Mark Vorobej (2006): A Theory of Argument

1 Summary of the book

This book is written for upper-level undergraduate students who have completed at least one course in logic, critical thinking or argumentation. Although the title suggests that the book provides a comprehensive theory, Vorobej deals primarily with the notion of argument, with the cogency of arguments and with how to develop a charitable reading of an argument and display it in a diagram. The book is not about argument schemes, argumentation indicators, dialogue, rhetoric or logical form. Nor is the book about argument evaluation. Norms are being discussed, but from the perspective of reconstructing arguments from a text. Part one of the book is called macrostructure and deals with arguments in canonical form (where they have a conclusion and a set of premises), with the cogency of arguments and with the analysis of so-called normal arguments. Part two is about the microstructure of arguments, i.e. with the more detailed patterns of evidential support. The book contains four hundred exercises with which students can examine the notions and definitions that the book introduces. Still, the book is not merely a textbook, but can also be considered as a scholarly contribution to the study of argumentation.

The first part, on the macrostructure of arguments, has three chapters. The first chapter is about the notion of an argument. Vorobej makes it clear that he conceives of an argument as an attempt by an author to convince an audience to do or believe something by an appeal to reasons or evidence. The audience he refers to is the intentional audience, i.e. the persons that the author himself has in mind in his attempt at rational persuasion. The aim of constructing the macrostructure of an argument is “to provide a perspicuous representation … of that argument’s macrostructure as it is conceived by its author”, giving in that way “primacy to persons over texts” (p. 18). To be clear, the task of argumentation analysis is not so much to determine what the author is in fact doing when offering an argument, in a more objective sense of the term, but to find out how the author “herself conceives of this particular exercise in rational persuasion” (p. 18). Learning to listen to authors is taken to be important for enabling the cooperative kind of relationships that are needed for a culture in which argumentation is to flourish. The chapter concludes with a discussion on charitable interpretations. When there are several readings that are equally well supported by the contextual and linguistic evidence, that reading should be favored that suits the author’s attempt at rational persuasion best.

The second chapter deals with cogency. In order for the author of an argument to achieve rational persuasion, his argument must be cogent. An argument A is said to be cogent, in a generic sense, for a person P in a context C “just in case it is rational for P, within C, to be persuaded to believe the conclusion of A, on the basis of the evidence cited within A’s premises” (p. 47). Cogency is further explicated with four conditions that yield Vorobej’s specific notion of cogency: It is rational for someone to be persuaded of the argument’s conclusion by way of its premises just in case: (1) it is rational for him to believe that each proposition in the premise set is true; (2) to believe that the premise set is relevant to the conclusion; (3) to believe that the premise set grounds the conclusion and (4) when it is rational to believe that the argument is compact. That a premise set is relevant is taken to mean that were the premises true, the set as a whole would provide evidence in favor of the conclusion. That a premise set grounds the conclusion means that were the premises all true, this would provide sufficient evidence for the conclusion. That a premise set is compact means that every proper subset of it would provide less evidence in support of the conclusion. What constitutes rational belief is based upon Richard Foley’s notion of reflective stability: “it’s rational for a person P to believe a proposition Q just in case, were P to reflect carefully upon Q in the context of her current overall epistemic state, she would eventually… form a settled conviction that Q is true” (pp. 61–62). As a proposition P can be rational for you to believe, while in fact you have not (yet) reviewed P critically from the perspective of your epistemic state, an argument can be cogent for you without the four cogency claims having been reviewed. Epistemic states are not restricted to beliefs but can also contain desires, fears, etc. When these emotions or wishes are not consciously accessible for a person, she can arrive at a reflective stability with respect to a proposition Q while, from a external perspective, her belief in Q is only explainable as the result of such an emotion or desire. Also in such circumstances, Vorobej holds it to be rational for this person to believe Q.

The third chapter starts with introducing the notion of a normal argument. An argument is normal when the author holds his argument to be cogent for himself as well as for the members of his audience. Vorobej then presents the normality assumption that is to be invoked when interpreting an argument: in the absence of evidence to the contrary an argument is normal. Given this assumption, it becomes easier for an analyst to find out what the argument of an author amounts to. We can make use of the requirements imposed on a normal argument, such as that it is rational for audience and author to believe that the premise set grounds the conclusion. So, we can use what we know about what is rational for these participants to believe when examining what argument is most likely to be expressed within some discourse. After a detailed exploration of the connections between cogency of the one hand and validity and reliability (a probabilistic version of non-deductive argument strength) on the other, this chapter closes with methodological matters. In this section, Vorobej points out that the objective of an argument analysis, i.e. to find out how an author believes he is convincing his audience rationally, can be understood in two ways. Either we are trying to understand how the author tries to rationally persuade his audience, given what we, as argumentation analysts, understand by “to rationally persuade”, or we are trying to understand how the author tries to rationally persuade his audience, given what he himself understands by “to rationally persuade”. Even though an author may employ a quite distinct notion of rational persuasion than Vorobej, Vorobej chooses the second option as appropriate in an examination of what it is that an author believes he is doing when presenting his argument. But, given that many authors do not make use of any sophisticated notion of cogency, Vorobej recommends, in the absence of contrary indications, to start from the thick cogency assumption: the assumption that the author uses the four cogency conditions as discussed above.

Part two of the book is about the microstructure of arguments, and more in particular with how we can represent, in diagrams, the way the author tries to rationally convince his audience. “[A] diagram that depicts the structure of an argument from the perspective of its normal author simultaneously captures how that author conceives of her own argument and how she believes others—specifically, those targeted by her argument—ought to view it as well” (p. 163). According to Vorobej, argument diagrams “are centrally concerned with displaying relevance relations” (p. 166). In addition to developing a method of argument diagramming, he also offers guidance for extracting a charitable reading from the text.

In chapter four, an argument is defined as convergent when every premise is independently relevant to the conclusion. That a premise or set of premises is relevant to a conclusion is taken to mean that the probability of the conclusion given the premises, is greater than the unconditional probability of the conclusion. A premise in an argument A is independently relevant to A’s conclusion if it is relevant to the conclusion “independently of any other proposition within A’s premise set (but not necessarily independently of all propositions whatsoever)” (p. 172). An author can put forward several converging premises in order to support his conclusion to a sufficient degree. Vorobej proposes to use squiggly elements in a diagram to indicate the analyst’s (called the artist by Vorobej) disagreement with some claim by the author. If he holds a premise false, he can convey this by making the diagram’s circular node representing this premise squiggly. Similarly, lines (representing a relevance relation) and arrows (representing a grounding relation) can be squiggled. When the analyst has several interpretational options and needs to be charitable he can investigate the drawbacks of each option by taking a look at the squiggles and use this information to choose the most charitable reading.

The fifth chapter deals with linkage. An argument is linked in case its premise set has at least two premises, is relevant to the conclusion and contains no proper subset that is relevant to the conclusion. The relevance as well as the grounding relation of an argument (for example a linked argument) can be vulnerable in the sense that when one premise is eliminated from the argument the argument comes to fail to be relevant for the conclusion, casu quo, to ground the conclusion. An argument is hypervulnerable when this holds for each premise of the argument. The policy of analyzing normal arguments in a charitable way is to prefer an invulnerable argument over a vulnerable and a vulnerable argument over an hypervulnerable.

The last chapter is about hybrid arguments, arguments that contain a supplementation relation within the premise set. A premise P supplements another premise Q when Q is independently relevant to the conclusion, while P is not, and when P and Q together provide stronger support for the conclusion than Q alone. This definition is introduced in order to capture cases where information is provided (P) that strengthens the support given by other premises (Q) to the conclusion. So if the conclusion “All the ducks on the pond are yellow” is independently supported by “Daphne has seen 100 yellow ducks on the pond” (which implies that the argument cannot be linked), then the premise “There are exactly 100 ducks on the pond” is not independently relevant to the conclusion but it strengthens the support the other premise gives to the conclusion. The supplementary premise is placed in the diagram, with a special notational device, on the same level as the premises that support the conclusion independently. This also holds for every supplementary premise that supports the connection between a supplementary premise and the conclusion. The book ends with an application of Vorobej’s theory to arguing from ignorance, hinting in that way at how his concepts and methods can be of use for fallacy theory and argument evaluation.

2 Evaluation

Vorobej carefully defines the notions that he needs for his method of argument analysis. The book is packed with subtle discussions about how to understand these various concepts. The definitions and interpretation policies can be applied to interesting argumentative discourses that are found in the exercises. In addition to being a suitable textbook for analytically minded students, the book is a valuable scholarly source when dealing with cogency and argument diagramming. To these general remarks, I want to add ten more detailed comments.

• 1. The students are taught to extract from an argumentative passages what the author himself conceives of as his attempt at rational persuasion. Vorobej even suggests at some point that it might be helpful to interview the author about his argumentative intentions. He forcefully incites us to listen to what other have to say. I am very sympathetic to and impressed by his examples of charitable analysis. However, at least as important as paying attention to the author’s beliefs and intentions is reading carefully what they in fact have expressed. So, I missed an account of the pragmatic and linguistic background that is needed for determining what utterances within a particular context actually convey.

• 2. Vorobej’s theory is epistemic in character. The kind of moderate epistemic relativism that makes cogency dependent upon context and the epistemic states of the persons involved is closely related to kind of dialectical relativism that makes the defensibility of a standpoint dependent upon the actual commitments of the interlocutor. However, by focusing on beliefs Vorobej needlessly narrows down the scope of his theory. Genuine arguments are also put forward outside of epistemic contexts, such as political debates, judicial examinations or negotiations.

• 3. Cogency has been defined with four clauses. However, the requirement of (positive, probative) relevance is made superfluous, both by the clause of groundedness as well as by the clause of compactness. To argue for the latter: if the argument is compact, then every proper subset of the set of premises would provide less evidence in support of the conclusion than the complete set of premises. Because the existence of a proper subset that provides less support than the superset implies that the superset at least provides some support, the set of premises cannot be completely irrelevant. So, compactness implies relevance.

• 4a. Even though argument is defined as an attempt at rational persuasion, and in that way as a dialogical event, the dialogical nature of persuasion is scarcely taken into account. There is no discussion of the kind of dialogical procedure that makes argumentation dialectically reasonable. Instead, Vorobej prefers the monological criterion of reflective stability in order to explicate the notion of rational belief.

• 4b. Because a proposition Q can be rational for you to believe, while in fact you have not (yet) reviewed Q critically from the perspective of your epistemic state, an argument can be cogent for you without you having reviewed whether the four cogency claims hold. Therefore, in order to realize (rational) persuasion, it does not suffice that the argument is cogent for the respondent. In addition, the respondent ought actually (in Vorobej’s terms) to review the argument critically until she arrives at a reflective stability with respect to the four cogency claims or at a point where she knows that such a stability is out of reach. She can only be said to be persuaded rationally if in fact her critical review leads her to accept the claims of the argument. Again, there is an irreducibly dialogical aspect in rational persuasion (cf. p. 68).

• 5. The compactness requirement on cogency plays a role that seems to be quite different from the other three requirements. Suppose an argument has all the virtues but being non-compact by having a superfluous premise P. Then, Vorobej holds, you might become persuaded of the argument’s conclusion, but not exactly by the all the evidence the author presents to you. It must be agreed that the superfluous premise P invites a critical response. Still, the author’s flaw is communicative in nature, rather than epistemic. This, again, pleads for an examination of the communicative requirements on arguments.

• 6. Given the importance of ex concessis arguments, the normality assumption seems to be too strong. When presenting an ex concessis argument, the author does not need to believe his premises or conclusion to be acceptable.

• 7. The book makes it insufficiently clear why the distinctions between the various types of diagrams matter. The main function of diagramming ought to be to facilitate argument evaluation. From such a perspective, it appears to be useful to put more emphasis on the distinction between, what is called in pragma-dialectics, multiple and non-multiple argumentation, i.e. between one attempt at persuasion and several attempts (van Eemeren and Grootendorst 1992). Instead, the discussions on argument diagramming mainly deal with relevance issues.

• 8. Vorobej rightly pays a lot of attention to premises that justify a support relation. But the way he proposes to diagram these so-called hybrid arguments is not convenient. Because a support relation is the kind of entity that can be supported in an argument, I suppose it can best be represented within the diagram with a separate node, x. The support for this relation can then be represented with a node y, put above this node x, and connected with an arrow from y down to x. If the proposition in x is left implicit by the author, it can be represented as an implicit element.

• 9. Like most textbooks that deal with diagramming (and like the software available for argument diagramming), Vorobej does not include a device for diagramming hypothetical reasoning, such as proposed by for instance Thomas (1981) and Fisher (2004). That is a pity, because a complex conditional proof or reductio ad absurdum can only be diagrammed without such a device in a cumbersome way.

• 10. Unfortunately, the book hardly contains any reference and neither does it have a bibliography.

Notwithstanding these critical comments, the book is a valuable contribution to the theory on reconstructing argumentation as well as a textbook that stands out with precision and subtlety.