ISSA Proceedings 2002 – A Pracmatic View Of The Burden Of Proof

logo  2002-11. A dialectical profile of the division of the burden of proof
In an earlier paper, entitled ‘Strategic maneuvering with the burden of proof,’ we have explained our dialectical perspective on the division of the burden of proof in a critical discussion (van Eemeren and Houtlosser, 2002). We did so by answering a series of interrelated questions from a procedural view of critical reasonableness: Why is there a burden of proof? A burden of proof for what? For whom? What exactly does the burden of proof involve? When is it activated? What means can be used to acquit oneself of the burden of proof? And when is one discharged? Because our responses were given in a critical rationalist vein, they are attuned to resolving a difference of opinion by critically testing the acceptability of a standpoint in the most systematic, thorough, perspicuous, and economic way. In the present paper we aim to complement this approach by offering a pragmatic solution for an important problem that may arise in ‘mixed’ disputes, where opposite standpoints are put forward regarding the same issue. The problem concerns the order in which the opposing standpoints are to be defended.

Making use of an analytic tool provided by Walton and Krabbe (1995), we describe the interactional situation in which our problem arises with the help of a dialectical profile. This profile specifies the moves that are admissible when dividing the burden of proof in a mixed dispute in the opening stage of a critical discussion. The profile starts from the situation that a mixed dispute has come into being in the confrontation stage between two parties. The profile includes both possibilities: the one in which the party that has advanced a positive standpoint is challenged first to defend this positive standpoint and the one in which the party that has advanced a negative standpoint is challenged first to defend this negative standpoint.


We are here concerned with the interactional situation that comes into being when a party, in response to the other party’s challenge (in turn 1), refuses (in turn 2) to defend his standpoint. When asked (in turn 3) why he does not want to defend his standpoint, this party can (in turn 4) challenge the other party to defend his opposing standpoint. As the profile specifies, in such a situation the other party has (in turn 5) two possibilities: either he concedes to begin defending his own standpoint or he rejects the challenge. If the other party rejects the challenge, the first party may (in turn 6) require an explanation why the other party does not want to defend his standpoint. At this point, the other party may (in turn 7) no longer return the challenge, because he has already challenged the first party in his very first move (turn 1). Instead, the other party may initiate a deliberation on the order in which the standpoints at issue are to be defended.

Thus the dialectical profile makes it clear that the problem of establishing the order in which two opposing standpoints are to be defended amounts to a procedural problem concerning who will be the first to assume the burden of proof in a mixed dispute. In the opening stage of a critical discussion, a deliberation may be started over the order in which the defenses should take place, and this deliberation is to be initiated by the party that has started the process of challenging. The dialectical profile also makes it clear why this procedure is so. It is only after the party that has been challenged initially (in turn 1) has returned this challenge (in turn 4), that the order of defense can become pertinent. The order of defense can only be made an issue by the other party in the subsequent turns (turn 5-7).

2. Acquiring a burden of proof
The dialectical profile clearly specifies how and when the order of defense can become an issue in a mixed dispute, but it does not specify how it can be decided what the order should be. In the various treatments of this burden of proof problem in the scholarly literature on argumentation, various kinds of would-be solutions have been proposed: epistemological, juridical, ethical, etc. In our pragma-dialectical approach we opt for a more general stance. We think that the way in which this problem is to be resolved depends in the first place on the institutional practice or context in which the discussion takes place. The opening stage of a critical discussion is designed precisely to accommodate the kinds of procedures and conventions that are operative in the various institutional practices and contexts. There are practices that are genuinely institutional, such as criminal lawsuits and parliamentary debates, and where fixed procedures determine how issues of order should be decided. There are also practices where no fixed procedures exist, but where nevertheless certain conventional rules are operative that are in agreement with the goals of the practice concerned. In a broader perspective, all everyday verbal interaction can be regarded as institutional in the Searlean sense (1969) that performing speech acts is a form of institutional, rule-governed behavior and specific types of speech acts in specific kinds of exchanges are subjected to specific kinds of conventions. If no genuine institutional procedures are operative in the context in which a discussion takes place, these specific kinds of conventions provide a pragmatic rationale for deciding on issues such as order of defense. In the remainder of this paper, we intend to explain what this pragmatic rationale consists of and how it can account for a certain decision on the order of defending when two opposite standpoints are advanced.

We start by presenting first two dialogues in which the parties advance opposing standpoint, and the first speaker requires the second speaker to defend his opposite standpoint first. The standpoint that introduces the issue is represented in italics. In the first dialogue, this standpoint involves an implicit accusation:
1. S1: My purple vase!
2. S2: Yes, what a pity, isn’t it?
3. S1: You dropped it!
4. S2: I did not!
5. S1: Make me believe you didn’t
6. S2: I beg your pardon?!
7. S1: Why not?
8. S2: Well, …

In the second dialogue, the standpoint is an informative assertive:
1. S1: Jan is leaving for Warsaw tomorrow
2. S2: When exactly?
3. S1: Ten a.m.
4. S2: Is that so?
5. S1: Yes, isn’t it?
6. S2: As far as I know, the train departs every odd hour

In a pragma-dialectical reconstruction of these dialogues as a critical discussion, the dispute can in both cases be characterized as mixed because the parties take opposite positions in regard of an issue: in dialogue (1), the issue is whether S2 has dropped the vase; in (2), the issue is whether the train leaves at ten a.m. In both disputes both parties have a standpoint of their own. Consequently, in both cases both parties have an obligation to defend their standpoints[i]. There is a problem, however. Temporarily or definitively (we cannot tell), the party whose standpoint is put forward first shifts the burden of proof to the other party, but this shift seems in case (2) more or less legitimate, but certainly not in case (1)[ii]. We think that by examining how in ordinary argumentative discourse a burden of proof is acquired and what the pragmatic rationale for attributing such a burden of proof can be, we will be able to explain this difference.

3. Relating the burden of proof to the pragmatic status quo
Reconstructing what people say and intend to convey in argumentative discourse as a series of moves in a critical discussion, as is aimed for in pragma-dialectics, amounts to an explicit analysis of these people’s ‘dialectical’ commitments to certain propositions. Such an analysis can only be achieved if the dialectical commitments of the parties involved in the discussion can be derived from the ‘pragmatic’ commitments that are inherent in the way in which they have expressed themselves in the discourse, whether explicitly or implicitly. These pragmatic commitments can be traced by making use of insight provided by theories of language use that focus on how mutual obligations are incurred and acquitted in verbal communication and interaction, such as the Searlean speech act theory and the Gricean theory of rational exchanges.

In the first place, Searlean speech act theory and Gricean theory of rational exchanges can be called upon in explaining the rationale for attributing certain pragmatic commitments to the participants in argumentative discourse. As Jackson (1995) observes, the Gricean maxims, in particular the Maxim of Quality (“Do not say what you believe to be false or that for which you lack adequate evidence”), support the general presumption that an assertion advanced in the discourse – and in our opinion this also goes for other types of speech acts – is acceptable. According to Jackson, this presumption is cancelled only if the interlocutor (1) has independent reasons to doubt whether the assertion is indeed acceptable or (2) whether the speaker is indeed behaving in a cooperative way, or (3) if the context indicates that the speaker himself deems his assertion less acceptable for the interlocutor (1995: 258). Ullman-Margalit (1983) expresses basically the same idea when she says that from a legal perspective an assertion being ‘presumptively acceptable’ means that the interlocutor is entitled to regard it as acceptable[iii].

In our opinion, the presumption of acceptability has an even more fundamental basis in the Interaction Principle. Van Eemeren and Grootendorst (1991) state this principle as a general prohibition against the performance of any speech acts that are not acceptable to the interlocutor. Unlike the Gricean maxims, the Interaction Principle involves a real requirement. A violation of this principle does not encourage alternative interpretations of what is said. On the contrary, such a violation obstructs the normal course of the interaction, and can even lead to sanctions[iv]. Anyone who performs a speech act is committed to complying with the requirement involved in the Interaction Principle, and this commitment gives rise to the presumption that the speech act that was performed is indeed acceptable. This presumption is similar to the presumption that motorists that approach a red light will obey the connected traffic rule and stop their car.

Until there are clear indications of the opposite, the interlocutor is thus entitled to regard the speech act performed by the speaker or writer as acceptable. If, however, there are indications that the speaker or writer has not fully committed himself to the requirement involved in the Interaction Principle, the situation is different. When, for instance, a speaker makes it known in advance that he anticipates opposition from his interlocutor, and – following up on this – the interlocutor does indeed express opposition to the speech act concerned, then the presumption shifts to the interlocutor. To regain the presumption of acceptability, the speaker has to adduce evidence that his speech act is acceptable after all. In other words, he has acquired a burden of proof. Only after the speaker has succeeded in acquitting himself of this burden, the presumption shifts back to his position. If the interlocutor then intends to maintain his opposition, he, in turn, should acquit himself of the burden of proof for his opposite position. This is the only way in which he can regain the presumption for his opposition (see Rescher, 1977).

We think that this analysis can be taken a step further by observing that it is reasonable to let the presumption of acceptability remain with a speaker as long as the speaker’s speech act does not go against the prevailing pragmatic status quo. This means that his speech act may not be at odds with the set of premises that are mutually shared by the parties involved in the interaction. This set of premises represents the ‘pragmatic’ status quo because rather than to warranted beliefs or the general state of knowledge in a certain field, as in the ‘cognitive’ or ‘epistemic’ status quo, it refers to the list of premises that the particular parties involved in the discourse explicitly or implicitly accept and that define their interactional relationship in the interactional situation at hand[v]. The pragmatic status quo is challenged as soon as one of the parties involved performs a speech act that is inconsistent with the shared premises, for example because the state of affairs presupposed by its identity or correctness conditions conflicts with one of more of the commonly accepted premises.

4. Violating the pragmatic status quo
When may a speech act be assumed to be inconsistent with one or more mutually shared premises? We think that Kauffeld’s (2002) analysis of the way in which a burden of proof is incurred in every day verbal interaction can be of help in answering this question. In Kauffeld’s view, it depends primarily on the nature of the speech acts concerned when people engaged in verbal interaction incur a burden of proof and what the burden of proof involves. This means that the illocutionary point of a speech act and the implications of having made this point in a felicitous way are of decisive importance.

In our view, Kauffeld’s account has the merit of complementing concerns with dialectical obligations in ideal situations with a pragmatic concern about the way in which burdens of proof are assumed in everyday verbal interaction. He achieves this complementation by showing how the performance of certain speech acts, i.e., proposing and accusing, can endow the speaker with certain probative obligations[vi]. We think that Kauffeld’s approach can be generalized and applied to all verbal interaction by means of speech acts. In our outline of how we think such a generalization can be realized, we adapt Kauffeld’s idea that certain speech acts may have implications that – possibly or presumably – go against the interlocutor’s interests. Our adaptation amounts to taking Kauffeld’s idea to mean that a speech act may have implications that go against the interlocutor’s view of the interactional relationship between the speaker and the interlocutor encompassed in the present pragmatic status quo.

According to our adapted account, a proposal would invite an adjustment of what the interlocutor until then took to be the shared expectation of how the interactional relationship between the communicators should be in the future; an accusation invites an adjustment of what the interlocutor so far regarded as the shared view of the relationship between them. In our conception of a pragmatic status quo, this would mean that both a proposal and an accusation have implications that are likely to be inconsistent with the list of mutually shared premises – or at least with what the interlocutor supposed the list to be.

Searle’s taxonomy of speech acts may be of help in determining which types of speech acts may have implications that run counter to the interlocutor’s view of his current interactional relationship with the speaker. ‘Commissives,’ for instance, can generally be expected to have implications that agree with the interlocutor’s view of the interactional relationship between the speaker and the interlocutor. ‘Directives,’ on the contrary, can easily have implications that are in disagreement with the interlocutor’s view. As a rule, promises do not introduce actions that the interlocutor will think inconsistent with agreed-upon desirables, but with requests this may quite well be the case. There is at least one class of speech acts in Searle’s taxonomy that contains both types of speech acts. This is the class consisting of the ‘assertives.’ Some assertives are designed to provide the interlocutor with information that he did not possess before but that is expected to be consistent with what he already knows, such as ‘informing’ and ‘explaining.’ There are also assertives, however, that aim to make the interlocutor accept a view that he did not accept before and that cannot be expected to be consistent with what he already accepts, such as ‘claiming’ and ‘accusing[viii].’

5. A pragmatic view on deciding the order of defense in a mixed dispute
Now we have explained what we mean by a pragmatic status quo and how we can determine whether or not a speech act may be considered to violate this status quo, we return to the problem of the order in which two opposing standpoints are to be defended in a mixed difference of opinion.

In the pragma-dialectical view of argumentative confrontation, the speech act that initially introduces the issue can acquire the status of a standpoint in a dispute in two ways: either the person who performed that speech act makes it clear that he anticipates that the interlocutor will not accept this speech act at face value or the interlocutor makes it known that he is not prepared to accept the speech act at face value by performing a counter speech act (see van Eemeren, 1987; Houtlosser, 2002). In the first case, there is no presumption attached to the initial speech act, because the speaker or writer makes it clear from the start that this speech act may go against the prevailing pragmatic status quo between him and the interlocutor. In the second case, the speech act concerned initially has a presumptive status, because for all the speaker or writer knows – and also for all we know – this speech act does not violate the prevailing pragmatic status quo. This presumptive status is, of course, canceled when the interlocutor opposes this speech act with a counter speech act.

Let us assume for a moment that the interlocutor opposes the speaker’s initial speech act with a counter speech act not only in the second case we discussed but also in the first situation, in which the speaker has made it clear that he anticipates such opposition. The interlocutor’s reaction then agrees completely with this anticipation. Both cases can now be regarded as involving the kind of interactional situation of maximal opposition that can pragma-dialectically be reconstructed as a mixed dispute: the two parties have assumed contradictory standpoints and each party has a duty to defend its own standpoint. All the same, there is an important difference between the two interactional situations. In the first case, the standpoint that initiated the dispute has no presumptive status from the start, whereas in the second case it has. And the interlocutor’s opposition has a presumptive status in the first case, but not in the second. In the second case it is, after all, precisely the interlocutor’s opposition that first challenges the pragmatic status quo that is up to then supposed to prevail.

What are the implications for handling the burden of proof of this discrepancy between these two different interactional situations in a mixed dispute? In ‘Strategic maneuvering with the burden of proof’ (2002), we have argued for a conception of the burden of proof as consisting in an obligation for a party in a dispute to defend its standpoint if challenged to do so, but we have also argued for the acknowledgement of an additional, procedural obligation that was pointed out by Hamblin (1970): the burden of initiative. Besides an obligation to defend a standpoint, a burden of initiative implies an obligation to defend this standpoint at this particular juncture of the discussion. Distinguishing the obligation to defend a standpoint from the obligation to defend it at this particular junction of the discussion allows for the existence of an interactional situation in which a certain party has an obligation to defend a standpoint, but is not required to acquit itself of this obligation now. That is, at a particular juncture, a party that has advanced a particular standpoint does not have the burden of initiative[ix].

It is precisely the additional obligation of having the burden of initiative that we just emphasized which makes for the difference in the burden of the parties in the two cases we just discussed. In the first case, the speaker has both an obligation to defend his standpoint and an obligation to start the defense. In the second case, he does have an obligation to defend his standpoint, but not an obligation to defend it immediately. He is only required to defend his standpoint after the interlocutor has defended his standpoint. Whereas the order in which the two standpoints are to be defended coincides in the first case with the order in which they have been put forward, in the second case it does not. The latter of the two dialogues we presented at the beginning of our paper is, not coincidentally, an example of the interactional situation in the second case:
1. S1: Jan is leaving for Warsaw tomorrow
2. S2: When exactly?
3. S1: Ten a.m.
4. S2: Is that so?
5. S1: Yes, isn’t it?
6. S2: As far as I know, the train leaves nine twenty

The first speaker’s assertion (in turn 3) has acquired the status of a standpoint because of the second speaker’s opposition (in turn 4). Nevertheless, the presumptive status of the first speaker’s assertion is preserved because at the stage in which it was performed there were no indications that he performed a speech act that could be regarded as going against the prevailing pragmatic status quo; consequently, this speech act cannot bestow a burden of initiative on him. First, the interlocutor should justify his opposition. Once he has done so, the first speaker’s assertive looses its presumptive status and this speaker is obliged to accept the burden of initiative. Then he cannot escape any longer from defending his assertive against the interlocutor’s opposition.

6. Conclusion
In this paper we have given substance to our pragmatic view of the burden of proof. Our claim was that a burden of proof is incurred as soon as a speech act goes against a prevailing pragmatic status quo. The concept of a pragmatic status quo can be specified in terms of a list of premises that are explicitly or implicitly accepted by the people who are having a dispute and define their current interactional relationship. Criteria for determining whether or not a burden of proof is incurred can be established by exploiting the idea that the performance of particular types of speech acts may have implications that go against the interlocutor’s view of this interactional relationship. Decisions on the order in which two opposite standpoints must be defended can be justified by giving a truly pragmatic interpretation of the burden of proof concept that differentiates between a conditional obligation to defend a standpoint and a burden of initiative.

[i] In a critical discussion, advancing a standpoint implies assuming a conditional obligation to defend the position expressed in that standpoint. When two opposing standpoints are advanced by different parties, both parties are required to defend their position.
[ii] In (2) it would indeed have been odd if S2 would in turn 6 have said that S1 should first prove that the train leaves at ten a.m.
[iii]  In the law, the notion of presumption is applied to situations in which something is an ‘impending issue.’ What to do, for example, when someone has been absent for more than seven years: Should this person be declared dead or not? For legal purposes, it is then presumed that this person is dead. Ullman-Margalit (1983: 148) emphasizes this feature when she says that “[p]resumption entitles deliberators to make an assumption that they are otherwise not entitled to make.” Jackson’s use of the notion of presumption conforms to the legal use on the condition that the acceptability of a speaker’s assertion can be considered an ‘impending issue.’ What to do when someone has said something: Accept it or not? The presumption is: accept, unless there is something that weighs against it.
[iv] The Gricean maxims, which are Jackson’s basis for the presumption of acceptability, are not rules in the same sense. Unlike violating a “real” rule, violating a maxim does not lead to any sanctions but to an interpretation of the speaker’s meaning that is different from the literal ‘utterance meaning’ (assuming the Cooperation Principle still applies). Thus, in a Gricean perspective, the fact that the Maxim of Quality is not violated does not warrant the conclusion that what the speaker asserts is presumptively acceptable. Given that none of the other maxims are violated either, and, again, the Cooperation Principle still holds, it is only warranted to conclude that nothing else was meant than was literally said.
[v] What Rescher (1975) and others have called a ‘cognitive status quo’ (or ‘epistemic status quo’) is in fact subsumed in our concept of ‘pragmatic status quo.’ The concept bears some relation to Walton & Krabbe’s (dialectical) concept of ‘dark-side commitments.’
[vi] “[I]n many kinds of illocutionary act, S does not, at least not typically, engage a larger obligation to provide, on demand, reason and evidence vindicating the truth and adequacy of her primary utterance. […] But, other things being equal, where S makes a proposal or levels an accusation, she cannot responsibly dismiss an addressee’s demand for proof” (Kauffeld 2002, italics by the author). For empirical confirmation of this theoretical observation in as far as it concerns ‘accusing,’ see van Eemeren, Garssen and Meuffels (2003).
[vii]  In his analysis of proposing, Kauffeld claims that the major reason for having to justify an act of proposing is that the one who proposes something is supposed to have good reasons for what he proposes and if he aims at having his proposal accepted he should inform the interlocutor of these reasons. In his analysis of accusing, Kauffeld suggests that a major reason for having to justify an act of accusing is that the accused party has a right to deny the accusation and can only do so properly if the accuser has provided reasons for his accusation.
[viii] The declaratives, in particular ‘language declaratives’ such as definitions and specifications, are likely to be open to the same problem, just as the ‘expressives.’
[ix] This is, in fact, a different way of making Rescher’s well-known distinction between an I(nitial)-burden of proof and an E(vidential) burden of proof.

Eemeren, F.H. van (1987). For reason’s sake: Maximal argumentative analysis of discourse. In F.H. van Eemeren, R. Grootendorst, J.A. Blair & C.A. Willard (Eds.), Argumentation: Across the Lines of Discipline. Proceedings of the Conference on Argumentation 1986 (pp. 201-215). Berlin: Walter de Gruyter.
Eemeren, F.H. van, Garssen, B., & Meuffels, B. (2003). “I don’t have anything to prove here”: Judgements of the fallacy of shifting the burden of proof. (This volume)
Eemeren, F.H. van & R. Grootendorst (1991). The study of argumentation from a speech act perspective. In J. Verschueren (Ed.), Pragmatics at Issue. Selected Papers of the International Pragmatics Conference, Antwerp, August 17-22, 1987. Volume I (pp. 151-170). Amsterdam/Philadelphia: John Benjamins.
Eemeren, F.H. van, & Houtlosser, P. (2002). Strategic maneuvering with the burden of proof. In F.H. van Eemeren (Ed.), Advances in Pragma-Dialectics (pp. 13-28). Amsterdam/Newport News, VA: Sic Sat/Vale Press.
Hamblin, C.L. (1970). Fallacies. London: Methuen.
Houtlosser, P. (2002). Indicators of a point of view. In F.H. van Eemeren (Ed.), Advances in Pragma-Dialectics (pp. 169-184). Amsterdam/Newport News, VA: Sic Sat/Vale Press.
Jackson, S. (1995). Fallacies and heuristics. In: F.H. van Eemeren, R. Grootendorst, J.A. Blair & C.A. Willard (Eds.), Analysis and Evaluation. Proceedings of the Third ISSA Conference on Argumentation. Vol. II (pp. 257-269). Amsterdam: Sic Sat.
Kauffeld, F. (2002). Presumptions and the distribution of argumentative burdens. In F.H. van Eemeren & P. Houtlosser (Eds.), Dialectic and Rhetoric: The Warp and Woof of Argumentation Analysis. Dordrecht: Kluwer Academic Publishers.
Rescher, N. (1977). Dialectics: A Controvery-Oriented Approach to the Theory of Knowledge. Albany, NY: SUNY Press.
Ullmann-Margalit, E. (1983). On presumption. Journal of Philosophy 80, 143-163.
Searle, J.R. (1969). Speech Acts. An Essay in the Philosophy of Language. Cambridge: Cambridge University Press.
Walton, D.N., & Krabbe, E.C.W. (1995). Commitment in Dialogue: Basic Concepts of Interpersonal Reasoning. Albany, NY: State University of New York Press.