructed into the form of a syllogism for evaluation.

The enthymeme potentially accounts for and evaluates the substantial and value-laden arguments of rhetoric. It is not limited to analytic arguments. The enthymeme is founded on the recognition that arguments are often substantial, uncertain, and problematic. But, because of the ambiguity of the standard of probability, the enthymeme does not provide a model of sufficient clarity.

Many of those who advocate the enthymeme would no doubt desire a rational and logical approach to argumentation and would not have rhetoric become entirely subjective or entirely lose its grounding in logic. To provide an adequate theory requires a systematic account of the theory of argumentation from a logical perspective.


99Induction and Probability

"As God has set some things in broad daylight," remarked Locke, "so, in the greatest part of our concernments, he has afforded us only the twilight, as I may so say, of probability."^ Philosophers have for centuries presented induction as the essential method by which to arrive at conclusions about the probability of a proposition. Induction is sufficiently clear and useful to provide general criteria for the analysis of arguments, but induction is not suitable to the analysis of all inference.

Model of InductionInduction is a process of drawing a conclusion from a

group of particular data. Kneebone describes the usualview that the process of inductive inference proceeds inthree steps. First, one considers a body of data. Then,one tries "to invent a more general hypothesis" to explainthe data. This hypothesis is then further tested against

25previously acquired knowledge. The evidence for an inductive generalization consists of a group of singular propositions which provide support for the generalization. The conjunction of the singular statements provides a logical reason for the generalization. Pollock argues that: "Theconjunction of singular statements clearly does not entail


the generalization, and yet it constitutes a logical reason for believing it."^® There should in turn be evidence for each of the particular propositions.

The particular propositions are often described as based on observation, but this need not be so. ̂ No a priori rule limits the support of inductive evidence. The particular propositions may also be supported by the testimony of various witnesses, or an inductive inference can support a particular proposition. For example, the particular proposition that horses have four limbs can support the inductive conclusion that mammals have four limbs. But the proposition that horses have four limbs can itself be backed inductively. Futhermore, a deductive inference can justify a particular proposition, which in turn can justify an inductive generalization. Induction produces new information; the conclusion is not contained in the premises in an analytic sense. Induction is not truth-functional and can account for a wide range of substantive arguments.

Induction is not formal inference. The pattern of inductive inference is not a form that an induction should meet in order to be sound; it is merely a definition of what inductions are like. There is no particular form that differentiates good inductive inferences from bad, and no formal rules to follow. There are various criteria of soundness, but induction is not in any sense formal inference.


101Induction is not truth-functional. More important,

induction is not truth-functional. It is, for example, a simple matter to produce an induction with true premises and a false conclusion, for example, "It hasn't rained all week, so we'll have a dry picnic today." Inductive premises justify the conclusion, but do not prove it to be true. A sound induction is reliable and makes belief in the conclusion more reasonable than if it were not supported by the inductive premises.^®

This issue is related to the so-called problem of induction, which seeks a deductive justification for the process of induction. The problem of induction may be characterized as a reductive problem. Induction is our source of knowledge for contingent generalizations. However, we have no source other than induction for most such generalizations? thus we cannot justify induction by inductively proving some general principle of induction. Furthermore, there is no non-inductive general principle that justifies the use of induction; that is, there is no a priori principle of induction. A reductive analysis, claiming that inductive generalizations only mean that evidence of such and such sort is available, robs inductions of their force.A generalization generated by an independent induction or by reduction would permit a truth-functional justification


102of induction. However, no such generalization can be established.

The standards of truth-functional logic are, however, simply not relevant to induction; it is not in need of truth-functional justification any more then the syllogism is in need of inductive justification.^ An induction, in and of itself, provides a good reason for a generalization. There is no more fundamental explanation of the soundness of contingent generalizations based on inductions. This is simply because, for contingent generalizations, induction is the most fundamental grounding.

Probability, in InductionInduction does not claim to produce universal truth or

certain conclusions; inductive generalizations may be mistaken. Almost all generalizations result from imperfect inductions.30 Induction can, if successful, establish its conclusions to be probably true. For example, one of induction's uses is to establish generalizations of cause and effect. Since the eighteenth-century empiricists, most writers agree that cause and effect relations can be identified only by imperfect inductions and are claimed to be probable, not certain.31

Events and circumstances can be more or less probable and the evidence can be of greater or lesser quality.


103There is no reason to require a priori a particular degree of precision of arguments; presumably, one will require a greater degree of precision for more important, or perhaps more delicate decisions. For example, Toulmin argues that, with regard to probability, "It causes only trouble if one thinks of the scientific applications of the term as the sole satisfactory ones."32 Why might this be so? First, in matters of controversy, the computation of precise probability figures is often out of the question. The evidence for many useful propositions is so ambiguous that a quantified probability judgment cannot be made precisely.

However, recognizing that probability is not inherently mathematical, the quality of inductive support is variable. An inductive generalization based on ten instances, for example, has greater justification, all other

33things being equal, than one based on two instances. Induction can give varying degrees of support to a general proposition.

Furthermore, inductive conclusions are defeasible. We can always imagine circumstances, however unlikely, under which it would no longer be justified to accept a conclusion drawn from an induction.

Probability in induction is comparative. If the evidence is any good at all, we might assign to an inductive


104generalization some such general evaluation as "very likely," "somewhat likely," "possible," or "remotely possible." Likewise, we often have grounds to call one event more likely than another. Such expressions indicate the degree of confidence that the evidence justifies in the statement, without requiring precision that may be unachievable in a given case. Locke agrees that probability can vary, "there being degrees herein, from the very neighborhood of certainty and demonstration, quite down to improbability and unlikeliness, even to the confines of impossibility."34 This argument does not rule out cases in which probability statements can be made with great precision; sometimes the evidence does permit mathematical ascriptions such as "The probability is 60%."

The application of inductive standards requires judgment and interpretation. But we can turn to the usual, well-known criteria of sound induction for help: fairsampling, careful observation, statistical analysis, and so forth. We should evaluate whether the particular information serving as evidence is typical and properly established. An induction must also evaluate any counterinstances tending to discredit the induction. A probability statement carries as part of its meaning, an implication about the quality of evidence available.


105Toulmin on Probability

Toulmin' s view of probability is apparently inconsistent with the view of probability expressed here, but these views can be reconciled if a few confusions are rectified. Most important, Toulmin confuses an account of the speech act performed in simple, first-person, present tense probability statements with a general account of the meaning of probability. It would be best to say that Toulmin adequately accounts for the force of such probability statements, but not for their propositional content.

Toulmin introduces a distinction between force and criteria of modal terms. The force of a modal term means "the practical implications of its use: . . . This forcecan be contrasted with the criteria, standards, grounds and reasons, by reference to which we decide in any context that the use of a particular modal term is appropriate."^ With regard to probability, the force can be explained by the speech act performed in assertions such as "probably such and such." Toulmin says, "We see how the word 'probably' comes to be used as a means of giving guardedundertakings and making qualified declarations of one's

36intentions." The criteria for the application of the term "probability" include statistical evidence or whatever other evidence backs up the probability statement. ̂ This account


106of probability explains adequately the usage of "probably" in contexts such as "The weather will probably be cloudy tomorrow." The forecaster is guardedly lending his authority to a weather prediction.

A critic might object that probability statements are about something, and that therefore an account of the force of such statements is not the same as an account of their meaning. To this, Toulmin replies that "The abstract noun 'probability'— despite what we learnt at our kindergartens about nouns being words that stand for things— not merely has no tangible [emphasis supplied] counterpart, referent, designatum or what you will, not merely does not name a thing of whatever kind, but is a word of such a type that it is nonsense even to talk about it as denoting, standing for, or naming anything. There are therefore insuperable objections to any candidate for the disputed title." He goes on to state that: "Conversely, there isno special thing which all probability-statements must be about, simply in virtue of the fact that they are . probability-statements." 3 8

The theory of meaning outlined previously suggests that meaning involves four constituents: the reference ofthe expression, its sense, its force, and the speaker's intention. The characteristic speech act performed in


107first-person, present-tense probability statements is the making of a guarded assertion; Toulmin1s theory accounts for the force of probability assertions and argues against the notion that "probability" has a referent. So far this is unobjectionable. But Toulmin gives no reason that the sense or propositional content of probability statements cannot be analyzed. In probability statements, the speaker intends to convey a recognition that justification- conditions have been met. Probability statements have justification-conditions depending in part on the use, or uses, of the word "probability," and understanding these conditions is part of understanding the meaning of sentences using that word.

Understanding the meaning of "probability" to involve both a characteristic speech act and a propositional content clarifies what otherwise would be some paradoxical situations that the critic might face. For example, suppose that I say "I give my authority (guardedly) to the assertion that there will be men on Mars by the end of the decade." On Toulmin*s account, this is the same as for me to say that "It is probable that men will be on Mars by the end of the decade." But this equivalence is implausible. Both of these sentences do perform the same speech act and have the same force. However, it seems intuitive,


108on any conception of logic, that from "It is probable that men will be on Mars by the end of the decade," we can infer that most likely men will be on Mars by the end of the decade. However, this does not follow from "I give my authority to the assertion that men will be on Mars by the end of the decade," regardless of whether my authority is guarded or not. In order to make the inference from the latter sentence, one would need to add some rider to the effect that my authority is good for something. Obviously I should not lend my authority to some matter on which I have no knowledge, but that is not the point.^ The point is that a logical implication is found in the statement that something is probable that is not found in a statement that merely gives my authority. A speech act analysis of force cannot convey an understanding of sense.

Toulmin's account should properly be called a useful explanation of the force typical of probability statements used to perform certain speech acts. Understanding the sense of probability statements based on induction requires an understanding of how induction justifies such statements. This view does not imply the erroneous theories that Toulmin criticizes.

Viewing matters in this way clarifies the relationship between probability and evidential support. Toulmin claims


109that: "Probable . . . is a term which keeps an invariantforce through a wide variety of applications. It is closely connected with the idea of evidential support, but is distinct from that idea. . . . If we go to the length of identifying support with probability, then and only then will the term become ambiguous."^® This position should better be that a probability statement implies something about evidence as part of its meaning. This does not imply the naive position that probability is identical to evidence.

Limitations of InductionUnlike traditional deduction, induction is not limited

to formal, truth-functional inference. Unlike the enthymeme, induction is founded on a fairly clear view of probability. Although induction is valuable, clear, and usable for analysis of arguments, it is not by itself sufficient for the needs of argumentation. It should be apparent that induction and deduction complement each other; neither can replace the other.

Inductive analysis cannot account for arguments from general principles, arguments taking the form of hypothetical implication or disjunction, and so on. Induction cannot justify the adequacy of many sources of knowledge. In general, induction is not a substitute for deduction and was


110never intended to be one. Although induction is unobjectionable in principle, it can never be the whole of logic.

ConclusionThis chapter completes the review of the traditional

models of logic. Modem theories of the enthymeme are vague and many are poorly developed from a logical perspective. Few purport to be primarily logical in the first place. The enthymeme allows for arguments of probability, but many theorists fail to explicate what they mean by "probability" or how the enthymeme can provide a logic of probability. However, much discourse seems to imply conceptual inferences of a non-inductive order: if the modelspresented so far are not adequate, some other model of deductive inference is indispensable to the theory of argumentation. A line of development was suggested for improving this essentially sound approach. Farrell's view, which proposes an interesting solution to the problem of multiple criteria, unfortunately does not address the need for logical criteria of soundness.

Induction provides a reasonably clear basis for inference under certain conditions. They are limited, however, to those instances in which a body of particular statements can be adduced in support of a generalization. Induction is too limited to account for all arguments, but it


Illaccounts for many important kinds of inference and will find a place in a more comprehensive theory.


1 1 2

Notes Chapter III

1. There have been a few advocates of formal logic in this century. Among those who advocate the use of some variety of traditional deductive logic for the evaluation of rhetorical argument is A. Craig Baird, Argumentation, Discussion, and Debate (New York: McGraw-Hill BookCompany, Inc., 1950), p. 156: "You should frequentlyreduce the reasoning to syllogistic form to test its validity and to clarify the thinking." Others who hold traditional deductive logic to be significantly useful for one purpose or another are Hugh G. Petrie, "Does Logic Have Any Relevance to Argumentation?" Journal of the American Forensic Association 6 (1969):55-60;Glen E. Mills and Hugh G. Petrie, "The Role of Logic in Rhetoric," Quarterly Journal of Speech 54 (1968):260-67; Gladys Murphy Graham, "Logic and Argumentation," Quarterly Journal of Speech Education 10 (1924):350-63.

2. See e.g. James H. McBumey, "The Place of theEnthymeme in Rhetorical Theory," Speech Monographs 3(1936):49-74? Lloyd P. Bitzer, "Aristotle' s Enthymeme Revisited," Quarterly Journal of Speech 45 (1959):399-408.

3. Aristotle, Rhetoric, trans. W. Rhys Roberts (New York: Random House, The Modern Library, 1954),1354al4.

4. Ibid., Bk 2, ch. 24; see esp. 1400b34.5. This is McBumey's view, see "The Place of the

Enthymeme."6. Charles S. Mudd, "The Enthymeme and Logical

Validity," Quarterly Journal of Speech 45 (1959):409-14.7. Aristotle, Analytica Priora, trans. A.J.

Jenkinson, in W.D. Ross, ed., The Works of Aristotle 1 (Oxford: At the Clarendon Press, 1928), 76a3; cf.Aristotle, Rhetoric, 1357a34.

8. McBurney, "Place of the Enthymeme," p. 57.


1139. Ibid., p. 52.10. Donald C. Bryant, "Rhetorics Its Functions and

Its Scope," Quarterly Journal of Speech 39 (1953):413; see also p. 406.

11. Douglas Ehninger, "Validity as Moral Obligation," Southern Speech Journal 33 (1968):216.

12. William Brock Brentlinger, "The Aristotelian Conception of Truth in Rhetorical Discourse," (Ph.D.Diss., University of Illinois, 1959), p. 78.

13. McBumey, "Place of the Enthymeme," p. 63.14. Bitzer, "Aristotle's Enthymeme," pp. 401, 403.15. Ibid., p. 405.16. Ibid., p. 408.17. Thomas B. Farrell, "Validity and Rationality:

The Rhetorical Constituents of Argumentative Form," Journal of the American Forensic Association 13 (1977):142.

18. Ibid.19. Ibid., P. 144.20. Ibid.. P. 143.21. Ibid., P. 145.22. Joseph W. Wenzel

Proceedings of the Summer Conference on Argumentation, ed.Jack Rhodes and Sara Newell (Falls Church, Va.: SpeechCommunication Association, 1980), p. 128.

23. Baird, Argumentation, pp. 155-56.24. John Locke, An Essay Concerning Human Understanding,

ed. Alexander Campbell Fraser (New York: Dover Publications,1959), 2: Bk. 4, ch. 14, par. 2.

25. G. T. Kneebone, "Induction and Probability," Proceedings of the Aristotelian Society, n.s. 50 (1950):34.


26. John L. Pollock, Knowledge and Justification (Princeton, N.J.: Princeton University Press, 1974), p. 37.

27. Contrast the logical empiricist position, see Daniel J. O'Keefe, "Logical Empiricism and the Study of Human Communication," Speech Monographs 42 (1975):169-83.

28. Simon Blackburn, Reason and Prediction (Cambridge: At the University Press, 1973), p. 31; see ch. 2, 3, and 5 for his view of induction.

29. On the problem of induction see Stephen Edelston Toulmin, The Uses of Argument (Cambridge: At the University Press, 19SS), pp. 235-240; bollock, Knowledge and Justification, ch. 8; S. F. Barker, Induction and Hypothesis: A Study of the Logic of Confirmation (Ithaca, N. Y.:Cornell University Press, 1957), p. 11; Roy Harrod, Foundations of Inductive Logic (London: Macmillan, 1974),ch. 2.

30. William Trufant Foster, Argumentation and Debating (Boston: Houghton Mifflin Company, 1908), p. 96.

31. Ibid., p. 184;. Graham, "Logic and Argumentation," p. 355; W. H. Walsh, Metaphysics (New York: Hartcourt,Brace & World, Inc., A Harbinger Book, 1963), ch. 7.

32. Toulmin, Uses of Argument, p. 89; see also Frederick L. Will, Induction and Justification: AnInvestigation of Cartesian Procedure in the Philosophy of Knowledge (Ithaca, N.Y.: Cornell University Press. 1974).pp. 64-67, 342-57.

33. This is a widely held opinion, see e.g. Keith Lehrer, "Justification, Explanation, and Induction," in Marshall Swain, ed., Induction, Acceptance, and Rational Belfef (Dordrecht: D. Reidel Publishing Company, 1970),p. 120.

34. Locke, Human Understanding, Bk. 4, ch. 14-15.35. Toulmin, Uses of Argument, p. 30.36. Ibid.. p. 50.37. Ibid., pp. 51-52.


11538. Ibid., pp. 51-52.39. Ibid., pp. 51-2, 90. The general line of my

argument is that of John R. Searle, Speech Acts: An Essayin the Philosophy of Language (Cambridge: At the UniversityPress, 1969), pp. 136-41.

40. Toulmin, Uses of Argument, p. 83.


CHAPTER IV

PROCEDURES AND FIELDS:THE DIALECTICAL VIEW

Arguments based on analogies, authoritative testimony, memory, value judgments, and so on, can all be sound logical arguments. But they are not sound by the standards of deductive logic: nor, according to the conventional viewsof logicians, are they logical arguments at all. Yet, the theory of argumentation seeks to evaluate the soundness of such arguments. A major development in the literature is the view that arguments can be examined according to their acceptability in a communication context of one sort or another when that context employs sound procedures. These contexts might be dialogues or rational enterprises. This approach does not primarily undertake to examine the argument's form, content, or evidential support; rather, it studies the process by which an argument does or does not gain the acceptance of a specified community.

Can Dialectics Substitute for Logic?The procedural approach suggests two answers to the

problem of multiple criteria. The first judges an argument sound if it is accepted in a communication setting. Let us call this a first-order dialectical theory of criteria.

116


117The second judges an argument sound because it meets criteria developed in a communication context using sound procedures. This is a second-order dialectical theory of criteria. The use of the terms first-order and second-order dialectical criteria parallels Butchvarov's first- and second-order criteria of evidence.

Acceptance as a Sign of ValidityHabermas argues that the criterion of truth is that an

argument is justified through discourse. He proposes that "Truth means the promise of attaining a rational consensus."* The rational consensus might not always be attainable in practice, but it is an ideal which makes sensible the notion of rational argument. Habermas specifically criticizes the correspondence theory of truth, which holds a statement to be true because of its accuracy in describing reality. This implies, perhaps, that a statement is not true because of some intrinsic quality of the statement, the evidence that is available to back the statement, or the matters that the statement is about. Instead, a statement is true becauseof the possibility of gaining rational consensus for the

2statement.Habermas' interpreters sometimes carry this position

even further. McCarthy, for example, says that "Rational consensus is the criterion of truth. The argument here


118amounts essentially to a criticism of all other proposed criteria, especially that of a correspondence between thought and reality." Although conventional standards of argumentation can be employed, "The claim is simply that, whatever the particular methods of problem resolution employed, a claim to truth is ultimately accepted or rejected in the light of a discursively attained consensus among those competent to judge."-* Aune argues that Habermas employs a notion akin to rhetorical validity. The potential of consensus is an essential ingredient of soundness: "Instead of relying upon good will, traditional notions of the rational in a given field, or the universal audience, Habermas makes argumentative validity dependent upon the rationality of society.

Now, why will such a view not solve the problem of multiple criteria? After all, an argument can be judged sound by a rational consensus even if the argument is not deductive: an argument’s potential acceptability in dialectic seemingly can be a first-order criterion of evidence. The dialectical theorist need not, we might speculate, entertain such concepts as argumentative form, categories of backings and grounds, patterns of inference, or logical types, except insofar as they shed light on the procedures to be followed in dialectical situations.


119Such a view, however, does not solve the problem of

multiple criteria. Habermas himself clearly recognizes that "One has to show what the consensus-producing power of the argument consists ins it cannot be in the mere fact that agreement can be achieved by argument, for this fact in itself needs explanation."5

The interlocuters in a dialogue must employ noncircular criteria when evaluating arguments. Let us suppose that a group of dialecticians, organized under the necessary circumstances, were able to employ faultless procedures. They cannot now employ a mystical standard by claiming, "We are discussing in an ideal speech situation. Therefore, whatever we agree on is true." On the contrary, they must employ logical criteria of soundness. A dialectical view cannot account for such criteria or explain why they are sound criteria; this is the task of logic. The context in which criteria are employed does not tell us what these criteria should be. It is more accurate to say that "The logical perspective . . . comes into full force under dialectical conditions" and that these conditions "are the optimum conditions for logical analysis, criticism, and judgment."6

Why, then, would a dialectical theory of criteria be attractive? To understand this, we must look back to the


120inadequacy of traditional deductive logic. Theorists of argumentation are coming to abandon logical issues because of the belief that a logical theory of criteria is not viable. This belief steins from a recognition of the limitations of traditional deductive logic and the theories of knowledge that have grown up around it. And, if a logical theory cannot account for rhetorical arguments, the best alternative might be to judge rhetorical arguments with reference to criteria such as consensus or intersubjective agreement.

Deductive analysis cannot show, for example, how knowledge about behavior can justify knowledge about other people's minds. Substantial inferences are beyond the reach of traditional deductive logic. Brummett, who advocates one of the most thoroughgoing dialectical theories, criticizes what he calls the mechanistic philosophy which claims that people justify opinions about reality by making deductions from private sensory experience. Brummett's summary of traditional epistemology reveals his diagnosis of the fundamental problem:

Why don't people apprehend objective reality in everyday life? If the observer doesn't immediately approach objective reality, mechanics argued, there must be some gap between the knower and the known, between people and reality.


121What is the goal of knowledge? To remove

those barriers between knower and known, or, in other words, to arrive at the truth. . . .

How to reach the goal of truth? Or, how best to remove those barriers between knower and known, appearance and reality? Mechanistic epistemology was based on two assumptions.First, reason and formal logic were considered most appropriate for the apprehension of an essentially mathematical reality. Second, objectivity was required for the observer.?

Traditional logic cannot show how to justify claims about the physical world on the basis of sense perception: itcannot resolve the gap between appearance and reality. Brummett astutely puts his finger on "reason and formal logic" as principal culprits.

Farrell's theory of consensually-oriented social knowledge is less specific about the weaknesses of the traditional approach, but he does criticize the "contradictions of extreme realism, radical empiricism, and logical positivism." He maintains that the traditional philosophical explanations of how we gain and justify our opinions must be rejected, and that "Contemporary philosophy has now moved away from the detached derivation of criteria for knowledge and toward the more inclusive study of human activity in all its forms— even as this activity informs the process of scientific knowing itself."® (emphasis supplied)


122In order to resolve the weaknesses of the traditional

theories, consensus theorists find it necessary to replace or supplement logical criteria with the criterion of consensus. Brummett states this quite directly: "Truth, forthe individual, is the extent to which the meanings of experience (that is to say, reality) of that individual are shared by significant others. Truth is agreement." Brummett equates reality with meanings, and argues forcefully that "People get meanings from other people through communication." He contends that "The only reality ever to be encountered is what is observed."® His solution to the reductionism of mechanistic philosophy is to seek an opposite reduction:

While mechanistic objectivity sought the eventual collapse of metaphysics and epistemology by coming directly to know all of reality, intersubjectivity seeks the same collapse but at this end of the philosophical tunnel. Intersubjectivity begins with the assumption that the study of what and how people know is nothing more or less than the study of the nature of reality.10Thus, Brummett begins by criticizing traditional

theories of knowledge for attempting to establish a deductive link between perception and an observable reality, and he concludes by reducing reality to shared experience. The gap between appearance and reality is abolished. This does not solve the problem at all: when we make claims


123about the physical world, we are not merely making claims about agreement and consensus. Brummett1s theory purports to tell us something exciting: it purports to tell ushow we know things about the world. But all it really does is to define the problem out of existence. The problem that Brummett confronts results from the recognition that deductive logic cannot explain how we can justify claims about the world on the basis of, for example, our perceptions about it. But the solution to this problem is not to produce a theory of rhetoric that reduces the problem out of existence; the solution is to come up with better notions about logic.

Farrell's solution is much like Brummett's, although perhaps a bit less extreme. He believes that it is necessary to consider "the collaboration of others as a criterion for knowing" and that "No criterion for knowledge can be polemically proclaimed; at the very least, it must require the cooperation of others in some form." His theory of social knowledge enables "each conscious person to place the content, direction, and intensity of personal knowledge within the context of an attributed distribution of public convictions."11

But, to echo Habermas, the grounds of agreement need to be explained, and a new consensus does not explain the first consensus. Sooner or later, criteria need to find


124their grounding in something other than consensus. The logician wishes to study evaluative criteria that can be employed during the course of communicative interaction.

One cannot help but sympathize with the dialectician's frustrations over the inadequacy of traditional logic. It is beneficial for dialectical theorists to recognize the inadequacies of traditional logic, but these inadequacies cannot be resolved by changing the question in the way they suggest.

Dialectics as the Foundation of CriteriaAnother version of the procedural view does not contend

that arguments are to be evaluated according to their acceptability in some context, but rather according to criteria developed in a particular context. An argument is held to be sound if it meets criteria developed according to sound procedures. This is a second-order dialectical theory of criteria.

This is much like the view advocated in the first volume of Toulmin's Human Understanding. Toulmin shares in common with Habermas an emphasis on the procedures employed during the course of rational criticism. Toulmin does not find the locus of rationality in logical criteria, but in procedures employed in the context of a rational enterprise:


125We must begin, therefore, by recognizing that rationality is an attribute, not of logical or conceptual systems as such, but of the human activities or enterprises of which particular sets of concepts are the temporary cross-sections: specifically, of the procedures by which the concepts, judgements, and formal systems currently accepted in those enterprises are criticized and changed.

Toulmin also says that "The intellectual content of any rational activity forms neither a single logical system, nor a temporal sequence of such systems. Rather, it is an intellectual enterprise whose 'rationality' lies in the procedures governing its historical development and evolution."13

A rational enterprise employs criteria of evaluation developed through and expressed in the procedures by which the enterprise conducts its business. An enterprise employs criteria of evaluation to suit its purposes, which vary from one enterprise to another and from one time to another:

Rationality, we shall argue, has its own "courts" in which all clear-headed men with suitable experience are qualified to act as judges or jurors. Within different cultures and epochs, reasoning may operate according to different methods and principles, so that different milieus represent (so to say) the parallel "jurisdictions" of rationality. But they do so out of a shared concern with common "rational enterprises", just as parallel legal jurisdictions do with their common judicial enterprises.14


For immediate purposes, Toulmin's view cam be interpreted as an account of how rationed criteria are developed, retained, amd discarded. Burleson, for example, takes such an interpretation of Toulmin:

The community of practioners in a discipline thus evaluates the merits of an argument according to how well it conforms with the discipline's ideals or goals and aids in resolving the discipline's set of unanswered questions. Ultimately, then, the "goodness" of an argument (i.e. the truth or validity of a claim) is determined by the consensual evaluation of a community of practioners acting in accord with their interpretation of the discipline's basic tasks.15

Such an evaluation must make use of logical standards.Burleson goes on to point out that "relativism is avoided"because there "is always a substantive basis for conceptualchange.

Basing his opinion on an interpretation of Habermas and Toulmin, Burleson advocates a similar view himself.He takes Habermas to advocate a theory of second-order criteria in which the rationality of specific logical criteria is tested against the standard of consensus. In favor of Habermas' view, so interpreted, Burleson writes:

In sum, working logics serve in the assessment of the rational legitimacy of claims within a specific field, while the logic of discourse ensures that such testing processes are themselves rationally legitimate. Thus, both a field-dependent working logic and a universal logic of discourse play essential roles in the process of testing and assessing claims to


validity, and rationality is demonstrated by adherence to the rules of both logics.

Burleson holds that the soundness of arguments is evaluated by reference to the quality of evidence and inference, as determined "within the confines of the particular standards of judgment developed in an historically evolving discipline."18 Thus, "The truth of a claim, in the final analysis, is inseparable from the social means through which

IQthe claim is assessed, tested, and justified."4*7Views such as Toulmin's and Burleson's might be thought

an apparent solution to the problem of multiple criteria.The soundness of criteria can be accounted for by reference to their grounding in rational enterprises and consensual procedures. All that this does, however, is to delay the logical problem; for, by what criteria are the criteria to be judged?

A rational enterprise, if it is to be rational, must have reasons for employing the criteria that it does.Simply saying that the criteria serve a field's purposes does not do the job. Having specified the purpose that is to be served, one still wants to know how to tell when a criterion serves the purpose. The criteria still need to be explained, accounted for, and justified on logical grounds. And, although these grounds might vary from field to field, or from time to time, it still might be possible


128to develop explanatory and critical categories of one sort or another, and to state a principle or principles to account for these categories: in short, to develop a logical theory of criteria. Such an account, or set of accounts, would have to be sensitive to the enormous richness and variety of criteria actually employed in various fields, as well as to the possibility that new concepts of evaluation might be developed to meet changing circumstances and purposes.

The need for a logical grounding for criteria can be made clear if we consider the position of arguers functioning within, for example, a rational enterprise. Suppose that they are faced with the problem of justifying evaluative criteria or procedures; Toulmin repeatedly discusses circumstances in which evaluative criteria are indeed challenged.20 A defender of these criteria cannot simply say "These are the criteria that the field uses;" "These criteria work;" or "We've all agreed to use these criteria." If a dialectician cannot give reasons for which the accepted criteria should be considered sound, it is difficult to see how these criteria can be justified. The notion of sound procedures does not serve logical purposes.


129Although Human. Understanding is an unfinished work,

there are numerous hints in the first volume to the effect that logical criteria are employed in rational enterprises and that there are reasons deeper than procedures to account for the soundness of logical standards. Toulmin advocates the aim of showing "how the intellectual authority of our concepts finds an ultimate source within the empirical matrices of understanding itself. Gut what directions he would take in such an investigation is not clear.

Fields of Argument The previous discussion glided over an important con

tribution of the dialectical or procedural view: thefield of argument. The appearance of Toulmin's Uses of Argument stimulated discussion of fields of argument.Since Toulmin introduced the concept of a field to suit logical purposes, it stands to reason, despite the arguments raised in the previous chapter, that a solution to the problem of multiple criteria might be found in his view that evaluative criteria are field-dependent. However, as the notion of a field has been developed, its utility for logical purposes is minimal: scholars have chosen to takea sociological view of fields inspired by Toulmin's


130Human Understanding/ rather than to develop the concept of a field from a logical point of view. To a great extent, the sociological view of fields, with its dialectical overtones, has come to displace the study of argumentation from a logical perspective.22

For example, fields are often defined in terms of academic disciplines.22 Willard characterizes a field as a school of thought;24 Rowland characterizes fields in terms of the shared purposes of the arguers in a rational enterprise.25

Definitions of this sort share one thing in common: they all define in terms of non-logical concepts. The Uses of Argument, on the other hand, defines a field in terms of the logical types of the propositions used in the field's arguments. Non-logical definitions fit closely with what our intuitions of a field tell us about what a field is, but they are not especially helpful for logical analysis and criticism.

Let us make a distinction between critical theories of logic and descriptive theories of logic. A critical theory provides an account of what standards and criteria are sound, what kinds of grounds can validly be offered to support claims, and how inferences can be drawn from those grounds to give claims sound support. A descriptive theory


13126explicates what standards and criteria are in fact used.

The concept of a field as a sociological entity of one sort or another yields useful descriptive explanation. For example, Rowland shows how the shared purpose of arguers in a field influences the types of arguments used and the rigidity of standards employed, but he does not undertake

27to show how the quality of standards should be determined.As fields are presently understood, investigation of fields does not tell us about the soundness of arguments; nor is it among the intentions of theorists such as Rowland to do so.

Toulmin's notion of a field of argument as a rational enterprise is minimally useful to a critical theory. His original conception of fields turns out to provide a more useful framework for a logical theory of argument? indeed, even those questions that it addresses but leaves unresolved are of greater interest to the logical perspective. Theorists of argumentation, influenced by the interesting but unfinished theory in Human Understanding, have paid less attention to the logical conceptions of The Uses of Argument.

Toulmin*s Original View of FieldsAs originated, the concept of a field of argument was

both logical and sociological. In The Uses of Argument,


132Toulmin*s examples of fields include such sociological entities as aesthetics, team-selection, physics, and mathematics. Toulmin defined a field in terms of the logical type of the propositions found in a field’s arguments:"Two arguments will be said to belong to the same field when the data and conclusions in each of the two arguments are, respectively, of the same logical type: they willbe said to come from different fields when the backing or the conclusions in each of the two arguments are not of the same logical type.^® Toulmin is not using "data" or "backing" in the more specific way outlined in his wellknown layout of arguments; in the present passage, "data" and "backing" seem to mean much the same. Toulmin does not define "logical type" but does provide numerous examples: some logical types are "reports of present andpast events, predictions about the future, verdicts of criminal guilt, aesthetic commendations, geometrical axioms and so on."JW

One reason that Toulmin holds formal logic incapable of handling substantial arguments is that many substantial arguments involve type-jumps in which the conclusion is of a different logical type from the data or backing.^1 We can get an idea of what a logical type is by examining some examples of-type jumps:


133We make assertions about the future, and back them by reference to data about the present and past; we make assertions about the remote past, and back them by data about the present and recent past; we make general assertions about nature, and back them by the results of particular observations and experiments; we claim to know what other people are thinking and feel* ing, and justify these claims by citing the things that they have written, said and done; and we put forward confident ethical claims, and back them by statements about our situation, about foreseeable consequences, and about the feelings and scruples of the other people concerned.32

This list makes it clear that "logical type" is a general concept that encompasses, at the least, distinctions of time and circumstance; of purpose (ethical and otherwise); generalizations and inductive evidence; thoughts and behavior. These distinctions are, for the most part, related to the content pr subject-matter of the propositions. Logical types are not particularly characterized by the forms of the propositions.33 Further, all of the examples of type-jumps are of cases in which a deductive inference could be made only by the application of reductive techniques. For example, statements about historical matters cannot be entailed by statements about the present: it is always logically possible for statements about present circumstances to be true, and for the statements about the past inferred from statements about the present to be false. A historian might make statements about ancient


134battle maneuvers on the basis of documents that exist at the present time. But these documents could be in error for any number of reasons, so that historical knowledge based on them could be false. Thus, a demonstrative inference is impossible. The problem can be resolved within the confines of traditional logic only by reduction: by anassumption that statements about the past only mean that certain evidence is available now. If this reductionist solution is rejected, scepticism concerning historical knowledge seems to be the only alternative.

Finding this dilemma between scepticism and reductionism unacceptable, Toulmin developed his theory of field-dependent substantial arguments, defining fields in terms of logical types. Logical types are the key factors used to explain the limitations of traditional deductive logic. The problem of multiple criteria of evidence can be stated as, "What understanding of evidence and reasons enables us to make sound substantial arguments— arguments that use type-jumps?" His answer grounds good reasons in the standards and criteria of various fields of argument.

This theory came afoul of a difficult problem.Toulmin falsely believed that the common fields of argument — philosophy, mathematics, team-selection, railroad- passenger relations, and the like, neatly correspond with


135logical types. For each field, it was supposed, there is a characteristic type of argument.

Willard, however, argues that a field is a sociological entity that operates according to patterns developed by the persons who work in that field. He claims that:

It is their aims, not their claims (or the entities to which their sentences refer) that makes up their special characteristics. . . .The type notion, then, cannot help us point to this or that characteristic to define a field and to mark it off from other fields.We require sociological notions to proceed clearly.35

Fields regularly borrow principles and standards from other fields; many standards are not the unique property of any field, and in general, types of arguments are more or less independent of the fields in which they are used. Thus,"To use types to define fields is to muddle two distinct purposes, viz., a philosophical/analytic aim and the social psychological empirical project I take to be central to field theory." Willard goes on to state that "The importance of this difference is that field theory cannot be expected to perform the functions of logic."36 One field might use various types of arguments, whereas different fields might use the same type of argument.

For example, various fields might use experimental evidence, statistical analysis, and so forth, and there


136might be no discemable difference in the way these standards are applied. A psychologist might justly criticize a sociologist for using incorrect statistical methods. Several different fields might make inferences from statements about the past to statements about the future.

A single field might use arguments of different logical types. For example, in history we might have argument by appeal to ancient documents as well as deductions or inferences from archaelogical findings. A historian might establish that an event took place by citing a letter describing the event, or by deducing from various known events that some other event must also have taken place.In psychology, some arguments are based on inference from theoretical principles, and other arguments on experimental findings. It seems highly implausible that a field is characterized by the use of a certain type of argument.

Arguments, and the propositions that make them up, vary greatly from one field to another, but these variations are not strictly correlated with fields; criteria are field-variant but not field-dependent. And if logical types do not bear some particular relationship to fields, if the concept of a field is not of any unique logical significance, Toulmin's claim that criteria are grounded in fields becomes a mere endorsement of the criteria actually


137employed in practice? or, more likely a mere descriptive account of logical criteria as they are actually employed.

The untenable position in The Uses of Argument can be resolved in two different ways. The first would be to emphasize the sociological aspects of a field. A field of argument is no longer to be seen as a logical construction. This is the direction followed in Human Understanding.Most theorists of rhetorical argument have followed the same course. Such a view, we have seen, does not address matters of logical criticism. The other alternative is to emphasize the logical characteristics of arguments, but not to assume that such characteristics are strictly field- dependent. Let us examine what sort of theory might be built on this second choice.

Toulmin's View of Logical Types: Future DirectionsToulmin*s well known model lays out the relationships

among data, warrant, backing, claim, modal qualifier, and possible exceptions (rebuttals)• Figure 1 outlines Toulmin's layout.

The warrant is an inference-license that permits the move from data to claim. The backing is often unstated— it is typically stated only if the warrant is challenged—


138

D(ata)

Figure 1 TOULMIN DIAGRAM

So, Q(ualifier), C(laim)

Since W(arrant)

UnlessR(ebuttal)

On Account of B(acking)


139but the backing for the warrant of a sound argument is always statable in candid discourse. In one of Toulmin's examples, the backing for "A man b o m in Bermuda will generally be a British subject" is "On account of the following statutes and legal provisions. . . . ” The backing for "A Swede can be taken to be almost certainly not a Roman Catholic" is "The proportion of Roman Catholic Swedes is less than 2%."

In both of these examples, the backing must come from somewhere; it must be based on something. In the first example, according to Toulmin, the backing comes from an examination of the statute books. The backing in the second example presumably comes from a study of Sweden's demographic records. Let us simply say that there is often grounding for the backing. Likewise, as many writers have pointed out, there must often be grounding for the data. Figure 2 modifies Toulmin's layout to reflect such grounding.

Furthermore, many grounds have grounds. Demographic data come from surveys; the information in statute books comes from first-hand reports of legislative proceedings. So, we can add grounds for grounds to the bottom and left- hand ends of Toulmin's diagram.


G(rounding)

140

Figure 2 EXPANDED TOULMIN DIAGRAM

D(ata) So, Q(ualifier), C(laim)

Since W( arrant)

On Account of B(acking)

G(rounding)

UnlessR(ebuttal)


Now, in a dialogue, matters rarely become so complicated. The interlocuters often agree on the soundness of a warrant, for example, without demanding backings, groundings for backings, or groundings for groundings. However, a sufficiently Socratic questioner can demand such unfolding of grounds if necessary.

If an argument, so pressed, is to be justified, the regression of grounds must be stopped somewhere. The argument's defender should have a basis for his opinions.The whole idea of sound argumentation is that one's claims are founded on sound considerations. But how are they founded? And in what way do grounds serve as good reasons: what is the nature of the logical relationship between data and claim? Between grounds and data? Between backing and warrant?

The answers implied in The Uses of Argument are that reasons are grounded in the myriad ways specified in various fields, and that reasons justify claims in various ways in different fields. For example, as Zarefsky reads The Uses of Argument:

Having rejected the universal validity standards of formal logic, and wishing also to reject a vicious relativism, Toulmin arrived at the "field" concept as a middle route. Arguments can be evaluated from a rational standpoint, but the criteria for what that standpoint consists of will vary with the field we are in.37


142As to the nature of data, backing, and their ground

ings, it would be no surprise to find these correlated with logical types. Let us consider a couple of Toulmin's own examples. Let us also, for the moment, follow Toulmin's example in Knowing and Acting by using the term "grounds" indiscriminately for data, backing, and grounds.These examples will help us to arrive at a non-field-dependent account of criteria.

1. We claim to know what other people are thinking or feeling, and justify these claims by citing the things that they have written, said, and done.39

That is, our grounds for knowledge about other people's minds come from knowledge of their behavior.Claims of one logical type, knowledge about other minds, are justified by statements about behavior.

2. We make assertions about the future, and back them by reference to data about the present and past.40

Statements about the future are of a logical type, and are characteristically justified by statements about the present and past. Indeed, since no one has first-hand knowledge of the future, the only grounds that can be offered for statements about the future are statements about the past and present.


143Logical types are characterized in terms of the types

of considerations that can be offered in support of certain types of claims: at least in simple cases, claims of onelogical type are typically supported by data and backing of another particular logical type.

So, what needs to be done to fill out the logical side of Toulmin's theory is, first, to clarify the nature of logical types, particularly those logical types that play especially important roles in justifying claims and to organize the different logical types that can be offered as data, backing, and grounds into useful categories.These should be based, insofar as possible, on the logical characteristics of the types. We must also show how propositions of one type can convey justification to claims of another type without resort to reductionist methods and find a way to end the regress of justifications modelled by the trails leading off the left-hand and bottom of the modified version of Toulmin's diagram.

ConclusionGiven the widely recognized inadequacy of traditional

logic, some scholars propose applying procedural standards to evaluate the quality of arguments. Dialectical theories, such as those of Habermas and Toulmin, would be used to fill in the gaps left by traditional logic and thus solve


144the problem of multiple criteria. Because traditional logic is not adequate, many scholars seem to believe that the logical perspective on argument cannot stand on its own, and they propose non-logical theories to serve the purposes of logic.

Notwithstanding the value of dialectical theories for other purposes, they are not capable of serving a purpose for which they were never well suited. Logical criteria are still required.

On the other hand, Toulmin's original conception of argument has substantial merits as a logical theory. He points the way to a conception of argument that brings logical standards more into accord with good logical practice. Further development of his insights can be of great value to the theory of argument.

Toulmin's model gives us a framework for the analysis of arguments from a logical point of view. It also poses an important problem that can no longer be met, because of the considerations given in this chapter, merely by an appeal to the actual standards of actual fields.Toulmin's original approach has not, however, lost its vitality. Resolving these issues would solve the problem of multiple criteria: it would show how the grounds orsources that are available, that rhetors cam and do use in


145marketplace argumentation, can lend justification to claims. The next chapter clarifies the nature of the grounds that rhetors can offer, and show how these grounds justify claims.


146Notes

Chapter IV

1. Jurgen Habermas, "Wahreitstheorien," in Wirklichkeit und Reflexion (Pfullingen, 1973). All citations are to a draft translation by R. Grabau, "Theories of Truth," mimeographed, n.d. The current reference is to p. 7.

2. Ibid., pp. 1-10.3. Thomas McCarthy, "A Theory of Communicative

Competence," Philosophy of the Social Sciences 3 (1973): 150. For reasons expressed later in the discussion, I have doubts that Habermas' position is actually so extreme. Furthermore, it does not need to be so extreme; no inherent conflict lies between logical and dialectical standards. Even if the more extreme interpretation of Habermas' view is correct, Habermas' theory could easily be reconstructed along simpler lines. See Thomas McCarthy, The Critical Theory of Jurgen Habermas (Cambridge, Mass.: The Mit fcress, 1̂ *78),pp. 291-310, in which McCarthy qualifies his earlier views.

4. James A. Aune, "The Contribution of Habermasto Rhetorical Validity," Journal of the American Forensic Association 16 (1979):104.

5. Habermas, "Theories of Truth," p. 22.6. Joseph W. Wenzel, "Jurgen Habermas and the

Dialectical Perspective on Argumentation," Journal of the American Forensic Association 16 (1979):94. Habermas argues clearly enough for a similar view in the second half of "Theories of Truth."

In addition to the views discussed here, Thomas J.Farrell, "Knowledge, Consensus, and Rhetorical Theory,"Quarterly Journal of Speech 62 (1976):1-14 and BarryBrummett, "Some Implications of 'Process' or 'Inter- subjectivity': Postmodern Rhetoric," Philosophy &


147

Rhetoric 9 (1976):21-51, especially pp. 25-32, both present thoroughgoing theories of first-order criteria from a dialectical or consensual perspective.

7. Brummett, nSome Implications," pp. 22-23.8. Farrell, "Knowledge, Consensus, and Rhetorical

Theory," p. 35.9. Brummett, "Some Implications," pp. 27, 29, 34.10. Ibid., p. 28.11. Farrell, "Knowledge, Consensus, and Rhetorical

Theory," pp. 3, 12.12. Stephen Toulmin, Human Understanding (Princeton,

N.J.: Princeton University Press, 1972) 1:33.13. Ibid., p. 85. Note that standards of evalua

tion are held to be relatively more formal within a conceptual scheme, less formal when conceptual change is considered. See ibid., p. 226.

14. Ibid., p. 95.15. Brant R. Burleson, "On the Foundations of

Rationality: Toulmin, Habermas, and the A Priori ofReason," Journal of the American Forensic Association 16 (1979) :11T.-----------------------------------------

16. Ibid., p. 116. Indeed, Burleson criticizes Toulmin for not giving an even more specific grounding for criteria in a consensual procedure.

17. Ibid., p. 126.18. Ibid., pp. 125-26.19. Ibid., p. 127.20. Toulmin, Human Understanding, 1:222-30.21. Ibid., p. 28. Or, Toulmin may simply have changed

his mind when he advocated a position much less relativistic than that outlined here, in Stephen Toulmin, Knowing &


148Acting: An Invitation to Philosophy (New York: MacmillanPublishing Co., Inc., 1976). The problems I have noted may simply be a lesson in the dangers of interpreting an unfinished work.

22. Charles Arthur Willard has published a series of articles arguing vigorously in favor of such a displacement, see e.g. "A Reformulation of the Concept of Argument: The Constructivist/Interactionist Foundations of a Sociology of Argument," Journal of the American Forensic Association 14 (1978):i22; "Field Theory: A Cartesian Meditation," Dimensions ofArgument: Proceedings of the Second Summer Conferenceon Argumentation, ed. George Zxegelmueller and Jack Rhodes (Annandale, Va.: Speech Communication Association,1981), pp. 21-43; "Some Questions About Toulmin's Viewof Argument Fields," Proceedings of the Summer Conference on Argumentation, ed. Jack Rhodes and Sara Newell [Falls Church, Va.: Speech Communication Association,1980), pp. 348-400.

23. Malcolm 0. Sillars, "Investigating Religious Argument as a Field," Dimensions of Argument, pp. 143- 51 views religion as a field; several of the papers in the same volume view legal argument as a field; e.g. Richard D. Rieke, "Investigating Legal Argument as a Field," pp. 152-58 and Janice Schuetz, "The Genesis of Argumentative Forms and Fields," p. 283.

24. Willard, "Some Questions About Toulmin's View of Argument Fields," p. 378.

25. Robert Rowland, "Argument Fields," Dimensions of Argument, pp. 61 ff.

26. For example, I take Walter R. Fisher, "Toward a Logic of Good Reasons," Quarterly Journal of Speech64 (1978):376-84 to be a descriptive account of the logic of value arguments, whereas George W. Ziegelmueller and Charles A. Dause, Argumentation: Inquiry and Advocacy(Englewood Cliffs, N.J.: Prentice-Hall, Inc., 1975),pp. 27-140 present a critical theory of rhetorical arguments.

27. Robert Rowland, "Argument Fields," Dimensions of Argument, pp. 56-79, esp. pp. 63-67.


149

28. More accurately, Toulmin sidesteps logical problems in Human Understanding, but will perhaps address them in the continuation of that work. He addresses such issues in Knowing & Acting.

29. Stephen Edelston Toulmin, The Uses of Argument, (Cambridge: At the University Press, 1958), p. 14.

30. Ibid., pp. 13-14.31. Ibid.. p. 167.32. Ibid., p. 219.33. The contrary view has gained general acceptance;

see e.g. David Zarefsky, "Persistent Questions in the Theory of Argument Fields" (Paper delivered at the Meeting of the Speech Communication Association, Anaheim, Ca., November 14, 1981), p. 4.

34. See Toulmin, Uses of Argument, pp. 229-32 on "Phenomenalism and Scepticism."

35. Charles Arthur Willard, "Argument Fields and Theories of Logical Types," Journal of the American Forensic Association 17 (1981):137.

36. Ibid., p. 141.37. Zarefsky, "Persistent Question," p. 3. The

above discussion follows Toulmin, Uses of Argument, pp. 104-13. --------- -----

38. Toulmin, Knowing & Acting, ch. 7.39. Toulmin, Uses of Argument, p. 219.40. Ibid.


CHAPTER V A THEORY OF CONCEPTUAL INFERENCE

Logic studies the kinds of evidence that people use, explains why that evidence should count as evidence, and explicates the rules of inference that explain how evidence supports claims. Since logic is a study of conceptual relationships, it should explain how the study of meanings accounts for evidence and inference.

Roderick Chisholm and John Pollock have suggested an interesting approach to these matters. Chisholm's epistemology is directed to explaining the foundations of justified belief. He suggests the study of three modes of inference: rules of induction, rules of deduction, andrules of evidence. Rational belief is based on the rules of evidence, which authorize the logical study and criticism of non-deductive conceptual relationships.'*' This suggests the existence of a logic of rational belief. Chisholm's rules of evidence give a rigorous basis for "inventio, the term applied in Roman rhetoric to the systematic investigative procedures by which rhetoric sought to turn up all the relevant arguments or considerations in any given situation," and which are immediately applicable

150


1512to practical problems. Pollock develops Chisholm's posi

tion clearly and systematically in Knowledge and Justification.̂

Logical reasons are reasons because of the relationships among the meanings of their terms and propositions. Meanings can often be stipulated by stating the conditions under which a proposition is justified. There is a logical connection between the claims of marketplace argumentation and the evidence that justifies them. The result is a way of understanding multiple, field-variant criteria.

The evidence that a rhetor presents will be of one logical type or another. A particular type of evidence draws on some source to establish the claim; so, the study of logic involves the study of sources. For example, the claim that "Prices are rising" might be supported by "The Bureau of Labor Statistics announcement of 31 March states that the Consumer Price Index rose by 5% last quarter."The evidence draws on the source of expert testimony. The claim that this table is brown draws on the source of visual perception. In general, the various propositions offered in evidence for various arguments can be categorized into logical types, each characterized by a

4source. Sources cannot be catalogued comprehensively,


152but a useful system can still be imposed to organize ourthinking about them.

The underlying epistemological principle at work isthat our usual sources of knowledge ought to be trustedunless there is a special reason not to trust them in aparticular case. Chisholm explains the point this way:

One important criterion— one epistemological principle— was formulated by St. Augustine. It is more reasonable, he said, to trust the senses than to distrust them. Even though there have been illusions and hallucinations, the wise thing, when everything seems all right, is to accept the testimony of the senses. . . . If on a particular occasion there is something about that particular occasion which makes you suspect that particular report of the senses, . . . then perhaps you should have some doubts about what you think you see, or hear, or feel, or smell. . . . In short, the senses should be regarded as innocent until there is some positive reason, on some particular occasion, for thinking5 that they are guilty on that particular occasion.

We are justified in holding beliefs based on sense perception unless there is some reason to believe otherwise. We are justified in holding beliefs based on the memory of past events unless there is some reason to believe otherwise. Some sources are highly reliable. In the case of some sources that are often wrong (memory, for example), we may frequently, or even usually, have reason to doubt. Prom the standpoint of logic, an argument is adequately grounded if the claim is derived by some inference from a usual source and if one believes no particular reason to doubt it in a particular case.® But to address the


153problem of multiple criteria we need to know why such arguments are sound and to clarify the nature of sources.

Logic can become the study of the general principles that should be observed in accepting propositions for rational belief and the exceptions to those principles.The principles will not be derived from academic, economic, or practical fields. They are instead the principles of sources of knowledge. Pollock explains:

Our knowledge can be separated into areas according to subject matter. These areas will include knowledge of the physical world, knowledge of the past, knowledge of contingent general truths, knowledge of other minds, a priori knowledge, and possibly knowledge of moral truths. The significance of these areas is that each has associated with it a characteristic source of knowledge.7

This list cannot be considered comprehensive. A theoryof criteria could be founded by categorizing sources ofknowledge and "defeaters," i.e. reasons for doubt.

The conceptual relationships of interest to theeveryday logic of predication are those of the conditionsunder which one is justified in applying a word to anobject, not the conditions under which one is correct in

Oascribing a word to an object. Traditional logic assumes the necessity of a truth-functional connection: if thetruth-conditions of the premises are met, then, the truth- conditions of the conclusion must also be met. But for many arguments, it is possible (not logically contradictory) for the premises to be true and the conclusion


154false. In these cases, traditional logic requires what Perelman and Olbrechts-Tyteca call "a perfectly unjustified and unwarranted limitation of the domain of action

9of our faculty of reasoning and proving." The key problem is to show how non-truth-functional arguments can be sound logical arguments. By considering meanings to be given by justification conditions, this thesis argues that logic can encompass sound but fallible arguments.

InferenceLet us say that a good reason for a claim justifies

that claim. Good reasons might, or might not, be conclusive. A good logical reason is a proposition or set of propositions that justifies a claim merely because of the meanings of the terms and propositions.'*'0

Inference and MeaningIn consequence of the extreme views of philosophers

of the ideal language school, such as Carnap, Russell, and Ayer, logic is widely considered to be a system of tautologies. The tautology is generally dismissed as uninformative and useless to arguments of any consequence. The tautology is typically represented as an argument that is true because of the meanings of the terms and true for all possible circumstances. A typical example is "If Mike is


155married, he is not a bachelor," which is true because of the meanings of the words involved.

Philosophers usually teach that such inferences are devoid of substantive probative value. But many useful and productive arguments are sound (but possibly not true for certain) because of the meanings of the terms. We can start with a simple example, adapted from Aristotle's discussion of probable propositions in the Prior Analytics.11

It might be said that "lovers show affection." If John and Mary are lovers, a reasonable inference could be made that they show affection to each other. This seems to be part or all of what is meant by calling them lovers. If they stopped showing affection, it would be difficult to still call them lovers. One can easily imagine an unhappy Mary saying, "John, you don't show me any affection; I don't think you're still my lover." This could reasonably be characterized as a conceptual argument. The argument is not based on an induction: if one encounteredcases of alleged lovers who never showed affection, one would not view them as counterexamples to the generalization. Instead, one would conclude that they are not really lovers. Thus, the argument that "If John and Mary are lovers, they show affection to each other" is sound merely because of meanings. Yet it could hardly be called a uselessly analytic argument or a pointless tautology?


156indeed, such arguments could repeatedly be useful to a couple discussing the course of their relationship. Nor is the argument truth-functional, because there are cases in which lovers would not show affection. For example, if John is extremely worried about his health, he might be too distracted to show affection to Mary even though he still loves her.

Another example, drawn from public affairs, illustrates the same point. Politicians typically say that a conservative opposes major change in our social system. Opposition to major change is at least part of the meaning of "conservative." If a major change is proposed, one would expect a conservative to oppose it. This is a conceptual argument. It is not inductive because one would not investigate the matter by cataloguing conservative politicians and looking to see if they usually support or oppose major changes. On the contrary, a politician who regularly and routinely supports major changes is not a conservative because of the meaning of the word. It is, of course, possible for a politician to claim to be a conservative even though he regularly supports major changes, but such a person would not be using the word "conservative" justifiably. On the other hand, the argument is not truth-functional, because circumstances sometimes arise in which conservatives support non-conservative programs.


157These are examples of arguments that a traditional

logician would reject out of hand. They are not founded on universal premises. They cannot operate truth- functionally because they are defeasible. Although they are sound because of the meanings of the terms and propositions used in developing the argument/ the conceptual relationships are not truth-functional. The traditional theories of logic and meaning cannot account for arguments of this sort. We can argue that these arguments are faulty, illogical, and illegitimate/ on the grounds that the traditional theory cannot account for their soundness. Or, we can revise the traditional theory. Revising the traditional theory seems the wiser course. All of these arguments give the impression of being sound, or at least sound if one grants the speaker's perspective on the problems at issue; the conclusions do indeed seem to follow reasonably from the premises, and arguments of this type are frequently found in discourse on the types of problems giving rise to rhetorical exigences.

These examples may seem too obvious to belabor at such great length, but, if numerous such arguments occur, the traditional view about tautologies will have to be abandoned. A large and important class of arguments progress from reasons to conclusion only by making use of the meanings of the terms and propositions. These arguments


158are defeasible, significant, potentially useful, and not necessarily inductive. The traditional theory of meaning implies that logical relations must be truth-functional. But, since many good logical reasons do not give truth- functional support to claims, the old theory must be wrong. The sense of a proposition can be given by stating justification-conditions. The best argument for this new theory is that it yields a more adequate logic.

Prima Facie ReasonsMany recent philosophers restrict the class of logi

cal reasons to conclusive, demonstrative reasons. However, Pollock argues that there are three kinds of logicalreasons: conclusive or deductive reasons, inductive

13reasons, and prima facie reasons. A prima facie reasonis a logical reason that is defeasible: "A prima faciereason is a reason that by itself would be a good reasonfor believing something, and would ensure justification,but may cease to be a good reason when taken together with

13some additional beliefs." A defeater is a proposition that provides a reason not to believe a logical inference. It may be that a statement P is a logical reason for believing another statement Q. A third proposition R is a def eater for the inference if the conjunction of P and R is not a logical reason for believing that Q.^ A type one def eater denies the conclusion of an argument. If my


159friend Steve calls and tells me that my wife is at the shopping center (P), this would be a prima facie reason for me to believe that she is at the shopping center (Q) . But I know that my wife is at home with me (R). Since I believe both P and R, I do not have a prima facie reason to believe Q. A type two def eater attacks the legitimacy of the inference. I might believe that the Japanese automobile industry is selling cars in America below cost on the basis of a statement by the President of the United States. But if I were to discover that the President had received a large contribution from the American automobile, industry, I would be less inclined to believe his statement. The second type of defeater would also be relevant if it seems to me that I see red, but know the room to be lit by colored lights.^

A prima facie reason does not produce a deductive inference, since it is entirely possible for a prima facie reason to be mistaken or defeated. However, if a prima facie reason P can be adduced for some statement Q, then there is a presumption for the conditional statement "If P, then Q." If one believes P to be true, then there is a presumption in favor of believing Q. However, in a valid deductive inference, the link between grounds and claim is beyond challenge: it is not reasonable to acceptthe premises of a valid demonstration and reject the


conclusion. It can, however, be reasonable to accept thepremises of a prima facie argument and reject the conclu-

16sion.Logical reasons. As an alternative to traditional

logic, Pollock contends that the conceptual relationships behind many logical reasons are given by the justification-conditions of the terms involved and not by their truth-conditions. Logical reasons are defined this way:

Whenever the justified belief-that-P is a good reason for one to believe that Q, simply by virtue of the meanings of the statements that P and that Q, we will say that the statement-that-P is a logical reason for believing the statement-that-Q.17One might object that this definition restricts logi

cal reasons to those instances in which one can deduce a statement from another statement solely by virtue of the meanings of the concepts in the statements. However, many arguments use contingent or substantial inferences.Pollock replies that a contingent reason can invariably be reduced to a logical reason. Let us suppose that I believe a proposition P which is a good but contingent reason for believing another proposition Q. Pollock argues that, if P is a good reason for believing Q, one can always give a reason to believe Q on the basis of P. This reason is equivalent to the warrant of the Toulmin model. The conjunction of P and the principle "If P,


161then Q" is a logical reason for believing Q. Althoughthis involves some restructuring of the original argument,that I believe Q on the basis of P, it shows that thelogical reason is a more fundamental type of reason than

18the contingent reason. Thus, logical reasons can have probative value in cases of substantial inferences.

Non-logical good reasons. Not all contingent reasons support arguments that are necessarily sound because of logical relationships? in other words, it is not always essential to go through the above procedure in order to produce a sound argument. I can justifiably assert that my desk is brown because it looks brown to me. Its looking brown to me is not a proposition and has neither truth-conditions or justification-conditions, but it justifies my statement that the desk is brown all the same.

I could say "The desk looks brown to me; things that look brown presumably are brown; therefore, the desk is presumably brown." This is a logical argument. But I can justifiably assert that the desk is brown without formulating such an argument. A child could justifiably assert and back up that my desk is brown without being able to formulate a logical argument.

Non-logical good reasons are those reasons that support a claim with something other than another claim.This is an important feature of what a later section calls


1 6 2

basic sources. But non-logical reasons can always be cast into logical reasons, that is, into prima facie arguments.

Justification-ConditionsAn adequate account of prima facie reasons requires

using justification-conditions to stipulate the sense ofpropositions as well as showing and explaining how theanalysis of sources yields the justification-conditionsfor propositions. The justification-conditions for aproposition are the conditions under which one my justi-

19fiably believe the proposition. We can distinguishbetween the universal justification-conditions and theessential justification-conditions. The simple case isfor the justification-conditions for ascriptions of aconcept (propositions of the form x is an a) :

The justification conditions of a concept are a fixed logical feature of that concept. By way of the essential justification conditions they tell us on what "immediate" grounds we can ascribe the concept to things . . . and by way of the universal justification conditions they tell us how to apply further knowledge we have acquired to the ascription of the concept in more indirect ways.20

The essential justification-conditions uniquely specifythe sense of the proposition; the universal justification-conditions are generally applicable patterns of inference.The essential justification conditions are expressed interms of the grounds that are unique and most immediatefor the proposition.


163Pollock gives an account of essential justification-

conditions for the simple statement "x has a red surface facing me;" which is justified on the basis of my belief that x has a surface that appears red to me, i.e. on the basis of perception. Sense-perceptions are the unique and basic grounds for beliefs about colors. It might be that I believe that x has a red surface facing me because someone told me so while my eyes were closed, but in this case I am simply relying on someone's sense perceptions. So, the essential justification-conditions for "x has a red surface facing me" and its complement are;

(1) "S sees x & part of the surface of x looks red to S"’"is a prima facie reason for £ to Eelieve that x has a red surface facing him? (2) "S sees x & no part of the surface of x looks red to S" is a prima facie reason for S to Eelieve that x does not have a red surface facing him.21 ”The universal justification-conditions are the same

for all concepts. These specify more indirect ways in which a proposition can be justified. The evidence of testimony is an example of universal justification- conditions. Under appropriate circumstances, testimony can prima facie justify virtually any proposition.

Let us take "John Hinckley is criminally insane" as a prototypical statement supported by the expert testimony of a court psychiatrist. Before I can assert this statement on the basis of expert testimony, certain precondi


164tions must be met; for example, I must hear (or read) theexpert's statement on Hinckley's sanity. Thus:

"John Hinckley is criminally insane" is justified for S on the basis of expert testimony if S hears it uttered by a duly qualified authority in a relevant field (psychiatry or psychology).

Defeaters unique to this type of justifications are wellknown and appear to fall into these categories:

1. The expert does not really have the relevant professional qualifications; e.g., might not be qualified in the correct sub-field

For example, the expert might be well trained in psychology but not in the psychology of criminals.

2. The expert did not undertake the necessary studies.

The expert might be capable, but could have neglected to interview Hinckley, or to administer the appropriate personality tests. Both defeaters (1) and (2) could be summarized in:

The testimony was given under circumstances in which expert testimony has proved unreliable in the past.But all defeaters might not relate to the expert.

A third defeater is:3. The field in question has proven unreliable

in making judgments of the sort in question.An opponent of the expert could argue that psychology has not developed the necessary methods to evaluate the criminally insane. Other examples of universal justification conditions are formal patterns of inference and categorical deductions.


165The key question for the logical criticism of a

proposition is to determine just what are the essential justification-conditions for that proposition. This is a fairly simple task for Pollock's prototypical case,"x has a red surface facing S.” This proposition can only be justified by vision, i.e. "x appears red to S" (or by means of remembering that x looked red, the testimony of someone who sees that x looks red, etc.). A complete theory must also explicate the universal justification- conditions applicable to all concepts.

It is critical to choose the correct level of generality for a statement of the essential justification- conditions. The proposition that "x is spherical" can be justified either by "x looks spherical to S" or by "x feels spherical to S." A more elegant formulation is that "S perceives x as spherical," which concisely states the essential justification-conditions for "x is spherical." But "x looks spherical to £" is not an adequate statement of the essential justification-conditions of "x is spherical," even though "x looks spherical to S" justifies "x is spherical.

Further, a statement of the essential justification- conditions does not mention contingent facts. That a bobcat makes a certain sound is not relevant to the essential justification-conditions of the concept of a bobcat.


166Someone had to know what a bobcat was in order to know that a bobcat makes that sound. It is possible that varying communities might construe the concept of a bobcat differently. For example, persons in a given culture might develop a concept of the "fearsome howling animal" and later discover, as a contingent fact, that it is a bobcat that makes the fearsome sound. Both concepts refer to bobcats, but the concepts differ according to the community. There is only one set of essential justification- conditions for a concept or proposition, and that is the set of conditions that uniquely specifies the sense of the proposition.

Multiple CriteriaThe tools are now available to solve the problem of

multiple criteria of evidence. The problem is to show how the various considerations given in support of claims in marketplace argumentation constitute evidence for those claims. The problem arises because of the lack of a deductive connection between most kinds of claims and the sorts of considerations that can be offered in favor of or against those claims. For example, as Brummett implied, deduction cannot resolve the gap between appearance and reality.^

The solution is that evidence logically supports a claim. The logical link is due solely to the meanings of


167the propositions given as evidence and of those stated as claims.

Under the traditional theory, logical relations can only be established by stating the truth-conditions. Therefore, in order to establish a logical connection between evidence and claim, traditional logic often has to reduce the truth-conditions of a claim to the justification-conditions of the claim.

However, the new theory of meaning abolishes this problem. Stating justification-conditions for a claim uniquely specifies the meaning of the claim. The logical relation between claim and grounds is immediate and does not require the intermediate steps of reduction or indue-

A i

tive generalization.The prima facie reason therefore provides a model

for nondemonstrative inference. Let us suppose that P is a prima facie reason to believe Q, and that a person believes Q on the basis of P. This person is justified in his belief of Q when he believes no propositions that are defeaters for the inference. Likewise, a rhetor may justifiably assert Q on the basis of P if he believes no relevant defeaters, and an audience may justifiably accept the inference if they believe no defeaters.

The logical gulf between evidence and claim is now dissolved. However, the claim is not equated with or


16 8

reduced to evidence. The essential justification-conditions of a claim uniquely specify the sense of the

25claim, but they are not the same as the claim.For many types of propositions, including most

propositions of value and policy, it is not reasonable even to ask for truth-conditions. For this reason, it is not reasonable to expect truth-functional logic to be relevant to the critical evaluation of many kinds of arguments. Most logicians' response to this situation is to claim that many arguments are simply not logical; that is, that an argument that fails to meet the standards of truth-functional logic must be defective. This thesis advocates instead that the limitations of truth-functional logic call for indicting the traditional view of logic.

The myriad sorts of evidence that can be adduced in favor of claims count as evidence because of the logical connection between a proposition and the justification that can be offered for the proposition. There are as many possible first-order criteria as there are possible arguments. But all of them count as evidence because of the logical connection: the first-order criteria all meetthe same second-order criterion. The arguments of marketplace argumentation are logical arguments subject to logical criticism.


169The fundamental problem for logic is no longer to

prove that marketplace arguments can really carry justification for claims. The task of logical criticism is now to correctly state the claim's justification-conditions and the exceptions to them, determine if these conditions have been met in a given case, and examine the claims of purported defeaters if there is a special reason to do so.

Prima Facie Arguments and Probable ArgumentsThe rhetorical tradition typically divides arguments

into necessary arguments and probable arguments. This conception, although sound in its basic insights, founders on the inability to provide a clear definition of probability (something clearer than "non-necessary") and a clear idea of the relevant criteria for probability- statements. The prima-facie argument appears to encompass all of what rhetorical theory has meant by "probable argument:" which is, presumably, the idea of an argument that supports its claim well but not well enough to be known for certain.

The prima facie reason is not a probable reason; neither is it improbable. The concept of probability does not enter into the concept of the prima facie reason. A deduction based on a probable argument might be represented in this way:

If P, then probably Q


170Or it might be expressed this way:

Probably (If P then Q)or:

28(If P, then Q) is true with probability r.A prima facie reason might be better symbolized in this way:

If P, then Q (Unless R, S, T, . . .)The difficulties associated with clarifying the nature of probable argument do not arise with prima facie arguments.

It would still be possible, perhaps with the aid of some minor redefinition, to equate probable arguments with prima facie arguments; indeed, such an equation squares well with ordinary usage. People often use "probably" interchangeably with "presumably." But using two technical terms to mean the same thing is pointless, so let us instead use "probable" to mean "more likely than not."In this sense, a probable argument is a subclass of the prima facie argument: "If P, then probably Q" is a claimthat can be established prima facie, presumably by inductive evidence.

The prima facie reason provides a model of useful arguments. It does not require certainty or measurable probability. It is not founded on an unrealizable ideal of formally necessary inference. And it is a simple, conceptually clear model of inference.


171Grounds and Sources

The evidence in a sound argument supports the conclusion because of the meanings of the propositions. The other task for a logical theory of argument is to discuss where evidence comes from.

Logical Types and SourcesA rhetor who makes a claim and justifies it adduces

evidence for the claim. The evidence can be expressed in a proposition. Thus, what Ehninger and Brockriede call the main proof line simply moves from evidence to claim. The evidence has to be of one type or another. Ehninger and Brockriede's "evidence" corresponds with Toulmin's "data." The warrant tells why evidence of a particular logical type supports the claim. For example, if the evidence for "Pedersen is (probably) not a Roman Catholic" is "Pedersen is a Swede," we want to know why statements about Pedersen's nationality justify a statement about this religion. The warrant and backing tell

27us.The evidence adduced for the claim of a good argu

ment will normally be something that the speaker believes justifiably. The speaker's knowledge must have a source. Thus, if I visit Mt. Washington and say that it is high, my evidence for this claim is seeing that it is high.


172The evidence implicitly appeals to the source. In

this example, the source is visual perception. This is another way of saying that a statement about visual perception is used to justify a claim about a characteristic of a physical object: a claim of one logical type supports a claim of another logical type. The function of the warrant and backing is to authorize the inference.To do this, the warrant must state the relevant logical characteristics of visual perception (the source) that enable statements about visual perception to justify statements about physical objects.

This point can be summarized in the principle that the warrant and backing indicate general principles or features of a source of knowledge by virtue of which data drawing on that source justify a certain type of claim.To recall Toulmin's discussion of type-jumps and re-work it into the terminology of sources is a useful way of making this insight concrete:

"We make assertions about the future, and back them by reference to data about the present and past" means that knowledge about the present and past is our source for knowledge about the future."We make general assertions about nature, and back them by the results of particular observations and experiments" means that particular observation and experiments constitute our source of knowledge for general assertions about nature; i.e., induction is our source for contingent generalizations."We claim to know what other people are thinking and feeling, and justify those claims by citing the


173things that they have written, said and done" means that behavior and signs of behavior constitute our source for knowledge about other people's thoughtsand feelings.28

A speaker giving evidence for a claim either explicitly or implicitly appeals to a relevant source or sources of knowledge.

Sources of knowledge may vary in generality. Sense- perception is our source for knowledge of the physical world; vision is a type of sense perception. Data from one source might in turn be supported by further claims based on other sources: if H claims that C on the basisof S's testimony, the data for this claim can be supported by evidence that H actually heard S's testimony and evidence that S had a basis for his testimony. Other evidence is not justified by further claims: "I see something red" is supported by my seeing something red, which is not a proposition.

Sources can also be distinguished according to whether they are unique sources for a certain type of claim. Vision is one unique source for statements about colors, but is not our unique source for statements about shapes: we can also learn about shapes by touch. It isimportant to identify unique sources because they let us specify the essential justification-conditions of propositions .


174Since the number of sources is unlimited, and the cor

responding number of logical types likewise is unlimited, there can be no question of discussing all of them. It is important to discuss the unique sources for important types of arguments and the basic sources which can terminate a regress of justifications.

Basic SourcesMost of the propositions that a rhetor might set

forth can be justified by other propositions. An inductive generalization can be justified by a group of particular propositions. A claim that Pedersen is not a Catholic can be supported by propositions about his Swedish ancestry and the percentage of Swedes who are Roman Catholic. That is, the data and warrants that will be given in a candid layout of an argument are propositions. In a sound argument, each of the propositions given as data or warrant can in turn be supported or justified.

However, the original claim can be justified only if the regress is terminable, only if the regress can be grounded. Logical justifications cannot be grounded in standards of effectiveness or consensus. So, some other type of grounding must be sought for claims. Let us say that basic grounds (coming from basic sources) underly all justifiable claims. Ordinary, marketplace arguments do


175not usually make explicit appeals to these grounds. The interlocuters in a dialogue usually reach agreement before probing such fundamental stages of the argument. Rhetors often find common ground with their audiences without needing to appeal to basic sources.

Nonetheless, if a rhetor or a dialectician (or their critics) wishes to lay out an argument, with all underlying justifications given in full, these basic grounds should be stated as the ultimate justifications for the claim. If no basic grounds are available, the claim is not justified.

The regress of justifications ends when a proposition can be justified without appealing to another proposition. Instead, the claim is grounded in a source. The history of philosophy teaches that there.are four such sources:

2 9external perception, self-awareness, memory, and reason. This list cannot be adequate; we must also often rely on the testimony of others. There are, presumably, sources for value arguments and aesthetic arguments.

Some of these sources are more basic than others. My memories are often memories of things that I have perceived. But this feature of sources is of no immediate significance. The important point is that all of these can justify propositions without requiring appeal to further propositions.


176Each basic source can ground a proposition in some

thing other than a proposition. Thus, my claim that I am happy at the moment is grounded in my feeling happy now (self-consciousness or self-reflection). The library room in which I am sitting has grey walls; I can support this claim by appealling to my perception of grey. My perception of grey and my feeling of being happy are not propositions; yet they can support propositions. Thus, the

30regress of justifications can reach an end. It is also important to distinguish sources according to their quality: some basic sources might not convey justificationbecause they are too unreliable. A subsequent section takes up this problem.

Sense PerceptionThe wall in front of me is brown. My justification

for this belief is not some other belief, but is simplythat the wall does indeed appear to me to be brown. Itis, of course, logically (truth-functionally) possiblethat the wall is not brown but that I am the victim ofsome peculiar optical trick, but I have no reason tobelieve this. As Chisholm noted, it is more reasonableto accept the testimony of my senses than to reject it,unless some reason leads me to doubt it in a particular

31case. And as Price notes, "There are perceptual mistakes, and there are several different kinds of them.


177But we have means of detecting them. Roughly speaking, wedo it by finding that the expected consequences of our

32belief do not occur.Sense perception is the essential source of all of

what we know about the world outside our own minds. We know some things about the world because, for example, we remember them, but these are memories of sense perceptions. So, the characteristics of the claims we can justifiably make about the physical world are inextricably tied up with sense perception.

Self-reflection or self-consciousness is the basic source for what we can justifiably claim about out own feelings, perceptions, ideas, and so forth. Selfreflection justifies certain propositions but is not itself a proposition. And, self-reflection is essential to the justification of all claims about one's own internal states and events. I feel my own feelings and think my own thoughts; no one else can do so for me, and no one else is aware of my thoughts and feelings in the same way I am. Other people are aware of my behavior and infer what I am feeling or thinking on the basis of that behavior. Further, they can tell me what they observe of me and what conclusions they draw from those observations.

Nonetheless, my own awareness of my feelings is primary: other people can point out to me that I am tired.


178I may not have noticed feeling tired until someone else pointed it out. But once I notice feeling tired, my justification for believing that I feel tired is that I feel tired, not that someone told me I feel tired.

Similarly, a child is taught to call himself hurt under certain conditions. But once he learns what it means to be hurt, his justification for claiming to be in pain will be the feeling of pain.

Self-reflection is our unique source of justification for claims about one's own feelings or thoughts, although it is subject to correction by others under certain conditions .

Memory. Memory is my source of knowledge for my personal past. As is the case with self-consciousness, other ways of knowing about my personal past are dependent on memory. However, they are not dependent only on my own memory, but also on the memory of others. For example, my parents' statements about my childhood are founded on their own memories.

An apparent counterexample is knowledge of artifacts. There are some things that I know about my past because of, say, photographs. I recently saw a photograph of myself at Natural Bridge when I was a boy. This provides evidence that I was at Natural Bridge, and does not seem to depend on memory.


179

Nonetheless, I remembered that I was at Natural Bridge after I saw the photograph. Furthermore, I remember what I looked like as a boy; otherwise, I would not recognize myself in the picture. I remember what Natural Bridge looks like; otherwise I would not recognize it in the picture either. The photograph seems merely to confirm the memory, but does not replace memory as a source of knowledge about my past.

My birth certificate is another matter, as its testimony about the time and place of my birth does not seem to depend on anyone's memory. In this case, there is no continuity of personal awareness. Cases such as this are properly regarded as borderline cases between personal memory and historical knowledge.

Reason. Reason is characteristically our source of knowledge of a priori truths— the theorems of deductive logic and mathematics, for example. It is the unique foundation for justification that depends entirely on truth-functional relations. For such relations, e.g. the relation that:

P is true;(If P, then Q) is true;Therefore, Q is true.

no more fundamental justification can be offered than the truth's self-evidence, or its derivation from self-evident premises.


180The truths of reason also yield universal

justification-conditions. It is a truth of reason that:P or R;Not-R?Therefore, P

So, if I know that I am in either Yorktown or Grafton, and that I am not in Grafton, I can justifiably assert that I am in Yorktown. If I know that my typewriter is not working either because it is dirty or dry of lubrication, and on closer inspection find that it is not dirty, I can conclude that it is dry of lubrication. These two inferences do not make use of any concept's essential justification- conditions .

The testimony of others. It is, on reflection, surprising how few of a rhetor's claims can be justified merely by the rhetor's own resources of perception or reason. Rhetors routinely rely on the opinions of others. Rhetorical theory traditionally distinguishes between testimony of experts and that of eyewitnesses. But the testimony of other people is not the unique source of any type of proposition. Nonetheless a rhetor can justify a proposition by appeal to the testimony of experts or witnesses. This testimony is not a belief or claim of the rhetor's, and therefore stops the regress of justifications .


181As Price states, "Indeed, each of us depends on

testimony for almost all that he claims to know aboutanything which is beyond the range of his own first-handobservation and memory; and one of the most importantfunctions of memory itself is the remembering of what wehave learned from other people by means of speech and

33writing." Price suggests that "Surely every person,just because he is a person, has at least a prima facieclaim to be believed when he makes a statement? Thisclaim is not of course indefeasible." The relevant policyseems to be, "Accept what you are told, unless you see a

34reason to doubt it." A reasonable person will accept the evidence of testimony as a prima facie but not conclusive reason for an opinion. Presumably, experience teaches one to give more credence to the testimony of some persons rather than others.

The bases of moral judgment. Although moral judgments are influenced by factual considerations, they involve considerations that cannot be explained by reference to, say, sense perception or memory. Nor can testimony alone account for moral judgments: we might saythat adultery is wrong because the church says so, but what grounds does the church have for its judgment?

To simplify the discussion of this issue, I will follow Rawls in assuming that people can acquire the


182complex set of faculties that permit them, in their reflective moments, to make reasonable decisions on matters of value. People can often make reasonable moral decisions in particular cases, and it is possible for the ethical theorist to develop principles that summarize or account for these decisions. These principles are, in a sense, inductions from individuals' considered moral judgments :

Let us assume that each person beyond a certain age and possessed of the requisite intellectual capacity develops a sense of justice under normal social circumstances. We acquire a skill in judging things to be just and unjust, and in supporting these judgments by reasons. Moreover, we ordinarily have some desire to act in accord with these pronouncements and expect a similar desire on the part of others.Clearly this moral capacity is extraordinarily complex. To see this it suffices to note the potentially infinite number and variety of judgments that we are prepared to make. The fact that we often do not know what to say, and sometimes find our minds unsettled, does not detract from the complexity of the capacity we have.

Now one may think of moral philosophy at first (and I stress the provisional nature of this view) as the attempt to describe our moral capacity; or, in the present case, one may regard a theory of justice as describing our sense of justice.35Kohlberg suggests that people develop an increasingly

sophisticated ability to make moral judgments. Childrenbegin by making moral judgments on the basis of reward andpunishment but ultimately learn to make judgments onhigher levels of moral development. The ability to makemoral judgments is developed by the influence of parents,


183siblings/ friends, and teachers. Kohlberg emphasizes thatthe ability to make higher order moral judgments can betaught. His research thus suggests that the capacities

36required to make moral judgments are learned skills.

Onigue Non-Basic SourcesSince it would be hopeless to try to list every type

of grounding that could be offered in support of a claim, this section gives only a few examples of sources to illustrate how they fit into the scheme.

The data and backings of arguments using these grounds can always receive further support. These are, however, grounds that uniquely characterize a type of proposition. Induction, for example, is the unique source for contingent generalizations. However, it is not a basic source and does not end the regress of possible justifications. Thus, the particular propositions that support an induction normally can each be justified. The justification-conditions of all contingent generalizations are all derived from the characteristics of induction in general.

The behavior of other people's bodies is our unique grounding for claims about other people's feelings and thoughts. We cannot directly think other people's thoughts or feel their feelings: we must make inferences from their actions. But each claim about someone's


184behavior can be supported by the evidence of testimony, perception, etc., so behavior is not a basic source.

Likewise, knowledge of the present is the unique source for knowledge of the past; knowledge of the past and present is the unique source for knowledge of the future. Legal records are the unique source for knowing current laws. Demographic surveys are the unique source for demographic data.

The number of non-basic sources, unique or not, is limitless. The theory puts no special limit on how broad or narrow a source can be: my neighbor is a source forgardening advice, which advice is an example of expert testimony, which is in turn founded on my neighbor's inductions over years of gardening experience. For the logical critic, of course, the broader and more general sources will usually be of the greater interest.

The Justification of SourcesOne cannot justify a claim of knowledge merely by

saying that it comes from a reliable source. Empiricistphilosophy traditionally attacks problems of justificationby attempting to show on the basis of general principlesthat a source of knowledge is incapable of error. Thisapproach is generally fruitless; few sources of knowledge

37can even pretend to be incapable of error. One cannot simply claim that a source of knowledge is reliable. The


185further question remains# how do we know that it is

38reliable? Unless this question can be answered# the presentation in this chapter is circular.

The above discussion of basic sources assumes that the regress of justification ends when a claim is justified by something other than another proposition; that is, when the justification is non-discursive. As far as rhetoric is concerned this may be enough: justificationsfind a non-discursive foundation.

But the further question remains of how to know when a non-discursive foundation is any good, how to know that a source really conveys justification, how to tell good sources from bad.

An inductive generalization resolves the difficulty. Price writes that, for example, "We all do think that some questions can be conclusively settled by the evidence of sense-perception, for example, the question 'Are there any matches in this box?' By looking inside, it is possibleto know or be certain that there are some matches there,or none. G. E. Moore was surely right in maintaining thatwe know many propositions of this sort to be true eventhough we do not know what the correct analysis of them is."39

We may not know immediately, or on the basis of a universal principle, that sense perception (or memory or


186

testimony) is reliable. But I do know many things on the basis of sense perception. For example, I know that my sweater is scratchy, that the wall in front of me is brown, that my typewriter is making noise, and so on. Although I may or may not be sure how I know these things, my belief in them is more reasonable by far than my belief in an abstract philosophical position that would lead to a contrary conclusion. For instance, if I act on the basis of such specific knowledge, my actions usually turn out as I expect. When they do not, I can often discover a reason. Further, the apprehensions of one sense often confirm another; the typewriter both looks and feels solid. An inductive procedure like this establishes a prima facie, defeasible argument in favor of sense perception. Similar arguments can be proposed for science:I do know many things on the basis of science; my knowledge of these specific cases is more reasonable than my belief in any abstract principle leading to a contrary conclusion. Such an argument cannot be constructed for astrology; I know very little if anything on the basis of astrology; on those occasions on which I have acted on the basis of superstitious predictions, matters rarely work out as I expect.

Many but not all sources can also be tested reasonably well by inductive comparisons with other sources. I


187cannot test my moral intuitions by comparing them against sense perception. But I can test my memories againstartifacts such as photographs and souvenirs. If such comparisons establish my memories to be reliable in caseswhere comparisons are possible, I am inductively justifiedin considering my memories to be reliable in cases where comparisons have not been raade.̂ ®

For these reasons, general principles about the reliability of particular sources of knowledge are best viewed as inductive generalizations, not as deductive conclusions.^

Any philosophical endeavor must have a starting- point. Traditional logical theory starts with general principles about logic and expects marketplace argumentation to conform to them. The resulting theory turns out to be inadequate. So, instead, this thesis begins with the assumption that marketplace arguments are by and large sound, and seeks to account for them by principles. These principles, once arrived at, serve for the logical criticism of rhetorical arguments.

The Toulmin Model and Logical TypesPollock's theory provides a solution to the problem

of multiple criteria. It shows how ordinary arguments can be construed as logical arguments. But Toulmin's layout of arguments is a more complete system for the candid


188layout and analysis of arguments. So, how does Pollock's conception fit with Toulmin's?

Data simply are the immediate considerations in favor of a conclusion. The warrant or backing embodies the justification-conditions. The warrant might embody basic, unique grounds, universal justification-conditions, or non-basic unique grounds. The warrant is a generalization of the justification conditions for a given logical type of conclusion.

The rebuttal clause states defeaters. (Pollock, if faced with a choice, would probably diagram defeaters next to the warrant rather than next to the claim, but this hardly matters.) Modal qualifiers are stated as appropriate; "presumably" is an implicit or explicit qualifier for all claims supported by prima facie arguments.Examples will clarify the relationships between Toulmin's and Pollock's approaches.

Figure 3 assumes that S sees T, and analyzes the argument for "T has a red surface facing S." Because perception is an example of a basic, unique source, (W) and (D) need no further backing. Toulmin would choose a simpler, less formal way to state the warrant, but this is purely a matter of style. The grounds are of the logical type of statements of sense-perception.


189

Figure 3DIAGRAM OF ARGUMENT BASED

ON A BASIC SOURCE

(D) x looks red to S

So, (Q) Presumably, (C) x has a red surface facing S

Unless (R)x is lit by colored Tights/S is hallucinating . 7 .

Since(W)

"Part of the surface of x looks red to S is a prima facie “ reason for S to believe that x has a red surface” facing him."


190Figure 4 analyzes an example of Toulmin's that uses

a unique, non-basic source. This argument uses a uniquesource because decisions about citizenship can only besettled by appeals to a nation's laws. The warrant hereis merely a special case of the justification-conditions.The warrant could equally well be stated as:

A man born in a nation of the British Commonwealth: (list of nations) or naturalized by the British courts will be a British subject (unless . . .).

The "unless" clause takes care of Toulmin's "generally." The new formulation of the warrant becomes a statement of the justification-conditions for "Harry is a British subject" with only a few changes in form:

"Harry was born in a nation of the British Commonwealth or was naturalized by the British courts" is a prima facie reason to believe that Harry is a British subject.

The warrant is easily reducible to the justification-conditions of the claim.

The backing for this example makes an appeal tostatutes and legal provisions. The testimony of law-booksis founded on the deliberations, orders, etc. of theappropriate law-making bodies.

The data that "Harry was born in Bermuda" can, ifnecessary, be supported by a new body of arguments. Thesewill make use of the memory and testimony of Harry'sparents, the hospital birth records, and similar sourcesof the sort used for historical knowledge.


191

Figure 4DIAGRAM OF ARGUMENT BASED ON A

NON-BASIC SOURCE

(D) Harry was bom in Bermuda

So, (Q) Presumably, (C) Harry is a British subject

Unless, (R) Harry is a naturalized American/ Harry has renounced his British citizenship . .

Since(W)

A man born in Bermuda will generally be a British subject

On Account of (B)

The following statutes and other legal provisions . . .


192A final example, depicted in Figure 5, deals with a

universal justification-condition. All horses are animals by definition. The conclusion follows by principles of classification. The backing simply appeals to the truths of reason. There are no defeaters or rebuttals because, if the premises are true, the truth of the conclusion is beyond question.

In all of these examples, the warrant states the basis on which data can justify a claim. All warrants appeal to one of the grounds or sources— basic or non- basic, unique or non-unique— on which people rely for knowledge. These sources enable us to move from data to claim. We can justify claims about objects on the basis of perceptions because perception is our unique source of knowledge about physical objects. We make generalizations on the basis of induction because induction is our unique source for contingent generalizations.

The Prima Facie CaseThe foregoing account treats relatively simple argu

ments in isolation from one another. In practice, competent rhetoric often introduces a battery of related considerations in favor of a conclusion. Pollock emphasizes the logical and epistemological aspects of prima facie reasons. Reseller's method of dialectics includes a conception of prima facie reasons emphasizing the process


193

Figure 5DIAGRAM OF ARGUMENT BASED ON

UNIVERSAL JUSTIFICATION-CONDITIONS

(D) Molly is a horse So, (Q) Certainly, (C) Molly isananimal

Since(W)All horses are animals

On Account of (B)

The definisions of "horse" and "animal"


1 94

of making out a prima facie case in the face of opposingarguments. In rhetoric, the concern is normally for theestablishment of a body of arguments for a proposition.Broadened in this way, the positions taken in this chapterprovide an adequate account of the bases for evaluatingthe soundness of rhetorical argument.

Rescher introduces the notion of a contention that is"evidentially sufficient" (abbreviated ES) and which maybe used in support of a case. An evidentially sufficientcontention need not be decisively established, but "willin general be defeasible." Such a contention is similarto Pollock's prima facie argument. Rescher continues that"such ES-contentions must be of a sort that (1) there is asubstantial presumption of truth in their favor, and (2)the evidential force they are able to lend to the thesisin whose behalf they are adduced is sufficiently weighty

42that the burden of proof is shifted in its favor."Rescher notes that "There is, in most probative con

texts, a standing presumption in favor of the usual, normal, customary course of things." We expect presumptions "in favor of the senses, of memory, the inductive groundrules, etc. We accept the claims derived from these

43sources as veridical until proven otherwise." Logically speaking, the burden of proof shifts when the case of the


1 9 5

opposite side, taken as a whole, prima facie supports that side's views.

Rescher introduces a useful distinction between a prima facie case and a prima facie contention. Pollock's theory deals exclusively with prima facie reasons, that is, with relationships among individual propositions. Rescher adds the notion that a group of statements, expressed according to a procedure in a dialectical context, together constitute a case for a thesis. This distinction is as important for rhetoric as for dialectics. Rhetoricians often recognize the use of a body of discourse in favor of a proposition; this is particularly so with arguments favoring a proposition of value or policy. An entire body of considerations might, for example, together constitute a case for the current arms limitation agreement between the United States and the Soviet Union; it is not merely a matter of a single good reason supporting the proposition at issue. One wishes to accept the best supported and most reasonable conclusion.

ConclusionThe epistemological theories of Chisholm and Pollock

assume that we can have good reasons for believing things even if those reasons are not demonstrative. We are justified in holding beliefs on the basis of sense perception, memory, the testimony of others, and so on unless


196

there is a particular reason not to do so in a particular case.

Chisholm and Pollock both argue forcefully that there must be a third type of logical inference in addition to deduction and induction. This third type of inference is logical, in that a good argument is sound because of the meanings of its propositions. But, unlike traditional deductive arguments, prima facie arguments justify substantial, significant inferences. The study of substantial arguments requires us to correctly determine how various grounds or sources justify certain types of inferences and to identify the possible exceptions or defeaters for such inferences. We are justified in believing a claim supported by a prima facie argument unless there is a particular reason to doubt the inference in the given case.

This approach escapes some of the confusions that have surrounded the notion of a probable argument. An approach of this sort must be supplemented by the reminder that a soundly backed conclusion is usually supported by a body of arguments that together constitute a prima facie case for the conclusion. This approach appears to be adequate for the analysis of value arguments and questions of policy as well as for simple propositions of fact.


1 9 7

Notes Chapter V

1. Roderick M. Chisholm/ Theory of Knowledge/ 2nd ed. (Englewood Cliffs, N.J.: Prentice-Hail, Inc., 1977)and The Problem of The Criterion: The Aquinas Lecture,1973 (Milwaukee: Marquette University Press, 1973).

2. Donald C. Bryant, "Rhetoric: Its Functions and Its Scope," Quarterly Journal of Speech 39 (1953):418; see also 414.

3. John L. Pollock, Knowledge and Justification (Princeton, N.J.: Princeton University Press, 1974).

4. Stephen Toulmin, Richard Rieke, and Allan Janik, An Introduction to Reasoning (New York: Macmillan Pub-lishing Co., Inc., 1979), pp. 37-39; Stephen Toulmin, Knowing & Acting: An Invitation to Philosophy (MacmillanPublishing Co., Inc., 121b), pp. 97-115; Stephen Edelston Toulmin, The Uses of Argument (Cambridge: At the University Press, 1958) , pp. 13-14. Toulmin uses the expression "logical type" to apply exclusively to propositions (claims, backing, grounds, etc.) Compare Gilbert Ryle, "Philosophical Arguments," Collected Papers (London: Hutchinson & Co., 1971), 2:194-211 which uses "logical type" indiscriminately for concepts, propositions, and arguments. I will follow Toulmin's usage.

5. Chisholm, Problem of the Criterion, pp. 23-24.6. See also Toulmin, Rieke, and Janik, Introduction

to Reasoning, pp. 121-25.7. Pollock, Knowledge and Justification, p. 6.8. Toulmin, Rieke, and Janik, Introduction to Rea

soning , p. 121-25; see also Nicholas Rescher, Dialectics:A Controversy-Oriented Approach to the Theory of Knowledge (Albany: State University of New York Press, 1977),pp. 25-37 and Pollock, Knowledge and Justification, ch. 2.


1 98

9. Ch. Perelman and L. Olbrechts-Tyteca, The New Rhetoric; A Treatise on Argumentation, trans. John Wilkonson and Purcell Weaver (Notre Dame, Ind.: University of Notre Dame Press, 1971), 1st paperback ed., p. 3.

10. I am not using the term "good reasons" in thequasi-ethical sense sometimes associated with the term.

11. Aristotle, Analytica Priora, trans. A.J. Jenkin- son, in W.D. Ross, ed., The Works of Aristotle (Oxford;At the Claredon Press, 1928), 1:70a.

12. Pollock, Knowledge and Justification, pp. 39-40.13. Ibid., p. 40.14. Ibid., pp. 40-42. This theory does, of course,

have its origins in a long line of philosophical literature. See e.g. R.G. Collingwood, An Essay on Philosophical Method (Oxford: At the Clarendon Press, 1933) , pp.151-75; Thomas Reid, Essays on the Intellectual Power of Man, ed. and abridged A.D. Woozley (Macmillan and Co., Limited, 1941), pp. 329-357; Ryle, "Philosophical Arguments," p. 194 notes his debt to Collingwood.

15. Pollock, Knowledge and Justification, pp. 40-42;see also Panayot Butchvarov, The Concept of Knowledge (Eveanston, 111.: Northwestern University Press, 1970),p. 268.

16. Pollock, Knowledge and Justification, p. 44.Both Pollock's conception of the prima facie reason are heavily influenced by Rogers Albritton, "On Wittgenstein's Use of the Term 'Criterion,'" Journal of Philosophy 56 (1959) :845-47 and Norman Malcolm, '^Knowledge of Other Minds," Journal of Philosophy 55 (1958):969-78.

17. Pollock, Knowledge and Justification, p. 34.18. Ibid., pp. 33-36.19. See this discussion of this matter in ch. 2,

pp. 55-57.20. John L. Pollock, "Criteria and Our Knowledge of

the Material World," Philosophical Review 76 (1967) :35.21. Pollock, Knowledge and Justification, p. 85.


199

22. Pollock, Knowledge and Justification, pp. 56- 70; 81-95, 98-119; Pollock, "Criteria," pp. 33-46. I am assuming here that "spherical" is an ostensive concept whose immediate sense cannot be further specified except by showing examples of spheres. If I am wrong, it may be that the essential justification-conditions for "x is a sphere" would be: *”

S sees and/or feels x and x looks and/or feels to Sexactly like an object the points of whose surfaceare equidistant from a common center point.23. Barry Brummett, "Some Implications of 'Process'

or 'Intersubjectivity': Postmodern Rhetoric," Philosophy & Rhetoric 9 (1976) :34.

24o Pollock, Knowledge and Justification, p. 20.25. Ibid., p. 44. This point leaves some room for

controversy. If a speaker believes P but has overwhelming evidence for a defeater R, and fails to believe R, one might not want to consider his belief in P to be justified. On the other hand, it is unrealistic to expect people to notice all evidence for any and all possible beliefs.

26. This formulation is favored by Glen E. Mills and Hugh G. Petrie, "The Role of Logic in Rhetoric," Quarterly Journal of Speech 54 (1968):263.

27. Douglas Ehninger and Wayne Brockriede, Decisionby Debate (New York: Dodd, Mead & Company, 1963), p. 103.They are, of course, using Toulmin, Uses of Argument,ch. 3.

28. Toulmin, Uses of Argument, p. 219.29. See Rene Descartes, "Rules for the Direction of

the Mind," in The Philosophical Works of Descartes, trans.Elizabeth S. Haldane and G.R.T. Ross (New York: DoverPublications, 1955) , 1:33.

30. Although the pattern by which I have conceivedof and organized sources is my own, my opinions were influenced especially by Chisholm, Theory of Knowledge and H.H. Price, Belief (London: George Allen & Unwin,Ltd, 1969).

31. Chisholm, Problem of the Criterion, pp. 23-24.32. Price, Belief, p. 103.


20033. Ibid., p. 113.34. Ibid., pp. 114-16.35. John Rawls, A Theory of Justice (Cambridge,

Mass.: The Belknap Press of Harvard University Press,1971), p. 46.

36. Lawrence Kohlberg, "A Cognitive-Developmental Approach to Moral Education," The Humanist 32 (1972):13-16.

37. For an excellent discussion of this issue, see Frederick L. Will, Induction and Justification: An Investigation of Cartesian Procedure in the Philosophy of Knowl-

'“thaca, N.Y.: Cornell University Press, 1974), esp.

38. Chisholm, Theory of Knowledge, pp. 63-64.39. Price, Beliefs, p. 102.40. A totally different approach to justifying

sources is to establish their causal reliability. For example, one might call on psychology to establish the efficiency of the mechanisms of memory. Arguments would then run along the lines of "S perceives x to be a, andwould not perceive x to be a unless x were (certainly,

probably) a." Although I am not optimTstic about theories of this sort, such a theory (if successful) is just as capable of establishing the reliability of a source as is the theory I have defended, and would solve the problem of circularity with which I began the discussion of this issue. A particularly good causal theory is in Fred I. Dretske, Seeing and Knowing (Chicago: The University ofChicago Press, 1969), ch. 3.

41. This brief synopsis of a difficult argument draws on the tradition of G.E. Moore as analyzed and summarized in Chisholm, Theory of Knowledge, ch. 7 and Toulmin, Uses of Argument, ch. 5. Moore's own positionis outlined in Philosophical Papers (London: George Allen& Unwin Ltd, 1959H

42. Nicholas Rescher, Dialectics: A Controversy- Oriented Approach to the Theory of Knowledge (Albany,N.Y.: State University of New York Press, 1977), p. 29.

43. Ibid.


CONCLUSION

Rhetorical discourse often gives reasons in favor of claims. Critics wish to evaluate the quality of those reasons. An important function of rhetorical theory is to guide the evaluation of the reasons that a rhetor can present to buttress claims. But given the inherent failings of the human capacity for knowledge, one cannot expect practical arguments to meet a philosopher's ideal of perfection. Still less can a rhetor, perhaps faced with the need to try to persuade others without time or resources to acquire the best evidence available or to evaluate its significance, meet the philosopher's limited standards. Traditional logic is not adequate to evaluate many marketplace arguments or, worse, arguments that do not meet the narrow standards of traditional logic might be thought to be necessarily faulty.

The major problems addressed in this thesis are that traditional inductive and deductive logic do not account for the standards typically relevant to the criticism of marketplace argumentation, that recognition of this inadequacy has led many theorists to abandon logical studies entirely in favor of dialectical and rhetorical views,

201


202and that as a result criteria adequate to judge the soundness of rhetorical arguments have not been available.

The inadequacy of traditional logic has long been well known to rhetorical theorists. On the other hand, attempts to find a suitable alternative to guide the evaluation of arguments have been frustrated. The reasons for their lack of success is that the deep reason for the inadequacy of traditional logic, its foundation in an inadequate theory of meaning, has received little careful attention.

Logicians traditionally assume that a logical argument is sound because of the meanings of the terms and propositions that it uses. But they also assume that the meaning of a proposition can be given only by stating its truth-conditions. This inevitably leads to the attempt to establish many, perhaps most, important conclusions by reductive analysis: by trying to show that the claim isabout the same things as the evidence. Statements about objects are merely statements about sense perceptions.The result is a trivial logic that misrepresents the nature of evidence and argumentation. This problem immediately affects the study of rhetoric because the rhetorical theorist or critic cannot for a moment ignore the practical, ambiguous, fallible arguments of marketplace


203

argumentation, with their implications for action and demands for immediacy.

The need left unfilled by traditional logic has led rhetorical theorists to seek non-logical logics. Some writers attempt to ground standards for soundness in the concept of efficacy: the quality of the argument isevaluated in part by its effectiveness in moving an audience. Others have tried to ground standards of effectiveness in the notion of sound dialectical procedures.But by their very nature, such approaches cannot meet the needs of logical criticism.

Dialectical and rhetorical standards are, properly speaking, intended to facilitate the study of the communication contexts and processes in which people argue. A dialectician might claim that the best procedures for exchanging arguments are those that guarantee fairness and equal opportunity for all parties. A rhetorical point of view studies the persuasiveness of the arguments: onepresumes that an argument that is persuasive for all human beings, or persuasive for a particularly appropriate group of people, or persuasive in contexts of special significance, is in some way a better argument.

Dialectical and rhetorical views yield significant insights into the nature of the situations in which people argue, the importance of their arguments, and the role of


2 0 4

argumentation in societies. Also, one certainly does expect that the best of competing arguments is the one that gains acceptance in the context of fair discourse; likewise one at least hopes that the best of competing arguments will be effective in moving an audience to action. But the dialectical and rhetorical views do not embody logical standards than can be applied within a communication context, and they do not develop criteria of soundness by which arguments can be judged.

All the same, rhetorical and dialectical standards have to stand alone unless a more adequate logic can be discovered. Even if they are conceptually flawed when applied as logics, the dialectical and rhetorical perspectives will have to do unless an adequate conception of logic can be developed.

Butchvarov reviews several models that attempt to explain and account for good reasons that are not logical reasons, i.e., not sound solely because of the meanings of terms and propositions. Toulmin's is the best developed, most promising account of multiple criteria that does not claim that the soundness of arguments is entirely due to meanings. Toulmin's original conception in The Uses of Argument does not assume that marketplace arguments are faulty because they are not deductive. Toulmin claims instead that good arguments are for the most part the


2 0 5

arguments that knowledgeable and qualified people in the appropriate fields judge to be good: the theory of critical evaluation must adapt itself to practice. His theory fails only in that logical types and evaluative criteria do not turn out to be strictly field-dependent in the way he thought, although they are field-variant. But his notion of logical types, divorced from strict field- dependence, provides a foundation for a more adequate theory.

If logical connections are explained by a superior theory of meaning, that the sense of propositions can be given by stating their justification-conditions, these problems are resolved. Sound marketplace arguments are sound because they are logical; that is, because of the meanings of their propositions. There is automatically a logical connection between a claim and the evidence for a claim, because the justification-conditions uniquely specify the meaning. Further, justification-conditions can be stipulated for propositions of value and policy, to which the application of traditional logic is problematic.

Logic no longer has the twin tasks of stating justification-conditions for a proposition and then proving that the justification-conditions really justify the proposition. The task for logic is now to correctly state the justification-conditions and see if they have been


2 0 6

met. Stating the essential justification-conditions for a proposition is a matter of studying the unique sources for various logical types of propositions and analyzing how these sources yield justification for the proposition. In most cases, justification-conditions can be stated in schematic terms. In other words, a test case can be analyzed, e.g. "x has a red surface facing S," and its justification-conditions determined. The justification- conditions for other propositions of the same logical type, in this case statements about color, will in general follow a similar pattern and can be stated in a similar way. Developing and applying such schematic statements of justification-conditions should become a principal task of logical theory and criticism.

A most important task is to develop explicit criteria for the evaluation of arguments over matters of value and policy. The proposed approach to logic is as well suited to arguments of value and policy as it is to propositions of fact, but much of the work involved in developing the details of the criteria for judging arguments over values and policies remains to be done. It is essential to identify the origins and backings for such arguments, to show how such arguments can be justified, and to develop clear statements of the essential justification-conditions for typical arguments of value and policy.


207

The major value of the new approach to logic is its great breadth of application. It becomes theoretically possible to evaluate arguments on matters of contingent fact and also arguments of value and policy. Logic is no longer limited to the analysis of tautologies and inductions. Logical studies can be adequate for the analysis of the broad range of marketplace argumentation.


208

BIBLIOGRAPHY

Albritton/ Rogers. "On Wittgenstein's Use of the Term'Criterion,'" Journal of Philosophy 56 (1959):845-57.

Anderson, Ray Lynn, and Mortensen, C. David. "Logic and Marketplace Argumentation." Quarterly Journal of Speech 53 (1967):143-51.------------------------

Aristotle. Analytica Priora. Translated by A. J. Jenkin- son. In The Works of Aristotle, edited by W. D. Ross, vol. 1” Oxford: At the Claredon Press, 1928.

Aristotle. Rhetoric. Translated by W. Rhys Roberts. New York: Random House, The Modern Library, 1954.

Arnauld, Antoine. The Art of Thinking: Port-Royal Logic.Translated by James Kickoff and Patricia James. Indianapolis: The Bobbs-Merrill Company, Inc., 1964.

Aune, James A. "The Contribution of Habermas to Rhetorical Validity." Journal of the American Forensic Association 16 (1979):104-111.

Baird, A. Craig. Argumentation, Discussion, and Debate.New York: McGraw-Hill Book Company, Inc., 1950.

Barker, S. F. Induction and Hypothesis: A Study of theLogic of Confirmation. Ithaca, N.Y.: Cornell University Press, 1957.

Bitzer, Lloyd F. "Aristotle's Enthymeme Revisited." Quarterly Journal of Speech 45 (1959):399-408.

Brentlinger, William Brock. "The Aristotelian Conception of Truth in Rhetorical Discourse." Ph.D. dissertation, University of Illinois, 1959.

Brummett, Barry. "Some Implications of 'Process" or'Intersubjectivity': Postmodern Rhetoric." Philoso- phy & Phetoric 9 (1976):21-51.

Bryant, Donald C. "Aspects of the Rhetorical Tradition— I: The Intellectual Foundation." Quarterly Journal of Speech 36 (1950):169-76.


2 0 9

Bryant, Donald C. "Rhetoric: Its Functions and ItsScope." Quarterly Journal of Speech 39 (1953):401-24.

Burleson, Brant R. "On the Foundations of Rationality: Toulmin, Habermas, and the A Priori of Reason." Journal of the American Forensic Association 16 (X979: IT2-27.----------------------------

Butchvarov, Panayot. The Concept of Knowledge. Evanston,111.: Northwestern University Press, 1970.

Carnap, Rudolph. Introduction to Symbolic Logic and Its Applications. Translated by William H. Meyer and John Wilkinson. New York: Dover Publications, 1958.

Chisholm, Roderick M. "The Problem of Empiricism."Journal of Philosophy 45 (1948):512-517.

Chisholm, Roderick M. The Problem of the Criterion: TheAquinas Lecture, 1973. Milwaukee: Marquette Uni-versity Press, 1973.

Chisholm, Roderick. Theory of Knowledge. 2nd ed. Englewood Cliffs, N.J.: Prentice-Hall, Inc., 1977.

Collingwood, R.G. An Essay on Philosophical Method.Oxford: At the Clarendon Press, 1933.

Davidson, Donald. "Truth and Meaning." Synthese 17 (1967):304-23.

Dretske, Fred I. Seeing and Knowing. Chicago: The University of Chicago Press, 1969.

Dummett, Michael. Frege: Philosophy of Language. 2nd ed.Cambridge, Mass.: Harvard University Press, 1981.

Ehninger, Douglas, and Brockriede, Wayne. Decision by Debate. New York: Dodd, Mead & Company, 1963.

Ehninger, Douglas. "Validity as Moral Obligation." Southern Speech Journal 33 (1968):215-22.

Ellis, Brian. "Epistemic Foundations of Logic." Journal of Philosophical Logic 5 (1976):187-204.

Evans, Gareth, and McDowell, John, eds. Truth and Mean- ing: Essays in Semantics. Oxford: Clarendon Press,19761


210Farrell, Thomas J. "Knowledge, Consensus, and Rhetorical

Theory." Quarterly Journal of Speech 62 (1976):1-14.Farrell, Thomas B. "Validity and Rationality: The Rhe

torical Constituents of Argumentative Form." Journal of the American Forensic Association 13 (1977) :142-49.

Fisher, Walter R. "Toward a Logic of Good Reasons." Quarterly Journal of Speech 64 (1978):376-84.

Foster, William Trufant. Argumentation and Debating.Boston: Houghton Mifflin Company, Riverside Press,1908.

Frege, Gottlob. The Foundations of Arithmetic: A Logico-Mathematical Enquiry into the Concept of Number. Translated by J.L. Austin. New York: PhilosophicalLibrary, 1950.

Frege, Gottlob. "On Sense and Reference." Translated by Max Black. In Peter Geach and Max Black, eds., Translations from the Philosophical Writings of Gottlob Frege. Oxford: Basil Blackwell, 1960.

Frege, Gottlob. Translations from the PhilosophicalWritings of Gottlob Frege. Edited by Peter Geach and Max Black. Oxford, Basil Blackwell, 1960.

Graham, Gladys Murphy. "Logic and Argumentation." Quarterly Journal of Speech Education 10 (1924):350-63.

Habermas, Jurgen. "Theories of Truth." Translated by R. Grabau. Mimeographed. N.d.

Habermas, Jurgen. "Wahrheitstheorien." In Wirklichkeit und Reflexion (Pfullingen, 1973).

Hage, Jerald. Technigues and Problems of Theory Construction in Sociology. New York: John Wiley & Sons,Wiley Interscience Publication, 1972.

Harrod, Roy. Foundations of Inductive Logic. London: Macmillan, 1&74.

Kant, Immanuel. Critigue of Pure Reason. Unabridged edition, Translated by Norman Kemp Smith. New York:St. Martin's Press, 1965.

Kneebone, G.T. "Induction and Probability." Proceedings of the Aristotelian Society n.s. 50 (1950):27-42.


211Kneupper, Charles W. "On Argument and Diagrams." Journal

of the American Forensic Association 14 (1978):181-86.Lehrer, Keith. "Justification, Explanation, and Induc

tion." In Induction, Acceptance, and Rational Belief. Dordrecht: D. Reldel Publishing Company,1970.

Lewis, C.I. An Analysis of Knowledge and Valuation. La Salle, 111.: Open Court, 1971.

Lewis, Albert L. "Stephen Toulmin: A Reappraisal." Central States Speech Journal 23 (1972):48-55.

Locke, John. An Essay Concerning Human Understanding. Edited by Alexander Campbell Fraser. 2 vols. NewYork: Dover Publications, 1959.

Malcolm, Norman. "Knowledge of Other Minds." Journal of Philosophy 55 (1958):969:78.

McBurney, James H. "The Place of the Enthymeme in Rhetorical Theory." Speech Monographs 3 (1936):49-74.

McCarthy, Thomas. The Critical Theory of Jurgen Habermas. Cambridge, Mass.: The MIT Press, 1978.

McCarthy, Thomas. "A Theory of Communicative Competence." Philosophy of the Social Sciences 3 (1973):135-154.

McDowell, John. "Truth Conditions, Bivalence and Verifi- cationism." In Truth and Meaning: Essays in Semantics, edited by Gareth Evans and John McDowell.Oxford: Clarendon Press, 1976.

Mills, Glen E., and Petrie, Hugh G. "The Role of Logic in Rhetoric." Quarterly Journal of Speech 54 (1968): 260-67.

Moore, G.E. Philosophical Papers. London: George Allen& Unwin Ltd, 1959.

Mudd, Charles S. "The Enthymeme and Logical Validity." Quarterly Journal of Speech 45 (1959):409-14.

O'Keefe, Daniel J. "Logical Empiricism and the Study of Human Communication." Speech Monographs 42 (1975): 169-83.


212O'Keefe, Daniel J. "Two Concepts of Argument." Journal

of the American Forensic Association 13 [ISli):121-28.Perelman, Chaim. Le champ de 11argumentation. Brussels:

Presses Universitaires de Bruxelles, 1970.Perelman, Ch., and Olbrechts-tyteca, L. The New Rhetoric:

A Treatise on Argumentation. 1st paperback ed. Translated by John Wilkonson and Purcell Weaver.Notre Dame, Ind.: University of Notre Dame Press,1971.

Petrie, Hugh G. "Does Logic Have Any Relevance to Argumentation?" Journal of the American Forensic Association 6 (1969):55-60.

Pollock, John L. "Criteria and Our Knowledge of the Material World." Philosophical Review 76 (1967): 28-60.

Pollock, John L. Knowledge and Justification. Princeton, N.J.: Princeton University Press, 1974.

Price, H.H. Belief. London: George Allen & Unwin, Ltd.,1969.

Quine, Willard Van Orman. Elementary Logic. Rev. ed.New York: Harper & Row, Publishers, Harper Torch-books, 1965.

Quine, Willard Van Orman. Mathematical Logic. Rev. ed. Cambridge, Mass.: Harvard University Press, 1955.

Quine, Willard Van Orman. Philosophy of Logic. Englewood Cliffs, N.J.: Prentice-Hall, Inc., 1970.

Rawls, John. A Theory of Justice. Cambridge, Mass.:The Belknap Press of Harvard University Press, 1971.

Rescher, Nicholas. Dialectics: A Controversy-OrientedApproach to the Theory of Knowledge. Albany: State University of New York Press, 1977.

Reynolds, Paul Davidson. A Primer in Theory Construction. Indianapolis: The Bobbs-Merrill Company, 1571.

Rhodes, Jack, and Newell, Sara, eds. Proceedings of the Summer Conference on Argumentation! Falls Church,Va.: Speech Communication Association, 1980.


2 1 3

Rowell, Edward Z. "Prolegomena to Argumentation." Quarterly Journal of Speech 18 (1932):1-13; 224-48; 381- 405; 585-606.

Rowland, Robert. "Argument Fields." In Dimensions of Argument: Proceedings of the Second Summer Conference on Argumentation, edited by George Zregelmueller and Jack Rnodes. Annandale, Va.: Speech Communication Association, 1981.

Ryle, Gilbert. Collected Papers. London: Hutchinson &Co., Inc., 1971.

Ryle, Gilbert. "Philosophical Arguments." In CollectedPapers. London: Hutchinson & Co., 1971.

Scruton, Roger. "Truth Conditions and Criteria." The Aristotelian Society supp. 50 (1976):193-216.

Searle, John R. Speech Acts: An Essay in the Philosophy of Language. Cambridge: At the University Press, 1969.

Sillars, Malcolm 0. "Investigating Religious Argument as a Field." In Dimensions of Argument: Proceedings ofthe Second Summer Conference on Argumentation, edited by George Ziegelmueller and Jack Rhodes. Annandale, Va.: Speech Communication Association, 1981.

Sztompka, Piotr. System and Function: Toward A Theory ofSociety. New York: Hartcourt Brace Jovanovich,Publishers, Academic Press, 1974.

Toulmin, Stephen. Knowing and Acting: An Invitation to Philosophy. New York: Macmillan Publishing Co.,inc~T9'76.

Toulmin, Stephen. Human Understanding. 3 vols. Princeton, N.J.: Princeton University Press, 1972-.

Toulmin, Stephen Edelston. The Uses of Argument. Cambridge: At the University Press, 1958.

Toulmin, Stephen, Rieke, Richard, and Janik, Allan. An Introduction to Reasoning. New York: MacmillanPublishing Co., Inc., 1979.

Walsh, W.H. Metaphysics. New York: Hartcourt, Brace &World, Inc., A Harbinger Book, 1963.


214

Weaver, Richard M. Ideas Have Consequences. Chicago:The University of Chicago Press, 1948.

Wenzel, Joseph W. "Jurgen Habermas and the Dialectical Perspective on Argumentation." Journal of the American Forensic Association 16 (1979):83-94.

Wenzel, Joseph W. "Perspectives on Argument," In Proceeding of the Summer Conference on Argumentation, edited by Jack Rhodes and Sara Newell. Falls Church, Va.: Speech Communication Association, 1980.

Will, Frederick L. Induction and Justification: AnInvestigation of Cartesian Procedure in the Philosophy of Knowledge! Ithaca, N.Y.: Cornell UniversityPress, 1974.

Willard, Charles Arthur. "Argument Fields and Theories of Logical Types." Journal of the American Forensic Association 17 (1981) :129-4i>.

Willard, Charles Arthur. "Field Theory: A CartesianMeditation." In Dimensions of Argument: Proceedingsof the Second Summer Conference on Argumentation, edited by George Ziegelmueller and Jack Rhodes. Annandale, Va.: Speech Communication Association,1981.

Willard, Charles Arthur. "A Reformulation of the Concept of Argument: The Constructivist/InteractionistFoundations of a Sociology of Argument." Journal of the American Forensic Association 14 (1978):l2l-40.

Willard, Charles Arthur. "Some Questions about Toulmin's View of Argument Fields." In Proceedings of the Summer Conference on Argumentation, edited by Jack Rhodes and Sara Newell. Falls Church, Va.: SpeechCommunication Association, 1980.

Wittgenstein, Ludwig. Preliminary Studies for the"Philosophical Investigations," Generally Known as the Blue and Brown BookST New York: Harper & Row,Publishers, Harper Torchbooks, 1965.

Wittgenstein, Ludwig. Tractatus Logico-Philosophicus.Translated by D.F. Pears and B.F. McGuxnness. London: Routledge and Kegan Paul, 1971.

Wright, Crispin. "Truth Conditions and Criteria." The Aristotelian Society supp. 50 (1976):217-245.


2 1 5

Zarefsky, David. "Persistent Questions in the Theory of Argument Fields." Paper read at the meeting of the Speech Communication Association, November 14, 1981, in Anaheim, California.

Ziegelmueller, George W., and Dause, Charles A. Argumentation: Inquiry and Advocacy. Englewood Cliffs,N.J.: Prentice-Hall, Inc., 1975.

Ziegelmueller, George, and Rhodes, Jack, eds. Dimensions of Argument; Proceedings of the Second Summer Conference on Argumentation ̂ Annandale, Va.: Speech Communication Association, 1981.


VITA

William Douglas Harpine was bom in Washington,D.C. on September 15, 1951 and graduated from Oakton High School in Vienna, Virginia in 1969. He attended the College of William and Mary from 1969 to 1973, graduating with the degree of Bachelor of Arts in philosophy. He entered Northern Illinois University in the fall of 1973, receiving his Master of Arts in speech communication in 1974. He was an instructor at Iowa State University for a year in 1975 to 1976.While at Iowa State, he wrote his paper on "Stock Issues in Aristotle's Rhetoric," which was later published in volume 14 of the Journal of the American Forensic Association.

In 1976, he went to the University of Illinois to pursue his doctoral degree. He was a University Fellow during his first year, and served as Director of Contest Debate and Director of Campus Forums.

He returned to the College of William and Mary in 1979 as assistant professor of theatre and speech. He held this post for three years before accepting an appointment as assistant professor of communication at the University of Akron in 1982.

216


217

Mr. Harpine married Elaine Clanton in 1977. live with their son David in Tallmadge, Ohio.

They


Foundations of the Theory of Argumentation

Documents

adjacent pages

blurred pages

missing pages

copyright owner

usersthis reproduction

blurred copy

xerographic copy

northern illinois university