Principia Rhetorica: une théorie générale de l'argumentation

Citation

Bade, D. (2009), "Principia Rhetorica: une théorie générale de l'argumentation", Journal of Documentation, Vol. 65 No. 3, pp. 515-521. https://doi.org/10.1108/00220410910952456

One of the most discussed topics of the past decade has been the semantic web and the various techniques that have been proposed and developed to make it a reality. Given the ubiquitous presence of rhetorical perspectives in virtually every field of academic research during the past‐half century, the complete absence of a rhetorical perspective in the discussions of the semantic web is surprising. While a search of the full text EBSCO Library, Information Science & Technology Abstracts with Full Text database using the words semantic, web and rhetoric produced zero citations when I began writing this review, the literature of information science and librarianship is teeming with discussions of issues that are clearly rhetorical. Examples are everywhere: discussions of the order of elements in a bibliographic display, relevance ranking, the language of indexing and description (natural language and controlled vocabulary), Dervin's “communication not information,” orientation towards the user, and the reference interview all involve issues of rhetorical practice. Most of these topics have never been discussed from a rhetorical perspective, but several publications of the last decade have used rhetorical theories in various areas of information technology (IT) research. Sapienza's (2007) paper on single‐sourcing, Potter (2008) on asynchronous learning networks, Argamon et al. (2007) on text classification, and Spinuzzi's (2005) paper on research techniques are excellent examples of how rhetorical theory can illuminate the design, implementation, use and evaluation of ITs.

Believing that “current rhetorical theories are insufficient” for analyzing the technique called single‐sourcing, Sapienza (2007) drew on early twentieth century poetics (imagism and acmeism) for his analysis. Like Argamon et al., Spinuzzi and Potter, he made no references to Michel Meyer, nor have I encountered any references to any of Meyer's works in the LIS literature. The present review is not a review of Principia Rhetorica as a book on rhetoric, but an introduction to the particular approach to rhetoric Meyer takes – one based on what he calls problematology and which as a search term also does not match anything in the EBSCO database – and its relevance to LIS.

Principia Rhetorica begins with a chapter on the definitions of rhetoric followed by a brief history of rhetoric leading to Meyer's “new vision of rhetoric” based on problematology. Meyer notes that the periods of greatest attention to rhetoric were periods of epistemological uncertainty: classical antiquity, the renaisance and the twentieth century. In earlier works as in the work under review here, Meyer compared the triplets of rhetoric (ethos, logos, and pathos) with the triplets of Saussure's speech circuit (speaker, language, and hearer), of information theory (sender, message, and receiver) and philosophy (self, world, and other). Here, he notes that ancient rhetoric emphasized ethos (the speaker), renaisance rhetorics emphasized pathos (the hearer/receiver/other), and twentieth century rhetorics emphasized logos (the world). A product of the renaissance, science emphasizes logos: language as hardware in the brain, messages in a technical system, the world as object of knowledge without reference to a knower. In contrast, Meyer's approach treats ethos, logos, and pathos as equal in importance, related through their participation in the problematic that gives rise to the communicative act.

Two remarks by Meyer serve nicely as demonstrations of his approach and its significance for current research in information systems. Contrasting rhetoric of Aristotle with that of our time, Meyer (1994, p. 9) remarks:

What one discusses are theses, (Aristotle), propositions, in the literal sense of the word to propose. However, that conception is out of date. What one debates, those matters upon which we reflect, are questions, and if there is no problem, there is no discussion (the full significance of this remark for philosophy may be found in his 1994 monograph Of Problematology).

His remarks on Descartes' four rules in the Discours de la méthode could easily have been written about the semantic web:

If one examines the source of these four rules, one immediately sees that they are concerned with transforming each stage of rhetoric into a precise rule of method. Descartes thus transposes inventio, dispositio, elocutio and memoria into strictly analytical steps, leading to the result that the result is a necessary result rather than simply a reasonable response […] The dissolution of the dialectic is total. The first rule is very clear in this regard: one must not permit anything that is problematic, since, according to Descartes, the problematic is the doubtful. The alternative is identified with the undecidable. The problematic must therefore be totally eliminated. One must find (inventio) the right response, that is to say the response that makes the problem dissappear. From there on it is the task of method rather than rhetoric to produce results that are no longer uncertain. Thus, one has judgements that no longer have any thing to do with interrogativity, these being filed away among all things dubious, because we find ourselves in an intellectual universe ruled by a founding self‐affirmation of the necessary and exclusive affirmation: A without the possibility of having non‐A (p. 45).

This transformation of the problematic, of possibilities, alternatives, disagreement, meaning, metaphor, polysemy, ambiguity, ignorance, the unknown, uncertainty, vagueness, deception and propaganda, into the “machine‐interpretable” triplets of resource description framework (RDF) is simply the latest version of the cartesian project to eliminate uncertainty. This transformation requires, as Meyer (1988, 1994) has argued at length in earlier works, obscuring or denying the problem which gives rise to the question that provokes the statement, proposition, response in the first place. The semantic web is predicated upon the notion that meaning can be made explicit, i.e. meaning is not problematic, all questions of interpretation will be eliminated. In the semantic web, there will be definitions and rules. The RDF and the functional requirements for bibliographic records alike assume a world that can be described entirely in propositions. The semantic web is the “philosophical grammar” of Descartes' contemporary John Wilkins in a new guise. And indeed Meyer's comment:

If one demands “What is X?” this evidently presupposes that X is some thing, a rather than b for example, since one must know what it is we are asking about and not confuse it with some other thing. So to say that X is a, or that X is b, is already to specify what it is that one is asking for (p. 26), […]

Echoes Harris' (2005, pp. 61‐2) criticism of Wilkins' universal language:

[…] a classification system that purports to reflect the “real” organization of the world of Nature cannot be proposed in advance of the experimental investigations necessary to determine that organization.

Moving from the renaissance into the twentieth century, Meyer's comments on Gadamer's hermeneutical approach also seem pertinent to the basic problems for understanding information seeking behaviour. When Meyer quotes Gadamer's remark “one cannot understand a text unless one has understood the question for which the text is a response” (p. 63). I am reminded of Collingwood's critique of propositional logic (specifically mentioned by Meyer several times in publications of the past 30 years) and his famous thesis about understanding historical events: in order to understand human action it is necessary to understand what the actor intended to do, what problem the actor was trying to solve. Gadamer himself argued that even though Collingwood was oriented in this direction, we still have no logic of question and answer. If the meaning of a statement relates to the question it answers, so must its truth. And as many of Collingwood's critics noted, discovering that question is a problem, a problem that will not be resolved in a triplet. It is also the problem in the design of any information system, i.e. matching a database to a query that is known to the search system only through the text entered into the system. It is the problem for which techniques as disparate as relevance ranking, keyword searching and controlled vocabulary have been proposed as solutions, but it is a problem, like historical understanding, which remains in the realm of the problematic.

Meyer then devotes two chapters to the distinguishing features of rhetoric and argumentation, followed by the chapters that make the book more than just another book on rhetoric. In the first of those chapters – Ethos, logos, pathos dans l'interaction rhétorique – Meyer offers up a discussion of the classical rhetorical concepts of ethos, logos and pathos that raises a multitude of questions for the parallel sender‐message‐receiver paradigm of information theory:

Is there anything more convincing for anyone than one's sense of humanity and capacity to respond to questions that others pose and to which they seek answers? […] The ethos is presented as the most profound expression of a common humanity, thanks to which we all alike respond to the great questions […] The ethos becomes more distant and “authoritative” as it places itself within the ad rem, and more malleable and adaptable to others, more compatible and sympathetic when it is rhetorical, distance there being by definition less conflictual. To create distance one must transform the ethos into a principle of authority that excludes all putting in question (p. 152).

The ethos, in becoming a principal of authority, goes from universal wisdom to particular knowledge, from humanism to technical competence (p. 153).

Reading this, I wondered how Meyer's “new” rhetoric would be pertinent to understanding human‐ or machine‐machine communication. Was Meyer's “new” rhetoric out of date before it was published? I think not, but questioning information science and the discourse of the semantic web in terms of ethos (and logos and pathos) will certainly raise some of those “great questions” that humans ask and to which they respond. What is ethos/authority in machine generated discourse? How is the query entered into an information system related to the inquirer's question? How is the response of the information system to that query related to the inquirer's question? To a human response?

Rhetoric, for Meyer, is the art of speaking and writing about the problematic. In an information system, there are no questions, only answers, for nothing is problematic: every response is an answer that can be accepted or rejected but never argued or debated (the problematic as the debatable being for Meyer the primary characteristic of human knowledge). This would seem to suggest that rhetoric plays no part in an information system, that logos (message) alone – to the exclusion of both ethos and pathos – matters in an information system. Yet, if we recognize that the computer is a primary auctoritas for our society, ethos comes back into the center of the discussion. Furthermore, the source of the data within the system and the designers of the software at work are both largely hidden elements that ought to be examined from the perspective of ethos as well. Do ontologies operating “under the hood” act as hidden ethos, making them unquestionable authorities? Will the interoperation of ontologies in the proposed semantic web distort the logos/message of the original author? Will this entail the elimination of pathos/hearer/receiver?

The next three chapters are devoted to discussions of ethos (Logique des valeurs, logique de la culture (Logic of values, logic of culture)), pathos (“Comment négocie‐t‐on la distance entre individus?” (“How is the distance between individuals negotiated?”)), and logos (La théorie des variations problématiques (Theory of problematic variation)), a chapter on social aspects of argumentation and concludes with the chapter “Métarhétorique.” The chapter on ethos concludes with the statement that “social life is founded upon the distance that must be continually negotiated at each encounter,” that distance being one of cultural and personal values which lead to different questions. The following chapter on pathos discusses the problem of interlocutors reaching understanding of each other's positions, if not agreement on them. Meyer's criticism of his teacher Chaïm Perelman's universal auditor appeared earlier in the book and in this chapter he contrasts the auditor/reader as imagined by the speaker/writer – pathos projectif, the projective auditor – with the actual or effective auditor/reader (pathos effectif). The assumed auditor/reader (in LIS terms, the user) of techniques such as natural language description, keyword searching and relevance ranking is none other than Perelman's universal auditor, a fiction, or, as Meyer puts it, the projective auditor, the reader as we (authors and programmers) imagine her to be, who may or may not have much to do with the actual auditor/reader/user:

In the ethos‐pathos‐logos relation the pathos projectif leads to a triple response: the understanding of what the speaker wants to say (“what is the question?” is the question to him), the adequacy of the response and the persuasive interest of the proposition. If I address myself to someone, I make that triple hypothesis: the difference will be eradicated by persuasion, my response will be judged adequate and the other will necessarily understand my point of view as well as my intentions. But the auditor is not forced to react as I imagine or hope he will. The pathos effectif, that is to say the real auditor, is moved by other parameters (p. 230).

Meyer's discussion of negotiating distance is largely concerned with a direct communication situation, whether in a court of law, mass advertising, reading or face‐to‐face speaking. Librarianship and information science are largely concerned with mediated communication, whether that mediation takes the form of document‐librarian‐user, document‐cataloguer‐IT‐reference librarian‐user or simply document‐IT‐user. I was led to see that the success of user supplied descriptions (tags, reviews, etc.) attests to the value of addressing a varied readership, an interesting technical response to a problem ignored by standardized thesauri as well as keyword searching. Most importantly, Meyer's problematological rhetoric provides a rigorous theoretical understanding of both the failures and the successes of a number of technical means for dealing with the problems of matching database and reader through a textual query, i.e. negotiating distance using a technical system, for the conflicts between the cataloger's assumption about the user and the actual user and the conflicts between system (designer's) assumptions about the user and the actual user can both be understood in terms of projective versus effective users.

Meyer's theory also provides insight into propaganda and spam:

A successful agent must offer responses that meets the values of the auditor, putting at a distance those that she rejects, exalting her emotions, even passions, and sincerely at that (p. 234).

However, he continues, the rhetoric most successful in that effort is that of the seducer, the manipulator. In the technical context of IT, we call that spam – when we recognize it. While programmers working on anti‐spam software will certainly agree with Meyer's remark that “in everyday life those failures and the subsequent adjustments are natural and inevitable” (p. 235), his statement is equally true for all communication.

The chapter on logos discusses the problematic of world as word: literature, law and political discourse, with a closing section on the language of advertising. As a key to literary style, he offers a “law of inverse problematicity” in which he states that “the more a problem is literally expressed in the text, the more the text will explicity provide the resolution” (p. 246). The corollary according to this law is that the more the problem is obscured, figuratively expressed, the less the text will attempt to resolve that problem. Applied to political and legal discourse, the law of inverse problematicity is restated. In legal discourse:

The more there is conflict, the more the discourse is affected by pathos. The more the discourse diminishes the problematicity of the questions, the more it expresses the resolution of the norm in terms of positive law (pp. 262‐3), […]

While in political discourse:

The more the question is expressly specified, the more the political discourse is literally conditional. The less overt the problem, the more it is taken over by ideological discourse that pretends to resolve everything (p. 263).

One of my few complaints about this book arises from having read his earlier books: Meyer's examples in Principia Rhetorica are the same examples he has proferred in publications of 1988, 1992, 1994 and elsewhere. By now, I am a bit weary of discussing the possible questions that could give rise to variations on the statement “Napoleon was the victor at Austerlitz.” A more diverse set of examples could have made his points not only clearer, but far more provocative. For instance, whether discussing ethos (sender), logos (message), or pathos (receiver) in information retrieval, the problem is generally approached as a problem of information and users. Yet, when confronted with the following responses to a query using the term “peace”, we can see how important it is to take account of ethos/authority in the evaluation of those results:

The State must see that public peace and order are preserved and, in their turn, order and peace must make the existence of the State possible (Hitler).

[…] the first premise for every truly human culture […] the individual's capacity to make sacrifices for the community (Hitler).

[…] socialism is sacrificing the individual to the whole (Goebbels).

But there is a direction to events, and the sacrifices of the present have not been in vain […] We are called still to spread liberty, to assure justice, to be the makers of peace (McCain).

Americans laid down their lives for this country and for the peace they were there to protect. We revere their service. We honor their sacrifice (Obama).

True partnership and true progress requires constant work and sustained sacrifice (Obama).

Socialism and work are inseparable from each other (Stalin).

Only under the reign of socialism can peace be fully established (Stalin).

Don't imagine that I came to bring peace on earth. I did not come to bring peace, but a sword! (Jesus).

My point here is not to smear Jesus, Obama and McCain by comparing their words to those of Goebbels, Hitler and Stalin, but to demonstrate that how we understand the meaning of peace (socialism, sacrifice, work […]) is dependent upon who we know as the speaker/writer, i.e. it is a rhetorical matter of ethos. ITs do take ethos into account whenever they rank a page by how “popular” it is, but in spite of a few such techniques, consideration of ethos is largely absent from IT system design. And how to incorporate matters of ethos/authority into IT is as much a political matter as it is technical, which is the topic of Meyer's penultimate chapter.

I began my own reading of Meyer's works in my efforts to understand what kind of a rhetorical strategies would be appropriate or necessary for creating machine operable metadata, and conversely, what strategies would be appropriate for library users facing a computer screen. Meyer addresses neither of these questions directly, but in his works and in particular this latest monograph Meyer has raised a host of questions that need to be addressed before tackling the particular questions of human‐ and machine‐machine interaction. Although there is a growing body of literature advancing a rhetorical approach to hypertext, internet writing and reading, human‐computer communication and related matters, the literature that I have examined is all based on older rhetorics emphasizing only one of the elements of the ethos‐logos‐pathos trilogy, and furthermore generally suffer from being focused on technology rather than the communication situation as a whole. Meyer's problematological rhetoric permits an understanding that is rooted deeply in human experience – difference, disagreement, ignorance, and alternative possibilities – focusing on the problematic and the ways in which we respond to questions, aspects of human communication that will not dissappear or become outdated with the next technical development. It is not misleading to characterize problematology as an anti‐ontology. Ontology attempts to state what is, while problematology enquires. The question is always prior and persistent; the answers we give (natural language, science, and ontologies) are debatable and ephemeral: like hardware and software, they are here today and gone tomorrow.

Agre (2002, p. 131) has described how practitioners of artificial intelligence (AI):

[…] believed that by defining their vocabulary in rigorous mathematical terms, they could leave behind the network of assumptions and associations that might have attached to their words through the sedimentation of intellectual history.

AI:

[…] inherited certain discourses from that history [of Western thought] about matters such as mind and world, and it has inscribed those discourses in computing machinery. The whole point of this kind of technical model‐building is conceptual clarification and empirical evaluation, and yet AI has failed either to clarify or to evaluate the concepts it has inherited (Agre, 2002, p. 141).

What happened instead, he argues, is that:

[…] AI has been left with no effective way of recognizing the systemic difficulties that have arisen as the unfinished business of history has irrupted in the middle of an intendedly ahistorical technical practice (Agre, 2002, p. 131).

Agre (2002) argued for “a critical technical practice: a technical practice within which such reflection on language and history, ideas and institutions, is part and parcel of technical work itself”. It is in that spirit and towards that end that we should read Principia Rhetorica. It is not just the meaning of words in a text that matters in information retrieval; authors and readers, information suppliers and information users are equally important agents. Principia Rhetorica does an excellent job of describing and theorizing these issues in terms of a problematological rhetoric.