How we should teach reasoning by sign, but don't.

• Author: Holmquest, Anne

This essay challenges the conventional interpretation of reasoning by sign as it usually appears in argumentation textbooks and is taught to most students. Devoted to Aristotle's Rhetoric, students of argument have neglected another ancient and rich explication of reasoning by sign in his Prior Analytics. The Prior Analytics derives reasoning by sign, the logic of invention, from a broader theory of logics in general. Aristotle's Prior Analytics can help us understand a neglected theory of sign reasoning focused on invention rather than justification, and can explain how a theory of signs underlies a theory of logic. What is advantageous about this approach? The legacy of Isocrates teaches us not to prefer those things which are merely theoretical, or merely practical, but to prefer theories that are practical. Aristotle's Prior Analytics is the only text on the practice of sign reasoning which is founded on an underlying theory of signs. Even recent theories of signs (e.g., Marcus, 2005; Queiroz & Merrell, 2005; Rosenthal, 2005) do not do this. What else is so advantageous? From this approach it will follow that the essence of sign reasoning is rhetoric, not philosophy. No other theory of this form of inference (e.g., Josephson & Josephson, 1994; Lipton, 2004; Walton, 2004a) can say this.

Charles Sanders Peirce is perhaps the only person to develop a theory and practice of sign reasoning out of the Prior Analytics (Holmquest, 1986). The Prior Analytics influenced Peirce to give the name abduction to the process of discovering a hypothesis to be tested, which differs from the process of justifying a hypothesis already discovered. The idea that abduction is reasoning by sign is so uncommon that it requires some preliminary explanation. Peirce believes that the purpose of reasoning is to discover, from the thought of what we already know, something else that we do not know. To Aristotle and Peirce, forms of reasoning differ as the part that represents an unknown in each form differs. Because nothing is known without middle terms, all arguments hinge on them. A middle term is an index. An index is a kind of sign. Thus, noting the general function of the middle as a sign can explain necessary and ordinary logics: deduction and induction (the logics of justification), and abduction (the logic of invention). Because Peirce inferred the concept of abduction from Aristotle, the centrality of the middle term as a sign leads to a broadened understanding that, in a sense, all reasoning is by sign. Thus, all logic is rhetorical. Peirce says that we cannot pay attention logically unless something is known through something else (Hartshorne & Weiss, 1932, pp. 387-388, 447-448). To Peirce a sign is anything that someone can take as significantly substituting for something else, and all reasoning is by sign, according to the transitive property of the relation of inclusion.

Whether we revere or criticize him, too often we neglect to read Aristotle carefully. The Prior Analytics should be read through Peirce's insight, and read for a theory and practice of sign reasoning that is the foundation of a theory of logic. Peirce resurrects and revises Aristotle's implicit concept of abduction by showing the relationships among major, minor, and middle terms in argument. Then, noting the middle term's importance in reasoning, he reinterprets Aristotle's conception of the middle as a sign in general and, thereby, replaces Aristotle's theory of predication with a broadened theory of signs, including reasoning by sign. According to Peirce, a sign can stand for something else only because this relation is mediated by another sign that translates and explains the first one; he calls this other sign an interpretant (Burks, 1958b, pp. 136-137). Finally, Peirce says that one sign is explained by another in the same way that one thought is explained by another (Hartshorne & Weiss, 1932, pp. 135-137, 1934, p. 416).

In taking this stance, I agree with Zompetti (2006) and Braet (2004) that rediscovery of Aristotle's texts prompts us to adjust our view of argument schemes. Zompetti (2006) reminds us that Aristotle's topoi were of two types-general and specific. Toulmin (1958) called these field-invariant and field-dependent arguments (pp. 14-15). Scholars often understand the field-invariant part of Toulmin's model as "a syllogism on its side" (J. L. Ericson, personal communication, November 19, 1975), representing enthymemes in rhetorical syllogisms. But topoi also enable specific arguments to be made enthymematically: These particular enthymemes are derived from specialized content of relevance to specific audiences. Zompetti (2006) quotes a famous passage from the Rhetoric "Topoi, according to Aristotle, are the places, or 'special regions the orator hunts for arguments as a hunter pursues game. Knowing where a kind of game (or argument) is to be found, he [sic] will hunt for it there'" (p. 20). Is this so different from understanding the cognitive schemes used by audience members? Both Zompetti and I believe that we can reconstruct the logic of ordinary language users using Aristotelian frames of reference. The notion of argument schemes implies that auditors test any given argument against evaluative criteria appropriate to that type of argument. Lines of argument may be evaluated similarly. Hastings (1963) distinguished restricted and unrestricted schemes; the latter include authority, analogy, and example. By its conventional interpretation, sign reasoning could be construed as an unrestricted argument scheme. As I will show, however, once Aristotle and Peirce are understood on the matter, we shall see that reasoning by sign is the scheme of schemes.

An objection might be raised against my use of the term abduction. I read Peirce very differently from Walton (2004a). As I will detail below, Peirce emphasized the relationship of the rule, the case, and the result. Walton's view of abduction, however, emphasizes dialogic arguments (2004a) and modusponens (2004b). Grossman (2006) has commented: "I cannot recommend [Walton's Abductive Reasoning] to generalist philosophers

looking for an efficient introduction to abductive inference. It is, however, a very worthwhile read for specialists in inference to the best explanation." So perhaps there is room for a different introduction. Contemporary accounts tend to equate abduction with inference to the best explanation (e.g., Josephson & Josephson, 1994; Lipton, 2004). In such accounts, rhetoric has little if any role. As Lyne (2005) has pointed out, rhetoric is not central in Walton's view. Interpreting abduction as sign reasoning, on the authority of Aristotle and Peirce, however, renders reasoning by sign broader than we heretofore have imagined: all reasoning is by sign, and it is inherently rhetorical.

In A Theory of Semiotics, Eco (1979) repeated a view widely held among philosophers, that ordinary language has no exact logic. This view places the theory of signs in an interesting dilemma:

[P. F. Strawson] says: "Neither Aristotelian nor Russellian rules give the exact logic for any expression of ordinary language, for ordinary language has not exact logic." The purpose of a theory of codes was to see if ordinary languages have, if not an exact one, at least one logic. Maybe the problem is not of finding a logic, if logic is only the theory of a formalized language. The problem is to find a semiotic theory which is surely different from a formal logic, but which is nevertheless able to dissolve the shade of skepticism suggested by Strawson's quotation, which may lead to the suspicion that natural languages cannot have a theory, which has to be refused if a semiotics is to subsist. (p. 300)

At least since the time of Ramus, Descartes, and Bacon, it has been a commonplace to criticize Aristotle. Eco's view that natural language has no exact logic but involves rhetoric makes his own theory of codes and sign production seem to fill a void in Aristotle. But, like Strawson, Eco too is speaking a commonplace. Of course Aristotle did not have an exact logic. What he had was an inexact logic, applicable to ordinary language and not different from a theory of signs. Reasoning by sign, in Peirce's sense, is the god of all cognitive schemes.

My argument is that Aristotle's Prior Analytics solves the problem that confounds Eco (1979), Walton (2004a), and others: Can ordinary language possess rationality that does not collapse into individual standards? Consider the example of Geico Insurance advertising. The makers of the its "caveman" commercials did not test them in focus groups. Yet, their success has spawned a spin-off television show and encouraged other advertisers to abandon focus groups as well. Advertisers are finding that ordinary people misrepresent their opinions, instead advancing arguments that they think will sound more impressive to the group, with the result that effective advertisements may be abandoned prematurely (Porter, 2007). Is it not rational to use a focus group in this field even though it was rational before? Did the producers possess rationality that the focus group did not? Or, did the results just collapse into a new individual standard?

Through Aristotle, Peirce develops a theory of signs that accounts for our ability to: (a) produce signs from existing codes, (b) generate signs in contexts where underlying codes do not yet exist, and (c) continually critique the production of signs, presently and in the future. These parallel deduction, abduction, and induction, respectively. Of course, the neglected process that demands recognition is (b) abduction. To generate signs in contexts in which underlying codes do not yet exist does not fall under the purview of logics of justification. It requires a logic of discovery.

To elaborate this argument, I first will survey approaches to reasoning by sign, showing that nearly all focus on the production of arguments from...