Rules for Argumentation in Dialogues

Abstract

In this article it is pointed out what kind of rules for communication and argumentation are required in order to make it possible to resolve disputes in an orderly way. In Sect. 16.2, Gricean maxims and Searlean speech act conditions are integrated in such a way that five general rules for communication can be formulated. In Sect. 16.3, starting from Lewis’s definition of convention, it is argued that the interactional effect of accepting is conventionally linked with the complex communicative act complex of argumentation. In Sect. 16.4, the rules for argumentation are placed in a dialogical perspective.

1 Introduction

Two people engaged in a discussion who try to resolve their dispute constitute the smallest possible community. They form a ‘community of minds’. Ideally, such a community of minds coincides with a larger group of people. But here we would like to concentrate on disputes between just two people.

A community, be it small or large, cannot survive, and, in fact, cannot exist if the members of that community try to avoid conflicts at all costs. The notion of community seems to suggest that agreement among the members of the community is preferable to disagreement. To a certain extent, this is true. However, it is sometimes necessary to motivate disagreement, otherwise no progress can be made and existing problems remain unsolved. On the other hand, motivating disagreement is counterproductive if it does not go together with providing the necessary means for settling disputes in a peaceful way. Any community needs established procedures for co-operation and co-ordination in order to reach solutions for disagreements.

We would like to point out what kind of rules for communication and argumentation are required in order to make it possible to resolve disputes in an orderly way. We hope to demonstrate that the rules presently being discussed are not completely alien to the rules which already exist among any given community of language users. In fact, they comply to a large extent with generally accepted rules for communication and argumentation. However, they are not completely identical with these rules, but constitute an extension and a critical regulation.

2 Rules for Communication

Verbal communication and interaction require the observance of various kinds of rules by the language users. These rules are pre-conditions for adequate communication and interaction. Four main categories of rules must be distinguished:

• 1. Syntactic rules for the production and interpretation of sentences and larger stretches of discourse.

• 2. Semantic rules concerning the meaning of the words and expressions which are used in these sentences and larger stretches of discourse.

• 3. Communicative rules for a recognizable and correct performance of the elementary and complex speech acts which are carried out in these sentences and larger stretches of discourse.

• 4. Interactional rules for an orderly and smooth conduct of the dialogues, conversations or other forms of (spoken or written) discourse constituted by the sequences of speech acts which are carried out in these sentences and larger stretches of discourse.

In order to make themselves understood, the language users must observe the syntactic rules of the language concerned (1). In order to make themselves understood, their formulations must be in accordance with the meaning of the words and expressions in the language concerned (2). In order to take part in verbal communication, they must observe the conditions for a ‘happy’ performance of their speech acts (3). In order to participate in verbal interaction, they must comply with a number of requirements for appropriate discourse (turn-taking, relevance, politeness, etc.) (4).

The categories of syntactic and semantic rules (1 and 2) refer to grammatical rules, the categories of communicative and interactional rules (3 and 4) to pragmatic rules. If the language users fail to observe the grammatical rules, they exclude themselves from the language community; if they fail to observe the pragmatic rules, they exclude themselves from the communicative community. The four categories are ordered hierarchically: (4) presupposes (3), and (3) presupposes (1) and (2). Of course, there are interrelations between all the four categories.

The rules for communication and interaction are social rules. Contrary to the laws of nature, they can be violated. Such violations may harm the comprehensibility or acceptability of the discourse in various degrees. In some cases, language users deliberately violate the rules in order to achieve a particular effect, as in indirect speech acts and conversational implicatures. Such an effect can only be achieved if the context, the situation or the general and specific background knowledge of the members of the community offer sufficient ‘compensation’ for the problems caused by the violation concerned.

The grammatical rules which are pre-conditions for adequate communication and interaction are the domain of linguistics proper. We shall discuss the communicative and interactional rules which are the domain of pragmatics. Major contributions to the theory of pragmatics are made by the philosophers Searle and Grice: Searle developed a theory of speech acts and Grice a theory of conversations. In order to reveal the similarities between Searle’s speech act conditions and Grice’s conversational maxim, we shall demonstrate how they can be integrated.¹

First, the Gricean Co-operative Principle must be re-defined into the more general and succinctly phrased Principle of Communication which states that language users be clear, honest, efficient, and to the point. The Principle of Communication summarizes the general rules which speakers and writers observe and which listeners and readers, when communicating, expect them to observe.

In practice, of course, it is not at all uncommon for one of the rules for communication to be broken, but this does not necessarily mean that the Principle of Communication has been abandoned altogether. If this is the case, however, then the person doing so is reneging on a basic convention of the community to which he belongs. Assuming that it is not clear that he is not in full control of his actions (he may be drunk for example), or that he cannot be held responsible for them, he will have to account for his defection or he will be faced with sanctions which may vary from an irritated reaction to a complete breaking off of the contact.

As an alternative to the Gricean maxims of Manner, Quality, Quantity, and Relation, the general rules which govern communication can now be rephrased in a Searlean way:

• 1. Perform no incomprehensible speech acts,

• 2. Perform no insincere speech acts,

• 3. Perform no unnecessary speech acts,

• 4. Perform no pointless speech acts,

• 5. Perform no new speech acts that are not an appropriate sequel or reaction to preceding speech acts.

The first rule implements the communication requirement “Be clear”. It corresponds to the recognizability conditions for the performance of speech acts: the propositional content condition and the essential condition. In order to be clear, the speaker (or writer) must formulate the speech act that he wishes to perform in such a way that the listener (or reader) is able to recognize its communicative force and to establish what propositions are expressed in it. This does not mean that he must be completely explicit, but it does mean that he is not allowed to make it impossible, or almost impossible, for the recipient to arrive at a correct interpretation.

The second rule implements the communication requirement “Be honest”. It corresponds to a part of the correctness conditions for the performance of speech acts: the responsibility conditions. It might be useful to note here that we refer to Searle’s sincerity conditions as responsibility conditions, in order to clarify what kind of commitments a speaker undertakes by performing a certain speech act, irrespective of the mental state he is in (1984, p. 195). The implication of the honesty requirement is that the speaker may be held responsible for having undertaken the commitments which are associated with the speech act concerned.

The third and the fourth rules implement the communication requirement “Be efficient”. They correspond to another part of the correctness conditions for the performance of speech acts: the preparatory conditions. The implication of the efficiency requirement is that a correct performance of a speech act must not be either unnecessary or pointless.

The fifth rule implements the communication requirement “Keep to the point”. It does not correspond to any speech act condition, nor does it refer to the performance of an individual speech act, whether elementary or complex. This requirement is concerned with the relation between several speech acts. The question here is whether, in the verbal and non-verbal context, the performance of a particular speech act is a relevant addition to the speech acts already performed. Thus, the relevance requirement “Keep to the point” relates to the sequence of speech acts and the function of a speech act in a particular speech event.

To fulfil the requirement “Keep to the point”, a sequel of speech acts or a reaction to a speech act must be appropriate. Precisely what comprises an appropriate sequel or an appropriate reaction is difficult to define in general terms. However, it is possible to indicate what this amounts to. Every speech act seeks to achieve the communicative effect so that the listener understands it, and the interactional effect so that he accepts it. So, the performance of a speech act expressing the fact that another speech act has been understood or accepted will be a relevant reaction. The same applies, of course, to the expression of non-understanding or non-acceptance. Giving reasons as to why something is or is not accepted, is also relevant.²

The rules for communication correspond to a large extent to Grice’s maxims. The main difference, which is also the main advantage, is that the maxims are now formulated as rules for the performance of speech acts. The first rule corresponds roughly to Searle’s propositional content condition and the essential condition. The second rule corresponds to his sincerity condition, the third and fourth rules correspond to his preparatory conditions, whereas the fifth rule does not have a counterpart in his conditions.

By integrating Gricean maxims with Searlean speech act conditions, both are enhanced. Compared to the maxims, the communication rules are more specific as a consequence of their connection with the Searlean conditions, and they are more general because they are no longer restricted to assertions, as they are with Grice. The speech act conditions also profit from it, because it has now been shown that the conditions for different speech acts are, in fact, specifications of more general rules for communication.³

The synthesis of Searlean and Gricean insights reveals the heterogeneous character of the original speech act conditions. Searle does not differentiate between their importance. In our revised version of his theory, we make a distinction between the propositional content and essential conditions on the one hand, and the sincerity and preparatory conditions on the other.

The need for this can be demonstrated by looking at the consequences of a violation of the various conditions. In the case of violation of the first two, no recognizable speech act has been performed, whereas in the case of violation of the second two, though the performance of the speech act is not quite successful, or happy in the full sense, a recognizable speech act is performed.

This crucial difference can be accounted for by realizing that there is a correspondence between the propositional content condition and the essential condition on the one hand, and Grice’s maxim of Manner (“Be perspicuous”), and our first rule of communication (“Perform no incomprehensible speech acts”) on the other. Violating these two conditions damages the recognizability of a speech act, whereas violating one of the two others affects its correctness because of insincerity, inefficiency, or irrelevancy. In order to express this difference terminologically, we refer to the first two as conditions for recognizability, and to the second two as conditions for correctness (cf. van Eemeren and Grootendorst 1984, p. 41).⁴

3 Interactional Effects

In what way are elementary or complex communicative (illocutionary) acts connected to their associated interactional effects (perlocutions)?⁵ We claim that this connection is, in a communicative community, to a certain extent, conventional. We know that in the literature on illocutionary and perlocutionary acts this is a matter of dispute. Hardly anyone disputes the conventionality of illocutionary acts, but opinions differ when it comes to the conventionality of perlocutionary acts. Austin and Searle take the standpoint that perlocutionary acts are never conventional, while Cohen (1973) allows for the possibility that perlocutionary acts may have just as good a claim to conventionality as illocutionary acts.

Let us begin by stating what we mean by conventionality. We shall start from a definition proposed by Lewis in Convention:

A regularity R in the behavior of members of a population P when they are agents in a recurrent situation S is a convention if and only if, in any instance of S among the members of P,

• 1. everyone conforms to R;

• 2. everyone expects everyone else to conform to R;

• 3. everyone prefers to conform to R on condition that the others do, since S is a coordination problem and uniform conformity to R is a coordination equilibrium in S. (1977, p. 42)

The nucleus of Lewis’s view is that a convention is a regularity in the behaviour of people brought about by a system of expectations. What, for example, is the significance of this view for the crucial relationship between the communicative act complex of argumentation and its associated interactional effect? First, it is necessary that a happy performance of the communicative act complex of argumentation regularly be followed by the occurrence of the associated interactional effect on the part of the listener, and, second, it is necessary that the speaker expects that this will happen. To what extent is this the case?

If regularity is the decisive criterion of conventionality, it is clear that the associated interactional effect of argumentation does not have much chance of qualifying to be called conventional, since there is no question of a regular occurrence of the associated interactional effect: one does not have to be a dyed-in-the-wool pessimist to dare assert that, in practice, an argument fails to be accepted just as often as it is accepted, and that the listener fails to be convinced by the argumentation at least as often as he is convinced.

Does this picture also apply for the speaker’s expectation? We believe not. The communicative act complex of argumentation and the interactional act of convincing maintain a bilateral relationship: argumentation is an attempt to convince, and for the performance of the attempt to convince, the speaker must argue. If the speaker had no confidence in his succeeding in convincing the listener with his argumentation, he would not have to argue. By arguing with the listener, instead of, for instance, giving him an order, the speaker indicates that he regards the listener as a reasonable judge who maintains the same standards for correct arguing as he does himself. Otherwise his argumentation would not comply with the correctness conditions for this speech act. This would mean that he would break a general rule for communication.

If speaker and listener have decided jointly to seek the resolution of a dispute, then it is in their interest to co-operate with one another and act in co-ordination. This means that, as far as possible, they must apply the same standards of judgment and that they must hold one another to these standards of judgments. If he wishes to fulfil the conditions for a happy performance of argumentation, the speaker will therefore prefer, in his attempt to convince the listener, to observe the same standards as the listener applies (or as the speaker thinks the listener applies) when making his judgment.

The speaker’s expectation that the listener will judge the argumentation by the same standards as himself, the fact that the listener may infer from the speaker’s decision to argue with him that the speaker expects him to apply these standards, and the fact that the speaker prefers to apply the same standards as the listener, indicates that the performance of the communicative act complex of argumentation is founded on the expectation that common standards are available for judging the argumentation and that these standards will be applied by the listener. This means that the occurrence of the associated interactional effect of argumentation may, from the point of view of the speaker, be called conventional in the sense in which Lewis uses the term.

The question which immediately has to be asked, of course, is to what extent the speaker’s expectation that the listener will apply the same standards of judgment is realistic. If we assume that, in conversations, a general Communication Principle operates, (and must operate to enable serious participants in a conversation to reach their objective), then it seems to us that, ordinarily speaking, the speaker may assume, precisely in the case of argumentation calculated to resolve a dispute, that the listener is taking a co-operative attitude and will, as far as possible, try to judge the argumentation by common standards. These common standards, observed by speaker and listener, will, in practice, not be based on an explicit accord between the participants in the conversation. According to Lewis, however, this is no reason for not referring to conventionality (1977, pp. 83–88). Following Barth, we propose to call such ‘implicit accords’, which are tacitly (‘implicitly’) accepted, semi-conventions (1972, p. 16). Since the speaker’s expectations, regarding the way the listener will proceed, are founded on the Communication Principle operating in conversations, we call these semi-conventions dialogical conventions.⁶

Naturally, such dialogical conventions can only apply to the behaviour of language users which they themselves can control. In his definition of convention, Lewis speaks of ‘a regularity R in the behaviour of members of a population P’. However, uncontrolled or even uncontrollable behaviour (such as automatic reflexes) is beyond the reach of conventions, and to some extent, this also applies to certain forms of ‘inner behaviour’—such as ‘considering’ and ‘feeling’—which are important for the achievement of the mental state of being convinced. The conventions of convincing can, therefore, be no more (and no less!) than act conventions, relating to the achievement of externalized, i.e. publicly stated, beliefs.

The dialogical act conventions for the conduct of discussions determine which speech acts are permitted. They regulate not the language users’ behaviour which is governed by grammatical rules, but their deliberate verbal acting. The conventions determine what the participants in the discussion may say and do in order to resolve a dispute. This means that, as regards argumentation, only the minimal associated interactional effect may be regarded as an effect to be achieved conventionally. The minimal interactional effect consists in the performance by the listener of the communicative act of acceptance. The optimal interactional effect would be that the listener is ‘really convinced’ (in a psychological sense), but this refers to a mental state which is inaccessible to direct observation and regulation.

In colloquial speech, the word convince is almost always used in the wider sense of striving after an optimal interactional effect. It will be evident that in this wide sense, the interactional act of convincing is not conventional. We use the term convince in the more limited sense of what we call the striving after the minimal interactional effect of acceptance. In this specific sense, which does not really conflict with the meaning of convince in colloquial speech, the perlocutionary act of convincing can be conventional.

4 Rules for Argumentation

If an argumentation theory is to be considered as a system of descriptive and/or normative rules for the performance of the communicative act complex of argumentation and the communicative act of acceptance, then, in our view, a dialogical design will be the most appropriate for that theory.

The speaker who performs the communicative act complex of argumentation is the listener in the case of the performance of the communicative act of acceptance, while, conversely, the language user who acts as listener in the communicative act complex of arguing is the speaker in the performance of the communicative act of acceptance. Moreover, the communicative act complex of argumentation, which is itself, qualitate qua, always a reaction to a particular utterance (or other sign) of doubt on the part of the listener, is always calculated to bring about in the listener the interactional effect that he react to the argumentation by performing the communicative act of acceptance. The smallest unit in the performance of argumentative communicative act complexes with effective minimal interactional effects is a completed dialogue in which the roles of speaker and listener are exchanged once and once only.

Argumentation is to be regarded as a communicative act complex at the textual level, and descriptive and normative argumentation theories must specify the rules determining the manner in which the speech acts performed by the speaker further or hinder, or ought to further or hinder, the performance of the communicative act of acceptance by the listener. An argumentation theory must provide the answer to the question in which cases particular communicative acts are (or ought to be) permissible in an argumentative dialogue and the question in which cases the associated acceptance interactional effect will (or ought to) occur. Only when this happens may we say that, in the argumentation theory, the ‘rules of the game’ are formulated for the performance of argumentative speech acts in discussions, and that this theory links up with the study of language use as it takes place in descriptive and normative pragmatics.

In a pragma-dialectical approach to argumentation, the idea of having a regulated discussion is considered as the basic principle of reasonableness. This requires the formulation of rules for such discussions. The dialectical aspect of this approach consists in there being two parties which attempt to resolve a dispute by means of a methodical exchange of moves, whereas the pragmatic aspect is represented by the description of these moves as speech acts.

In what way does the formulation of normative rules for critical discussions, as rules for the performance of speech acts, facilitate a natural connection to the descriptive conditions for performing elementary and complex speech acts in argumentative discourse? These conditions are closely connected with all kinds of general rules which govern everyday discourse and conversation, such as Grice’s maxims (1975), and the rules for turn-taking as described by conversation-analysts (cf. Levinson 1983; Edmondson 1981).

Our normative discussion rules can be seen as dialectical regulations of the rules that already apply in ordinary discourse. Of course, this is a simplification, but it draws attention to the fact that proposing normative rules for critical discussions has more ties with reality than some people think. To give an example, one could refer to the similarities between the starting point in the ideal model that the participants in a critical discussion must strive for the resolution of a dispute on the one hand, and the commonly accepted conversational fact that in ordinary conversation there is a preference for agreement among the interlocutors for the other.

If two language users jointly attempt to resolve a dispute by engaging in an interaction of speech acts, according to the rules, then their discourse can be referred to as a reasonable discussion. The rules of our ideal model for reasonable discussions specify what sorts of speech acts the participants in a critical discussion have to perform at the four stages of such a discussion, in order to contribute to the resolution of the dispute.⁷ The rules prescribe at what stage of the discussion the discussants are entitled, or indeed obliged, to perform a particular speech act.⁸

Starting from Searle’s taxonomy of speech acts (1979, pp. 1–29), it can be said that all kinds of assertive speech acts can be used to express standpoints and argumentation, and to establish the results of the discussion. The use of directive speech acts is restricted to challenging somebody to defend his standpoint and requesting him to put forward argumentation in support of it. Commissive speech acts are used to accept (or not accept) a standpoint, or argumentation, and to agree upon the division of dialectical roles in the discussion and upon the discussion rules. Finally, language usage declaratives, such as defining, precizating, amplifying, and explicitizing, can be helpful in avoiding a variety of misunderstandings. It should be noted that other types of declarative and all expressive speech acts are not listed in the model, because they don’t contribute directly to the resolution of a dispute.⁹

5 Conclusion

As we have argued elsewhere, the discussion rules described in our model derive their reasonableness from a twofold criterion: problem-solving validity and intersubjective or conventional validity (cf. Barth and Krabbe 1982, pp. 21–22). Together, the argumentation rules form an adequate procedure for resolving disputes which is intersubjectively valid for discussants who wish to resolve their disputes.¹⁰

The question remains to be answered how many people in the real world, if any, can be expected to live up to this strict canon of reasonableness for discussions? Are they people of flesh and blood, or are we talking about saints who live in a fictional world? We think real people are, in principle, not only perfectly capable of observing the discussion rules as formulated, but also act upon these rules when they try to resolve a dispute by means of a discussion. But, of course, not always, and not always completely without interference by unreasonable elements. Nobody is a saint and reasonableness is a matter of degree.

The normative rules of the pragma-dialectical model are, at least partially, congruous to the system of norms ordinary language users have internalized anyway. Empirical research has already shown many similarities—and also some differences (cf. Jackson and Jacobs 1981, 1982). In a future publication we shall come back to this in more detail.¹¹