Conceptualizing Argument Quality Via Argument Structure

Gregory W. Boller, Memphis State University
John L. Swasy, American University
James M. Munch, University of Delaware
ABSTRACT - This paper discusses the construct of argument quality. It discusses the Areni and Lutz (1988) approach of using a Fishbein model of belief strength and belief valence as dimensions of the argument quality construct. This paper argues for a more specific use of the term argument, and reviews the Toulmin (1958) model of argument structure. We suggest that Toulmin's concepts of evidence, authority, and probability are important structural elements that need to be considered to better understand the concept of argument quality. Also, implications for conceptualizing argument elaboration in terms of the Toulmin structure are noted.
[ to cite ]:
Gregory W. Boller, John L. Swasy, and James M. Munch (1990) ,"Conceptualizing Argument Quality Via Argument Structure", in NA - Advances in Consumer Research Volume 17, eds. Marvin E. Goldberg, Gerald Gorn, and Richard W. Pollay, Provo, UT : Association for Consumer Research, Pages: 321-328.

Advances in Consumer Research Volume 17, 1990      Pages 321-328

CONCEPTUALIZING ARGUMENT QUALITY VIA ARGUMENT STRUCTURE

Gregory W. Boller, Memphis State University

John L. Swasy, American University

James M. Munch, University of Delaware

[Authors acknowledge support from the Avron B. and Robert F. Fogelman Academic Excellence Fund.]

ABSTRACT -

This paper discusses the construct of argument quality. It discusses the Areni and Lutz (1988) approach of using a Fishbein model of belief strength and belief valence as dimensions of the argument quality construct. This paper argues for a more specific use of the term argument, and reviews the Toulmin (1958) model of argument structure. We suggest that Toulmin's concepts of evidence, authority, and probability are important structural elements that need to be considered to better understand the concept of argument quality. Also, implications for conceptualizing argument elaboration in terms of the Toulmin structure are noted.

INTRODUCTION

The concept of argument quality is important although it has not been the direct focus of much interest by consumer researchers. Perhaps this is because the simple issue of whether strong messages are more persuasive than weak is not interesting. Nevertheless, argument quality plays an important role in research involving the Elaboration Likelihood Model (see Petty and Cacioppo 1986 for a description and review of ELM). Operationally, "argument quality" provides a reference point for assessing how other communication factors affect persuasion. For example, the persuasive effects of rhetorical questions have been inferred from the fact that in low involvement conditions, rhetoricals enhance differences in the persuasiveness of strong and weak message versions. Under low involvement conditions, rhetorical questions encourage receivers to diligently consider the true merits of the message arguments (Petty, Cacioppo and Heesacker 1981). Integration of the logic and compellingness of the message along with activated and generated beliefs underlie acceptance of the message conclusion.

However, Areni and Lutz (1988) point out that many researchers have not been clear or consistent regarding the conceptual meaning of argument quality. Definitions of argument quality typically rely on idiosyncratic manipulations of message content that ultimately affect message valence. For example, Hunt, Smith and Kernan (1984) manipulated argument quality by defining strong arguments as statements of tangible product features and weak arguments as statements of intangible features.

The thrust of the Areni and Lutz (1988) paper centered on showing that the argument strength manipulation employed in the often cited Petty, Cacioppo and Schuman (1983) study was confounded with manipulations of the desirability-of product attributes/benefits used in the message. The weak message focused on less desirable attributes/benefits (e.g., "floats in water," and "designed for the bathroom") whereas strong messages involved more desirable product features/benefits (e.g., "blade sharpness," and "smooth shave"). In terms of the Fishbein and Aizen (1975) model of belief structure, the valence (ei) of the attributes/benefits presented in the strong message was more positive than that of the weak. It is clear that the distinguishing difference between the strong and weak messages was not simply their logical compellingness.

The results of the Areni and Lutz (1988) study suggest that the conceptual meaning of argument quality mostly rests in the argumentative structure contained in the persuasive message. The purpose of this paper is to continue Areni and Lutz's discussion of the argument quality construct by (1) extending their conceptualization of argument structure, and (2) considering the potential relationship between argument structure and argument elaboration.

DEFINING ARGUMENT QUALITY IN TERMS OF ARGUMENT STRUCTURE

Following the Fishbein and Ajzen perspective, Areni and Lutz (1988, p. 198) define a persuasive message as "a set of arguments concerning beliefs that link the object with positive and negative consequences and evidence in support of those arguments." They go on to say,

Of particular importance in the present situation is that each target belief has associated with it a belief strength (bi) and a valence (ei). That leads to the characterization of argument quality as comprised of two separate components, argument strength and argument valence. Argument strength is defined as the audience's subjective probability that the attitude object is associated with some outcome or consequence.

In their view, argument strength is defined in terms of a syllogism (McGuire 1960) created by the supporting beliefs (belie& formed by the presentation of evidence). Areni and Lutz (1988, p. 199) write,

...Argument strength may be defined in terms of the likelihood that the conclusion of a logical syllogism is accepted. If the conclusion is ascribed the role of target belief in Fishbein and Ajzen's model, then Fishbein and Ajzen's support beliefs can be viewed as the premises in a syllogistic structure underlying the target belief.

In principle, we agree that the conceptual meaning of argument quality primarily resides in the structure of the argument. And in general, the syllogism provides a good starting point for examining the essence of argument structure. However, we believe that the usefulness of the syllogistic model is limited. Specifically, the syllogistic model does not provide clear distinctions between various components of a message, cannot adequately represent the complex structures found in most marketing messages, and does not provide much direction for examining potential relationships between argument structure and argument elaboration.

As an alternative, we present a different conceptualization of argument structure that overcomes the limitations associated with the syllogistic- model. And as an introduction to this model, we find it instructive to first consider the intended function of argument.

The Intended Function of Argument

Willard (1979; 1981) writes that the function of argument is to provide reasons for the assertion of particular claims, positions or beliefs. Kneupper (1979) and Burleson (1979) maintain that the function of argument is to establish a relationship between an asserted claim and its supporting reasons or evidence. In other words, argument serves to support the assertion of a particular claim.

In terminology more appropriate for marketing applications, Brockriede (1975) writes that the function of argument is to provide rationale for a leap from existing beliefs to the adoption of new ones. That is, arguments give people objective and rational reasons to create a particular set of beliefs.

According to Burleson (1979), arguers first "lay claim" on existing beliefs of those toward whom the argument is directed. That is, arguers make assertions of the nature, "here is something you don't know," or "here is something that is contrary to what you already know." Next, arguers attempt to validate these asserted claims by providing a carefully structured presentation of assertions in support of those claims (Toulmin 1958; Toulmin, Reike and Janik 1984). Generally speaking then, arguments are linguistic procedures which are intended to establish the validity of an asserted claim through the presentation of a highly structured set of reasons in support of that claim.

It is important to recognize that arguments are not simple claim assertions (e.g., Petty, et al. 1983), nor are arguments statements of evidence. An argument is a type of message presentation in which one class of assertions are used with intent to support assertions of another class. Certain assertions are ascribed the status of "evidence" when, and only when, they are used to support other assertions of a more general nature. Assertions which require logical support in order to be believed are ascribed the status of "claims." Specifically then, argument can ultimately be both defined and revealed by the logical relationships amongst message assertions (i.e., structure).

The Toulmin Model of Argument Structure

Instead of describing argument structure in terms of the syllogistic model (e.g., Areni and Lutz 1988; McGuire 1960), we use Toulmin's (1958) jurisprudence model. This model of argument structure is widely accepted by many argument theorists and forensic specialists, relatively simple to understand, and very applicable to marketing contexts (see Deighton 1985). Moreover, the Toulmin model offers a distinct advantage over syllogistic models of structure in that it allows one to examine the specific types of assertions used in an argumentative message.

Toulmin (1958) and Toulmin, Reike and Janik (1984) state that there are six crucial elements in an argument--claim assertions, grounds, warrants, backing, qualifiers, and rebuttals. For ease of discussion, we collapse these six elements into four components--claim assertions, evidence (grounds), authority (warrants and backing), and probability (qualifiers and rebuttals).

Claim Assertions Perhaps the best way to understand the Toulmin model, is to work through an example. Assume that we wish to construct an argument to support the claim, "Eating Kellogg's All Bran cereal may reduce your risk of rectal cancer.' This claim is an assertion about benefits of this brand that we put forth for public acceptance. Specifically, we want our target audience to believe this benefit claim. But suppose we simply made this assertion, and then left it at that? Although some people might believe this claim, most probably would not.

To support and validate our claim, we need to construct an argument. We need to offer reasons why people should believe that eating Kellogg's All Bran cereal may reduce their risk of rectal cancer. Specifically, we need to logically present evidence in support of our claim.

Evidence (grounds) In the Toulmin model, grounds are the specific facts that a communicator produces to support a claim. Stated differently, grounds are pieces of evidence which a communicator relies upon to "clarify and make good" an asserted claim (Reinard 1984). In advertisements, we often find that brand benefit claims are supported by statements about product characteristics (e.g., product attributes).

To keep our example simple, consider only one piece of evidence. For the grounding, we state that, "All Bran is made of 100% natural fiber." Thus, our argument reads: "Eating Kellogg's All Bran cereal may reduce your risk of rectal cancer because All Bran is made of 100% natural fiber."

However, this one piece of evidence may not be sufficient to validate our asserted claim. Some people reading our argument might ask, "What does natural fiber have to do with rectal cancer?" In argument theory terms, they are asking, "On what authority are you presenting this particular piece of evidence as support for your claim?" Now we need a warrant.

Authority (warrants and backing) Warrants are statements which serve as the "rational authority" needed to logically connect the evidence with the asserted claim (Toulmin, Reike and Janik 1984). They are statements which serve as the "justification...employed by a communicator to demonstrate that the claim legitimately can be inferred from the grounds" (Reinard 1984, p. 207).

Warrants are not intended to strengthen the grounds. That is, they are not meant to change a receiver's belief that the evidence is indeed factual. Instead, they are designed to show that using particular evidence to support a specific claim is appropriate and legitimate (Toulmin 1958). In essence, warrants specify a particular domain of evidence that can be legitimately used to support a claim. In order for a claim to be judged as valid, the evidence presented as support must be viewed as legitimate evidence. Or, in jurisprudence terms, the evidence presented in support of a claim must be "admissible evidence."

Returning to our example, we use the following statement as our warrant. "Diets rich in fiber can help reduce the risk of rectal cancer." Logically now, our argument reads: "Diets rich in fiber can help reduce the risk of rectal cancer. Eating Kellogg's All-Bran cereal may reduce your risk of rectal cancer because All-Bran is made of 100% natural fiber.'' Alternatively, we might order the argument, "All-Bran is made of 100% natural fiber, and since diets rich in fiber can help reduce the risk of rectal cancer, eating Kellogg's All-Bran cereal may reduce your risk of rectal cancer."

Now, is the argument strong enough to establish the validity of our claim? Someone reading our argument could ask, "How do we know that diets rich in fiber can help reduce the risk of rectal cancer?" That is, some people may perceive our warrant to be a relatively weak "rational authority." So, to further strengthen our warrant, and consequently enhance the validity of our claim, we can provide factual evidence called backing that supports the warrant For instance, we might use the statement: "Medical studies have proven that people who eat fiber rich foods are 60% less likely to develop rectal cancer than those people who don't."

Now, our more complex argument reads: "Medical studies have proven that people who eat fiber rich foods are 60% less likely to develop rectal cancer than those people who don't. And since diets rich in fiber can help reduce the risk of rectal cancer, eating Kellogg's All-Bran cereal may reduce your risk of rectal cancer because All-Bran is made of 100% natural fiber."

Our argument regarding the Kellogg's All-Bran claim is nearly complete at this point. We have provided reasonably strong support to establish the validity of our claim. But there still may be some people who do not accept our claim. These people might recognize circumstances that would render our claim invalid. If we do not acknowledge, at least in part, the existence of these circumstances, some people may reject our claim because they perceive us to be untruthful by reason of omission.

Probability (qualifiers and rebuttals) The probability component of argument is not necessary to validate a claim. This component is used by communicators when they want to acknowledge that their claim assertions are not "absolute" (cf. Burleson 1979; Habermas 1979; Jaccard 1981). In essence, this third component allows a communicator to tell an audience that he/she is being "truthful." And given that most consumers view advertising as extremely partisian, the use of qualifiers and rebuttals is often recommended to enhance credibility (cf. Deighton 1985).

You may have noticed the qualification we included in the latest version of our claim. We asserted that "eating Kellogg's All-Bran cereal may reduce your risk of rectal cancer." By using words such as "may" and "probably," we communicate the existence of circumstances that will render our claim invalid. We could take an additional step by providing the audience with one or more of these particular circumstances. That is, we could provide a rebuttal. For instance, we might say, "Genetic tendencies, however, sometimes outweigh factors such as diet." With respect to advertisements, rebuttals are generally included at the end of the ad, or in a nonprominant place as in a footnote.

Our completed argument can now be read as follows: "Medical studies have proven that people who eat fiber rich foods are 60% less likely to develop rectal cancer than those people who don't. And since diets rich in fiber can help reduce the risk of rectal cancer, eating Kellogg's All-Bran cereal may reduce your risk of rectal cancer because All-Bran is made of 100% natural fiber. Genetic tendencies, however, sometimes outweigh factors such as diet."

To recap our example, we first asserted a claim, "Eating Kellogg's All Bran cereal may reduce your risk of rectal cancer." Recognizing that many people would not believe this claim, as it stands, we constructed an argument to support our assertion. We used evidence (one specific grounding), authority (a warrant with backing), and probability (a qualifier and a rebuttal). We used these three components of argument to support and validate our brand benefit claim.

Building arguments according to the guidelines of the Toulmin model is a relatively easy task. In fact, the model might well be used by advertising copywriters as a tool for constructing argumentative ads. However, the Toulmin model is also useful as a diagnostic device. In particular, this model of argument structure can be used to examine the nature of strong versus weak manipulations of "argument quality."

REEXAMINING MANIPULATIONS OF ARGUMENT QUALITY

Areni and Lutz (1988) convincingly argue that many manipulations of argument quality are really manipulations of claim valence. That is, many researchers who puport to manipulate argument quality actually end up manipulating the inherent desirability of brand benefit claims. For example, Petty et al. (1983) used claim assertions such as "unsurpassed sharpness" and "eliminates nicks and cuts" for their "strong arguments." In contrast, they used assertions such as "floats in water" and "designed with the bathroom in mind" for their "weak arguments." Interested viewers will most likely see these latter assertions as weak since they are not usually sought-after benefits of razors. Areni and Lutz point out that while valence manipulations produce interesting effects, manipulations of argumentative logic (i.e., structure, in our framework) are more useful within the pragmatic contexts of marketing communications. As such, Areni and Lutz present possible manipulations of argument strength, while holding claim valence both constant and positive. Their argument versions can be seen in Figures 1 and 2.

It is important to note that while Areni and Lutz primarily define argument quality in terms of the subjective probability receivers attach to a target belief, recognize that this probability is really an outcome of processing a particular argumentative structure. Using the Toulmin model allows us to examine the particular argumentative structures created by Areni and Lutz. The structural analysis of the arguments used to support the "unsurpassed sharpness" brand claim is pictured in Figure 1. The structural analysis of the arguments used to support the brand claims "prevents rusting" and "eliminates nicks and cuts" is pictured in Figure 2.

Diagramming these arguments with the Toulmin model reveals some interesting observations about the nature of strong versus weak arguments, as formulated by Areni and Lutz. First, Areni used Lutz used "authority assertions" to legitimize the relationship between the primary claims and the evidence, and changed the specificity of these assertions as one method of manipulating argument strength. Second, Areni and Lutz introduced additional evidence in support of their primary claims, and changed the valence of this additional evidence as a second way to manipulate argument strength. These will be explained in more detail below.

Manipulating Authority

Turning first to Figure 1, we find that the "advanced honing method"-to-"unsurpassed sharpness" relationship is authorized by several assertions. In the strong version (Panel A), the assertion "...honed more cleanly and smoothly..." serves as a warrant, while the assertion "...recent scientific breakthrough" serves as backing for the warrant. In other words, a description regarding the nature and the effects of the honing method serves to make the "advanced honing method" assertion stand as compelling evidence in support of a sharpness claim.

Turning to the weak version (Panel B), we find that Areni and Lutz changed their authority assertions. Specifically, the assertion "...better than the others" now serves as the warrant, while the assertion "new production process..." now serves as backing for the warrant. In essence, Areni and Lutz manipulated the specificity of their "authority assertions." In the first case, their warrant and backing were detailed, and specifically related to the concept "honing method." In the latter case, the warrant and backing were vague, general and not clearly related with the concept "honing method." Examining the structural diagrams in Figure 2, the reader should note that a similar situation exists with the second set of strong and weak arguments. There is one exception however. The weak argument diagrammed in Figure 2 possesses a warrant without backing.

In general, Areni and Lutz's manipulation of "authority assertions" is consistent with Toulmin's (1958) ideas regarding argument strength. By manipulating warrants within the argument, Areni and Lutz have effectively altered the structural logic (i.e., the logical relationship between the evidence and the claim).

Manipulating Evidence

In addition to manipulating the specificity of authority assertions, Areni and Lutz also manipulated the nature of evidence offered in support of their primary claims. In Figure 1, note that Areni and Lutz introduced a second set of evidence which consists of assertions regarding Edge's sharpness in comparison to other razors, as rated by two prominent athletes. In the strong version (Panel A), the athletes rated Edge superior to Schick, Personna, Bic and Gillete. In the weak version (Panel B), the athletes rated Edge superior only to Personna. Examining the structural diagrams in Figure 2, the reader should again note that a similar situation exists with the second set of strong and weak arguments.

The Areni and Lutz manipulation of evidence is essentially the same type of argument strength manipulation used by Petty et al. (1983). That is, instead of manipulating claim valence, Areni and Lutz manipulate evidence valence. Because Areni and Lutz did not provide warranting assertions for the second evidence-to-primary claim relationship, their manipulation reduces to a "positive" versus a "less positive statement of fact about the brand."

Summary Contention

Based upon our structural analysis of the Areni and Lutz arguments, we offer this summary thought: Perhaps argument quality is a term better used to describe the structural integrity of an argumentative message. With this conceptualization, we can now distinguish between the outcome of argument processing (e.g., strength) from the characteristics of the argument. Defining argument quality in terms of structural integrity allows us to avoid circular confusions such as "strength-as-believable" and "quality-as-valence." That is, we can avoid tautological definitions such as "strong arguments are anything that elicit a positive response." But more importantly, defining argument quality in terms of structural integrity forces us to consider the necessary relationships between argument structure and argument elaboration.

FIGURE 1

STRUCTURAL ANALYSIS OF THE ARENI AND LUTZ (1988) ARGUMENTS FOR THE BRAND CLAIM OF "UNSURPASSED SHARPNESS:" STRONG AND WEAK VERSIONS

FIGURE 2

STRUCTURAL ANALYSIS OF THE ARENI AND LUTZ (1988) ARGUMENTS FOR THE BRAND CLAIMS OF "PREVENTS RUSTING," AND "ELIMINATES NICKS AND CUTS:" STRONG AND WEAK VERSIONS

IMPLICATIONS FOR UNDERSTANDING ARGUMENT ELABORATION

To date, the most influential approach to understanding argument elaboration is the Elaboration Likelihood Model (Petty and Cacioppo 1981). ELM distinguishes between central and peripheral routes to attitude change. The critical feature defining the two routes is that central processing is inferred when attitude change appears to based upon the receivers' diligent consideration of information he/she feels is central to the true merits of the message. By examining argument elaboration within the framework of argument structure, the meaning of central versus peripheral becomes clearer. For example, Bitner and Obermiller (1985) explain that when a message does not provide central cues, or the highly motivated individual does not have experience, he or she may elaborate in a "central" manner cues that would otherwise be peripheral. While there is little research describing the "central" processing of peripheral cues, Bitner and Obermiller may be suggesting that the receiver extracts his or her own "evidence" to support asserted claims (cf. Kardes 1988). As such, understanding the nature ("central" versus "peripheral") of message processing may be a matter of first describing the argumentative structure provided for the receiver, and then, that generated by the receiver.

Given the examples discussed in this paper, it is evident that structural manipulations of argument quality can take many possible forms. And a critical question that should be asked is, "Which component parts of an argument structure manipulation will influence the nature of receivers' elaboration?" It is important that any changes in structure be fully understood in terms of their relationship to elaboration measures of interest. For example, Eagly (1974) criticizes the use of broad recall indexes as measures of persuasion (i.e., elaboration) because they do not reflect what aspects of the message are actually encoded and elaborated upon. Thus if a researcher, for example, uses evidence (or lack thereof) to manipulate argument quality, then measures of evidence recalled may be more indicative of "central" argument elaboration than a more general measure of number of claims recalled (see Munch and Swasy 1988).

It is important to recognize that viewers exposed to argumentative messages must, themselves, construct a representation of the argumentative structure (Brockriede 1975, Hample 1980; 1981). That is, they must identify the evidence-v ia-authority and probability to claim relationships throughout the course of message exposure. Once these logical relationships have been identified, viewers may offer their own qualifications, discount the presented evidence with examples of counter evidence, or challenge the logical relationships between evidence and claims with their own warrants. If we are to better understand persuasion and argument elaboration, we need to begin to see the necessary relationships between the logical structure of the argument and the argumentative structure of-receivers' processing.

REFERENCES

Areni, Charles S. and Richard J. Lutz (1988), "The Role of Argument Quality in the Elaboration Likelihood Model," in Advances in Consumer Research, ed., M. Houston, Association for Consumer Research, 197-203.

Bitner, Mary J. and Carl Obermiller (1985), 'The Elaboration Likelihood Model: Limitations and Extensions in Marketing,'' in Advances in Consumer Research, eds., E. C. Hirschman and M. B. Holbrook, Association for Consumer Research, 420-425.

Brockriede, Wayne (1975), "Where is Argument?," Journal of the American Forensic Association, 11, 179-192.

Burleson, Brant R. (1979), "On the Foundations of Rationality: Toulmin, Habermas, and the a priori of Reason," Journal of the American Forensic Association, 16, 112-127.

Deighton, John (1985), "Rhetorical Strategies in Advertising," in Advances in Consumer Research, eds., E. C. Hirschman and M. B. Holbrook, Association for Consumer Research, 432436.

Eagly, Alice H. (1974), "Comprehensibility of Persuasive Arguments as a Determinant of Opinion Change," Journal of Personality and Social Psychology, 29, 6, 758-773.

Fishbein, Martin and Icek Ajzen (1975), Beliefs, Attitudes, Intentions and Behavior: An Introduction to Theory and Research, Reading, MA: Addison-Wesley.

Habermas, Jurgen (1979), Communication and the Evolution of Society, Boston, MA: Beacon Press.

Hample, Dale (1980), "A Cognitive View of Argument," Journal of the American Forensic Association, 16, 151-158.

Hample, Dale (1981), "The Cognitive Context of Argument," Western Journal of Speech Communication, 45, 148-158.

Hunt, James M., Smith, Michael F. and Jerome B. Kernan (1985), "The Effects of Expectancy Disconfirmation and Argument Strength on Message Processing Level: An Application to Personal Selling," in Advances in Consumer Research, eds., E. C. Hirschman and M. B. Holbrook, Association for Consumer Research, 450-454.

Jaccard, James (1981), "Toward Theories of Persuasion and Belief Change," Journal of Personality and Social Psychology, 40, 2, 260-269.

Kardes, Frank R. (1988), "Spontaneous Inference Processes in Advertising: The Effects of Conclusion Omission and Involvement on Persuasion," Journal of Consumer Research, 15, 2, 225-233.

Kneupper, Charles W. (1979), "Paradigms and Problems: Alternative Constructivist/Interactionist Implications for Argument Theory," Journal of the American Forensic Association, 15, 220-227.

McGuire, William J. (1960), "A Syllogistic Analysis of Cognitive Relationships," in Attitude Organization and Change, eds., M. J. Rosenberg et al., New Haven, CT: Yale Press, 65-111.

Munch, James M. and John L. Swasy (1988), "Rhetorical Question, Summarization Frequency, and Argument Strength Effects on Recall," Journal of Consumer Research, 15, 3, 69-76.

Petty, Richard E. and John T. Cacioppo (1981), Attitudes and Persuasion: Classic and Contemporary Approaches, Dubuque, Iowa: William C. Brown Company.

Petty, Richard E. (1986), "The Elaboration Likelihood Model of Persuasion," in Advances in Experimental Social Psychology, ed., L. Berkowitz, New York, NY: Academic Press, 123205.

Petty, Richard E. and Marlin Heesacker (1981), "Effects of Rhetorical Questions on Persuasion: A Cognitive Response Analysis," Journal of Personality and Social Psychology, 40, 3, 432440.

Petty, Richard E. and David Schumann (1983), "Central and Peripheral Routes to Advertising Effectiveness: The Moderating Role of Involvement," Journal of Consumer Research, 10, 3, 135-146.

Reinard, John C. (1981), "The Role of Toulmin's Categories of Message Development in Persuasive Communication: Two Experimental Studies on Attitude Change," Journal of the American Forensic Association, 20, 206-223.

Toulmin, Stephen (1958), The Uses of Argument, Cambridge: Cambridge University Press.

Toulmin, Stephen, Reike, Richard D. and Allan Janik (1984), Introduction to Reasoning, New York, NY: Macmillan.

Willard, Charles A. (1979), "The Epistemic Functions of Argument: Reasoning and Decision-Making from a Constructivist/Interactionist Point of View," Journal of the American Forensic Association, 15, 211-219.

Willard, Charles A. (1981), 'The Status of the Nondiscursiveness Thesis," Journal of the American Forensic Association. 17. 190-214.

----------------------------------------