Presumptions in Speech Acts

Abstract

The aim of this paper is to explore the viability of accounting for presumptions as a subtype of verdictives, within the framework of the Austinian approach to speech acts. The available set of felicity conditions is examined and worked out, in order to try and account in particular for a main feature of presumptions, namely, their function in shifting the burden of proof. In order to extend the Austinian framework as required, the notion of pragmatic presupposition accommodation is shown to be a useful device.

1 Introduction

The concept of presumption in the field of argumentation theory is usually introduced with a historical reference to Whately’s Elements of Rhetoric (1846) and his discussion of the probative duties and responsibilities incurred by the participants in a deliberative discourse (Hansen 2003; Kauffeld 1998). As Lilian Bermejo-Luque (2016) has noticed, there are two main views in the literature about presumptions, the inferential and the dialogical ones. According to the inferential view, presumptions are the conclusions of inferences drawn on the basis of a presumption rule and a premise stating a presumption-raising fact. The dialogical view on presumptions takes these to be claims categorized by a precise dialogical status. In contrast, Bermejo-Luque claims that presumptions can be characterized as non-inferential speech acts. Sharing this intuition with her, my aim is to further explore the presumptions-as-speech-act hypothesis, taking as a background the standard framework of speech act theory. Yet there is a point in which I will depart from her views. Bermejo-Luque contends that the correctness of presumptions does not depend on the pragmatics of dialogical procedures. Contrastingly, my aim is to explore the viability of accounting for presumptions within the framework of an Austinian approach to speech acts, whereby illocutions are conceptualized as resulting from communicative exchanges constrained by certain conditions.

Before starting, it is also obliged to mention Walton’s first speech-act account of presumptions. In his (1993), he characterized these as a speech act “half way between assertion and (mere) assumption (supposition)” (p. 125). Moreover, he gave a detailed set of “essential conditions” for the speech act of presumption in dialogue (grouped in subsets of preparatory, placement, retraction and burden conditions). These were seen as the procedural and dialogical conditions for raising presumptions. In my view, the intuition underlying this approach is correct. My main objection to this pioneering model has to do with the implausibility that such complex sets of rules can be seen as underlying the performance of competent speakers. Moreover, it seems to me that Walton’s first speech-act account of presumptions was not in fact theoretically independent from an argumentative frame. On the contrary, he seems to have been presupposing an argumentative setting where presumptions should have a role. Only under this presupposition makes sense that the retraction conditions embed the notion of “giving a good reason” (p. 139). This requirement was nevertheless not sufficiently supported. Remember that, for Walton, presumptions where a type of assertive speech act. It would be necessary at least to show how the requirement might be integrated within the framework of speech act theory, on pain of being subject to the objection of intellectualizing in argumentation-theoretical terms the well-founded notion of speech act.

My aim in what follows is to try and approach a concept of presumption in speech-act theoretic terms. This approach by no means aims to question the inferential account. It is intended to just show the plausibility of an alternative speech-act approach able to explain those pragmatic aspects of presumptions that cannot easily be accounted for in terms of premises and conclusions. My contention will be that at least some pragmatic aspects of presumptions can be accounted for in terms of illocutionary acts and their felicity conditions. Before being in a situation to present my own analysis, however, I am going to consider two prominent accounts of presumptions within the field of argumentation theory, namely, those due to Ullman-Margalit and Walton. My aim is to benefit from the significant work already available and approach the act of presuming taking into account what seems to me some relevant and well-established findings.

2 Main Features Characterizing Presumptions

Within the literature on presumptions, two main features have been identified as essential to characterize presumptions: (1) their function in communicative exchanges to enable the line of dialog to go ahead, even in the absence of sufficient evidence or reasons to support a claim or conclusion; and (2) their function in shifting the burden of proof and, as Walton puts it, “the shifting of the burden of rebuttal” (1993: 140).¹

The first aspect is already to be found in Ullman-Margalit (1983). She limits her inquiry to presumptions viewed as “an assumption made in advance of practical deliberation” (p. 143), where they play the role of rational prerequisites to arrive at a decision about action. Interestingly, when she discusses the particular example of a jury having to make a decision, she remarks that in the absence of sufficient evidence, or in cases of conflicting evidences, presumptions are allowed as long as no evidence or no sufficient evidence to the contrary is produced (p. 146). In line with Ullman-Margalit, Walton has noticed that presumptions in everyday dialog are very common devices used to assist dialog to move forward towards reaching its goal by argumentation used for that purpose. In his (1993), he already claimed that a presumption can be justified (“in reasoning”) on a practical basis, “on the grounds that it can enable the line of reasoning to go ahead, even in the absence of absolute knowledge of what will happen in a particular situation where some commitment to action or inaction needs to be made.” (p. 129) Also in his new dialogical theory (Walton 2008), where presumptions are conclusions resulting from a type of inferential reasoning, this aspect is vindicated. In my view, this important function in communication is directly related and may help to explain the second essential trait of presumptions to be here considered.

For a second aspect of the pragmatics of presumptions to be taken into account concerns its function in shifting the burden of proof and, as Walton put it, when considering the essential conditions for the speech act of presumption “the key idea is the shifting of the burden of rebuttal” (1993: 140) More importantly, also in his new dialogical theory he still suggests that in everyday dialog presumptions can be seen as a kind of speech act that shifts roles and burdens. He claims that presumptions are linked “to the shifting back and forth of an existing burden of proof in a structured dialog setting.” (p. 214). This essential feature has in fact been acknowledged and highlighted by most specialists working on presumptions, notwithstanding their different views of the dynamics of such effect.

Walton’s focus on everyday dialog is worth of notice. In legal argumentation, particularly in presumptions of law the procedure is usually institutionalized so as to regulate the turn-taking of participants and other aspects of the interaction. This regulation affects, among other things, the assignment and shifting of the burden of proof. In non-formal, everyday argumentative interactions, however, no such explicit arrangement is available. This makes it disputable that the very notion of presumption should require being associated to that of the burden of proof, which in its turn seems to presuppose an argumentative setting. One reason to oppose this widely accepted view has to do with the fact that this feature seems to challenge a straight approach to presumptions qua illocutions, particularly so if presumptions are to be seen as a subtype of assertive speech act—for which, in principle, no argumentative setting is called for. Nevertheless, it seems to me that the need to account for this important function in shifting the burden of proof is also what motivates the main intuition underlying the inferential account. For only in the course of a dialogical interaction seems a requirement concerning the burden of proof or rebuttal to make sense.

I take it to be correct that the very notion of presumption can be characterized by connecting it to a conventional effect, namely its shifting the burden of proof, to the extent that this effect can be seen as a defining feature that distinguishes presumptions from other assertive speech acts. Nevertheless, I do not think that this assumption necessarily entails that the communicative interaction is explicitly argumentative. For reasons to be considered below, presumptions can be seen as a subtype of verdictives, where the usual speaker’s obligation to support the utterance (by providing appropriate data, evidence, reasons, etc.) is shifted to the interlocutor.

This second aspect of presumptions, their shifting the burden of proof (or the burden of rebuttal) allows the exchange to go ahead under the presumption, that as such does not need to be supported by the interactants. Only when new evidences or reasons make it unreasonable to stick to the presumption (a circumstance that can adopt the form of the presentation of a rebuttal, but does not need to be so), shall the interactants reassess and restate their positions. From a speech-act theoretic perspective, what needs to be accounted for is the characteristic obligations and shifting of burden that presumptions involve.

To recapitulate, I propose that the following points can be considered well established and should constitute the points of departure of our exploration in what follows. Firstly, presumptions can be viewed as a type of assertive speech act, close to (but not identical to) assumptions or suppositions. Secondly, an important function of presumptions in communicative interactions is to enable the dialog go ahead, even in the absence of sufficient evidence or in cases of conflicting evidences, so as to facilitate the interaction. Thirdly, this function is made possible by means of the peculiar shifting in the burden of proof that characterizes presumption. In all that, I am here endorsing the analysis put forward by the above mentioned accounts.

Nevertheless, and in contrast with them, the analysis I am going to suggest, if acceptable, aims to show that an explicitly argumentative setting is not necessary for the act of presuming to take place qua communicative action. This claim in no way attempts to question other, well established models within the field of argumentation theory. Undeniably, the act of presuming can and play a role in many argumentative dialogues. Moreover, as the inferentialist models have showed, a presumption can be reconstructed as the conclusion of an inference drawn with the help of an inference rule. The analysis I am putting forward can be seen as complementary to those models. My reconstruction does not focus on the underlying inference possibly backing the introduction of a presumption, but on the pragmatic conditions and effects that characterise the act of presuming, as seen from a participant’s perspective.

Still, I think that my account can be relevant to argumentation theory in the following respect. As I am going to argue, a correct presumption is a reasonable endorsement that a speaker presents to an addressee (or audience) and that, if acknowledged by him or her as such, assigns a dialectical obligation to the addressee him or herself. The endorsement can be seen as reasonable to the extent that it is based on reasons, or might be supported by explicitating the reasons on which it is based. To that extent, the pragmatics of the speech act of presuming deploys an underlying dialectical structure of obligations to justify and acknowledgement of those obligations. It is this dimension of speech, as applied to the particular case of presumptions, what I am trying to elucidate by means of the present analysis.

3 The Framework of Speech Act Theory: The Austinian Approach

Now, my proposal is to see presumptions as a subtype of verdictive speech act,² within the framework of the Austinian approach. In order to explain the shifting in the burden of proof (or rebuttal) that presumptions characteristically accomplish, I will resort to the notion of Austinian presupposition accommodation, i.e. Austinian accommodation of felicity conditions, as this notion has been worked out within the above-mentioned framework.

According to Austin (1962) seminal work, speech acts are actions having certain effect on the social and interpersonal world of the interactants, an effect that he described as conventional.³ The Austinian approach to speech acts takes it that this effect is conventional due to the fact that it is brought about on the basis of an agreement, possibly tacit (and possibly institutionalized), among the relevant participants (Sbisà 2002: 433). Moreover, speech acts are seen as context-changing social actions (Gazdar 1981: 68), whereby making a speech act can be seen as bringing about changes in the social and interpersonal environment. The Austinian account contends that whereas locutionary acts bring about representations of states in the world, illocutionary acts produce changes in the normative states and relationships of the participants, namely, changes that affect their commitments, obligations and duties, as well as their rights, entitlements and powers, and other similar deontic stances (Sbisà 2002, 2006; Witek 2013, 2015) Briefly, the contention is that illocutions modify the commitments and entitlements of the interactants. Thus it is claimed that the central function of illocutionary acts is to bring about changes in the normative structure of social and interpersonal reality. Thus, following the Austinian account of speech acts,⁴ the pragmatic force of an illocutionary act is to be explained in virtue of its conventionally determined effects. These effects in turn should be described by reference to the normative states of the participants and the changes that the speech act brings about. Different types of pragmatic force are to be characterized depending on the type of conventional effects on the normative stances of the interactants.

To better explain why I think that this framework can be of help for an account of presumptions as speech acts, some clarifications can be in order here concerning the more general setting of speech act theory. As stated above, my contribution attempts to show that presumptions can be accounted for in purely pragmatic (illocutionary) terms and thus that presumptions can be seen as speech acts. The account is not aimed to call into question the inferentialist one, but to show how in conversational settings, where no presumptive rules of inference have been introduced (not in an explicit or conscious way), presumptions are nevertheless available as a speech-act conversational move. In my view, Austin’s notion of verdictive speech act allows us to understand why the corresponding illocution entails a justificatory obligation. Moreover, and considering other prominent approaches to presumptions, I take it that presumptions qua verdictives transfer this justificatory obligation to the addressee; how this move can take place is what remains to be elucidated within an Austinian approach. From this perspective, my contention is that the existence of justificatory obligations is inherent to the pragmatics of speech. Austin’s notion of verdictive nicely accounts for the justificatory obligation that is constitutively linked to these speech acts, without presupposing an inferential rule.

This is also a reason to prefer Austin’s vs. Searle’s perspective on speech acts. There are many points in common between both approaches and, in particular, it can be considered that Austin’s verdictives and Searle’s assertives overlap to a great extent, although not to all extents. For Searle, assertives are essentially speech acts whose illocutionary purpose is “to commit the speaker (in varying degrees) to something’s being the case, to the truth of the expressed proposition.” (1975: 12) Moreover, he declares that this class will contain most of Austin’s expositives and many of his verdictives. Within Searle’s taxonomy, presumptions could be seen then as assertives in informal, conversational contexts. In in institutional or socially institutionalised contexts, the act of presuming should in my view be typified as a declarative (i.e., an act able to “bring about some alteration in the status or condition of the referred to object or objects solely in virtue of the fact that the declaration has been successfully performed”, 1975: 17). This treatment, if correct, would be fully in line with the analysis here suggested, in what concerns the classification of presumptions within a given taxonomy.

My main reluctance to endorse Searle’s framework has to do with a conceptual point concerning the normativity of speech. Both Austin and Searle agreed that successful speech acts, correctly performed, produce an illocutionary effect. For Austin, this effect must be seen as conventional. For Searle, the means to attain it are conventional (the set of constitutive rules for the use of the force indicator device); yet the illocutionary effect is in itself not so, at least not in the general case. Presupposing that other conditions hold, the essential rule is stated in terms of the speaker’s illocutionary purpose, and the illocutionary effect amounts to the hearer’s uptake of this purpose. For Austin, the hearer’s uptake was a necessary condition, but he stressed the further condition that the effect of the speech act should take place as a result of the procedure being followed, as a conventional effect. In contrast to this notion of conventionality, in some of his writings Searle has explicitly accepted that the illocutionary effect be interpreted as the hearer’s uptake of the speaker’s intentions: “Illocutionary effect is a matter of understanding the utterance, and is roughly equivalent to Austin’s ‘illocutionary uptake’.” (Searle 1992: 140). We should remember here that uptake was for Austin only one necessary condition for a speech act to be successfully performed, the second one being the bringing about of a conventional effect.

In my view, Searle’s tenet makes it difficult to account for the main traits of presumptions concerning the kind of obligations and rights, and other normative stances that the interlocutors engage in. Even if, from Searle’s perspective, the speaker’s commitment to the truth of his/her statement can give rise to an obligation to justify, this cannot be seen as part of the illocutionary effect of the utterance. Moreover, it is difficult to considerer as part of the illocutionary effect what can be seen as one of the main specificities of the act of presuming, namely, its shifting of the burden of proof. For one thing, the fact that this shift takes place requires more than the speaker’s intentions that it be so. It requires the addressee’s recognition of his/her obligation in the context, given his/her acknowledgement of the meaning and force that the speaker’s utterance has had in the context.

Notwithstanding these considerations, and although Austin’s attention to the normative aspects of speech acts is more detailed than Searle’s, the above-stated tenets cannot be attributed to him in a straightforward way. This is why I am here applying and further extending what I have already termed the Austinian approach to speech acts (Sbisà’s 2002, 2006; Witek 2015). Sbisà’s original interpretation of Austin’s views proposes to consider the conventional effect of speech acts as conventional in that this effect depends on the interactants’ agreement that it has taken place: “I suggest (…) that the bringing about of conventional effects depends on agreement about their coming into being among members of the relevant social group” (Sbisà 2002: 48) Moreover, these effects are understood as changes in the normative relationships of the interactants, namely, in their mutual obligations and rights, and other similar deontic stances (cf. Sbisà 2006). Therefore, and taking into account the previous considerations, in order to extend this framework to presumptions, I will take a point of departure in Sbisà’s characterization of verdictives. This move can allow us a first, tentative approach, in which the usual felicity conditions for verdictives should integrate the two main features of presumptions as identified above.

4 Presumptions as a Subtype of Verdictive Speech Acts

According to Austin, verdictives are typified as the act of giving a verdict or finding, formal or informal, final or provisory, “upon evidence or reasons as to value or fact” (1962: 150, 152) In her reworking of this notion, Sbisà has proposed that verdictives are characterized by the following interactional effects:

A speaker performing verdictives in an interactional situation such as a public debate presents him or herself as willing to take on responsibility for the correctness of his or her claims and as acting on the basis of some kind of reliable and testable cognitive competence. The audience is supposed not merely to acquire the speaker’s findings, but also some of his or her evidence or reasons and thereby of his or her competence. (Sbisà 2006: 167)

Taking this extended analysis as a starting point, I take it that verdictives are illocutionary acts that consist of the issuing of a (possibly provisional) finding on the basis of evidence or reasons (i). They presuppose, as a preparatory condition, that the speaker has access to the elements required for a justified finding (data, criteria, and the like) on the relevant subject matter, so that she must count as cognitively competent (ii). They commit the speaker to giving evidence or reasons for her verdictive if requested (iii), and give license to the addressee to issue analogous verdictives on the same subject matter (iv).

My intention in what follows is to explore and give support to the suggestion that presumptions are a subtype of verdictive speech act. For the sake of simplicity, we are going to consider presumptions that are or can be made explicit in the context of a communicative exchange, so that it makes sense from the beginning to consider them to be a type of speech act.⁵ In such cases, my suggestion is that felicity conditions for presumptions should include,

• 1. the speaker’s power/right/entitlement to endorse the presumption and to legitimately expect that the addressee also endorses it, and

• 2. the participants’ license to issue analogous verdictives on the same subject matter.

Condition (1) can be seen as a basic, general requirement for any pragmatically correct verdictive, as based on the Austinian analysis presented above. If the condition is not satisfied, the speech act can be said to be unhappy or failed and thus not correctly performed.⁶ Condition (2) takes into account the response that the act, being correctly performed, conventionally invites on the part of the addressee, also in agreement with the analysis.

My justification to introduce these conditions is the following. According to Austin, a verdictive is uttered on the basis of some evidence, data, reasons, etc. In the particular case of presumptions, and given what we could describe as its counterfactual character (something is presupposed in the event that it is not proven otherwise), we may think that it must be a basis of reasons what allows one to state a presumption. Thus, the precondition that the speaker may count as cognitively competent amounts to his/her power/right/entitlement to endorse the presumption, on the basis of some available reasons to do so. Furthermore, it is this very availability of reasons what entitles the speaker to a legitimate expectation concerning the addressee’s endorsement.

To a certain extent, the precondition stating that there should be some reasons giving support to the presumption may be seen as answering to an intuition which is close to the inferentialist model. In particular, the present reflection can be seen in line with both Ullman-Margalit’s and Kauffeld’s appeal to the kind of justificatory basis that lends support to the presumption (more on that below). The difference between the inferentialist model and the present account is that, from the point of view of the pragmatics of communication, no concrete rule of inference must be supposed in order for a presumption to be taken as correctly performed. The reasons giving support to the presumption can be very different in kind. Yet this is the basis that authorizes the introduction of the presumption as a reasonable endorsement.

The notion of reasonableness has been broadly discussed in the literature on argumentation.⁷ Here, I am using this notion in an intuitive, pre-theoretical sense. It is applied to a verdictive speech act that counts, for the participants in the communicative exchange, as reasonable enough to be endorsed. Thus the speech act is reasonable (or not) from the participants’ point of view. My claim only amounts to saying that, in interpersonal exchanges in informal contexts, the speaker (whenever s/he is seriously following the procedure for correctly performing a presumption, and is not acting on secondary intentions) puts forward his/her utterance as a reasonable endorsement. This amounts to considering, from a participant’s point of view, that there must be some reasons available to the participants which justify the presumption.

Yet this does not by itself decide the question of whether the utterance must be taken as a presumption. From the point of view of the Austinian approach that is being worked out here, the addressee’s response plays a decisive role for the utterance to reach its intended illocutionary effect. For only if the addressee accepts the utterance as a presumption, and thus as a reasonable endorsement, can the speech act of presuming be said to have achieved its conventional effect and thus to have been correctly and fully performed. This fact does not exclude the possibility that, in the course of the exchange, the addressee may be willing to rebut the accepted presumption and to act accordingly, since accepting an utterance as a reasonable presumption does not commit one to accept it as true to the facts or correct (righteous, pragmatically appropriate, etc.) Finally, one may observe that in formal and institutional contexts, in contrast, presumptions are introduced and sustained by some authority. To that extent, their being reasonable presumptions is additional to their being in force and thus to their reaching the conventional effect they have.

Now, coming back to our hypothesis, we can examine more closely conditions (i)–(iv) as stated in the definition of verdictives, in order to try to determine whether or not they answer to presumptions. Condition (iv) above corresponds in a straightforward way to condition (2). But we may wonder whether conditions (i)–(iii) still hold. Condition (i) seems to be weakened in the following sense. In presumptions, as stated above, the reasons or evidence backing them do not need to be definitive or conclusive, but just enough as to make the finding reasonable. In this sense, the type of verdictive corresponding to presumptions could be said to be a reasonable endorsement (of a propositional content p). Moreover, condition (ii) invokes the speaker’s cognitive accessibility to the source of authority or to the grounding that supports the presumption. There will be cases in which this source or grounding is of a social or institutional character, so that the speaker who introduces the presumption can be seen as deferring her cognitive competency to this social or institutional setting.

This point has been nicely approached by professor Kauffeld, even if from an inferential point of view, when he claims that “To presume that p, in the ordinary sense of the term is to infer that p on the supposition some agent has made, is making, or will make it the case that p, rather than risk criticism, retribution, etc. for failing to do so” (2009: 3) Somewhat in accordance with this idea is also Ullman-Margalit, when she considers the particular case of conversational presumptions of truth, sincerity, and appropriateness of contribution to interpersonal communication. In those particular cases, these normative presumptions have to do “(i) with the question of which sort of error is morally or socially more acceptable, and (ii) with the moral or social evaluation of the regulative effect on people’s behavior of the presumption rules being instituted and operative” (1983: 161) Notice that in this second type of deference, in contradistinction to Kauffeld’s, it is not required that anyone be able to make it the case that p. What is merely required is that endorsing p be seen as preferable to not endorsing p, for reasons of a moral, practical, etc. character. There will be other cases, however, in which deference to an authority or to a social or practical regulation is not needed.⁸ All what is required by conditions (i)–(ii) is that either the speaker is herself able to provide the basis of evidence or reasons that support the presumption as a reasonable endorsement, or that she is able to defer this obligation to a social or institutional instance.

Notwithstanding this, it seems that condition (iii) is difficult to comply with. As a verdictive, presumptions seem to contradict the felicity condition mentioned above, namely, they do not seem to commit the speaker to giving evidence or reasons for her verdictive if requested. As we had highlighted, presumptions free the speaker from such burden, assigning at the same time a similar but opposite burden to the addressee concerning the questioning, rebuttal, weakening, etc. of the presumption. We can formulate this (reverse) condition in the following terms:

(iii)’ A presumption creates the obligation on the part of the addressee to give evidence or reasons for his opposition to endorse the presumption, whenever he or she does oppose it.

Now, the question to be explored is whether and how the Austinian approach can account for this central aspect of presumptions. A first attempt to solve this problem could consist of trying to identify, among the different types of speech act, the type that best answers to condition (iii)’. It seems to me that the type more appropriate in this respect would be that of exercitives. For, as we are going to see, exercitive speech acts can be characterized as acts which assign an obligation, duty, commitment, etc. to the addressee, by virtue of the influence, power, authority, etc. accorded to the speaker.

According to Austin, “An exercitive is the giving of a decision in favour of or against a certain course of action, or advocacy of it.” (1962: 154). Sbisà has proposed to reinterpret these speech acts as follows:

Exercitives [are] illocutionary acts consisting of the exercise of authority or influence. They presuppose some degree of authority or authoritativeness on the part of the speaker and assign or cancel rights or obligations to or from the addressee. (2006: 165)

As in the case of verdictives, we may say that the speaker’s authority or authoritativeness can be (although does not need to be) subsidiary to that of an external source or grounding. This analysis would allow us to consider presumptions as a type that shares a main trait of verdictives (since verdictives carry with them the burden of justification) and of exercitives (since exercitives assign an obligation to the addressee). But this solution could be subject to an important objection. Firstly, in those cases in which the addressee succeeds in challenging the presumption, because he or she convincingly shows that it was not sufficiently reasonable or because he or she has rebutted it, the speaker’s authority is cancelled with the presumption. Yet we had suggested that the speaker’s authority or authoritativeness (to assign an obligation to the addressee to the effect that it is the addressee who must challenge the presumption) could be inherited from an external source or grounding. How can this cancellation be effected in the local context of the interaction? The shifting in the burden of proof presupposes that the speaker’s authority or authoritativeness was from the beginning conditional, but it is not clear how the felicity conditions of exercitives could and should codify this.

A second attempt to deal with presumptions in terms of other types of speech act might be the following. Searle (1969) famously proposed a type of speech act, that of declaratives, to have a double direction of fit, that of words-to-world (like assertions) and that of world-to-words (like directives and commissives). Although Austin’s and Searle’s taxonomies do not completely coincide (and they answer to different criteria), for present purposes we may wonder whether presumptions could be seen as a declaration in Searle’s sense. For, according to him, “Declarations bring about some alteration in the status or condition of the referred to object or objects solely in virtue of the fact that the declaration has been successfully performed.” (1975: 17). In the case of presumptions, the suggestion would be that the particular alteration they bring about is precisely a shift in the burden of proof. This suggestion is nevertheless subject to an important objection. Taking presumptions to be declarations may be considered appropriate in certain cases, particularly in those in which the presumption is grounded on moral or legal reasons and there is an institutional framework supporting it. But it is doubtful that more informal presumptions as those raised in everyday interaction do require an institutionalized setting of that kind.

Alternatively, a second attempt to account for presumptions as a subtype of verdictive could take into account the notion of Austinian presupposition accommodation, as worked out by Witek (2013). This concept takes a point of departure in David Lewis’ notion of presupposition accommodation, understood as a rule-governed process whereby the context of an utterance is adjusted to make the utterance acceptable. Applied to illocutionary acts, the notion of Austinian presupposition accommodation aims to capture a rule-governed context-adjusting process whose function is to repair the context of an utterance to make the utterance a felicitous performance of a particular type of illocution. We should remember that, according to Austin (1962), the felicity of an illocutionary act presupposes (in the sense of conditions that must be seen as satisfied) that (a) the circumstances for the performance of the act are appropriate, and that (b) the speaker of the act is endowed with an appropriate illocutionary role (particularly, with the required normative power/authority/authoritativeness).⁹ Conditions (a) and (b) are then the conditions that can be accommodated by means of the Austinian presuppositions of an act. Witek has put forward the following rule:

If at time t speaker S makes binding illocution I, and if the felicity of I requires presupposition F to be satisfied by the objective context, and if F is not part of the objective context just before t, then -ceteris paribus and within certain limits- presupposition F becomes part of the objective context at t. (2013: 10)

Here, a binding illocution is a felicitous illocutionary act bringing about the conventional effect that is constitutively linked to this type of illocution. The Austinian presupposition is to be understood as a contextual state of affairs that is required by the felicitous performance of the illocution and is determined by the conventional procedure under which the act is performed. Austinian presupposition accommodation allows the interactants to adjust the context so as to make that certain conditions hold. This move has the effect of turning an utterance into a felicitous illocutionary act.

How is this process of presupposition accommodation related to presumptions? Remember that our problem was to account for the peculiar shift in the burden of proof that characterizes presumptions, and this aspect seems to go against condition (iii) for verdictives. In our attempt to explain this aspect of presumptions, we had stated the reverse condition (iii)’ (stating that a presumption creates the obligation on the part of the addressee to give evidence or reasons for his opposition to endorse the presumption, whenever he or she does oppose it). But in so doing, we had to acknowledge that this reversion seemed to approach presumptions to exercitives. Now, in order to account for the speaker’s authority or authoritativeness, our proposal is that the notion of Austinian presupposition accommodation can be appropriate and a useful device. Given condition (iii)’, the felicity of the speaker’s presumption requires that she be seen in the course of the interaction as endowed with the corresponding authority or authoritativeness. This in turn depends on the participants agreeing on this very fact, an agreement that can be implicit or tacit –in which case it is the interaction itself, and the response given by the addressee, what ultimately determines the accommodation. But presupposition accommodation can take place only against a background of pragmatic conventions, namely, the Austinian felicity conditions that jointly determine the performance of an act as a felicitous illocution of a given type.

For illustration purposes, we can imagine the following alternative dialogues between A and B about a third colleague of them, C.

(Example 1)

• A. Do you think it a wise idea to let C to manage the accounts?

• B. C has the required expertise. Besides, I presume his honesty.

• A. Ok.

(Example 2)

• A. Do you think it a wise idea to let C to manage the accounts?

• B. C has the required expertise. Besides, I presume his honesty.

• A. I wouldn’t, remember that in his previous position as a bookkeeper disciplinary proceedings were opened against him that haven’t yet been closed.

In Example 1, B can be said to have performed two conjoint verdictive acts. In the second one, “I presume his honesty”, the explicit performative “I presume” indicates the force with which the speaker is putting forward his utterance, as a necessary step to search his addressee’s agreement. For only if this condition is satisfied can the speech act of presumption be said to have been satisfactorily performed. Still, as a verdictive and according to the standard analysis, B has committed himself to providing evidence or reasons for his verdictive if requested. A’s answer indicates that she agrees with accepting B’s authority or authoritativeness to introduce the presumption as a reasonable endorsement in the consequential succession of events. In this way, A is conferring to B’s utterance, with her agreement, the status of a felicitous presumption, and moreover discharging B of justifying his verdictive.

In Example 2, however, A does not accept the presumption. The standard analysis of verdictives would take it that it is B who should support his questioned utterance by giving appropriate evidence or reasons. Nevertheless, the explicit performative as uttered by B prompts from A, together with her (indirect, yet explicit) refusal, an explicit reason that supports her refusal to agree with the verdictive as a reasonable endorsement. In so doing, she is answering to the obligation of carrying the burden of proof. This shift in the burden of proof can be pragmatically explained, as argued above, by an accommodation of the pragmatic presuppositions that are available in the context concerning the status and role of the participants. With her last turn of reply, A shows that she accepts to carry herself the burden of proof, thus tacitly recognizing the principled authority or authoritativeness of B to issue a presumptive speech act.

To the extent that B’s presumptive speech act is finally questioned and not agreed upon by A, it could be said to be a failed, infelicitous speech act. In terms of the Austinian approach, the lack of agreement in an informal, not institutionalized context (as is the case here imagined) determines that the utterance does not attain its (intended) conventional effect, namely, it does not acquire the force of a working presumption in the succeeding interaction. Yet, in order for A to express her refusal, she has tacitly acknowledged that the speech act was a presumption (since she has assumed the burden of proof), though a defective one (given that the utterance was rejected as an unreasonable endorsement).¹⁰ This outcome, paradoxical as it may seem, is nevertheless standardly assumed as a fertile result of speech act theory. An utterance can be said to be a defective, flawed or infelicitous speech act, yet be recognized as a performed speech act.¹¹

To sum up, under the proposed account presumptions are viewed as a subtype of verdictive acts, in which the Austinian felicity condition (iii) is accommodated by means of a process that confers to the speaker the authority or authoritativeness that is required by the alternative condition (iii)’.

5 A Note on Conversational Presumptions

The notion of conversational presumptions is introduced by Kent Bach and Robert M. Harnish in their (1979). As is well known, they follow Grice (1989) in viewing communication as an inferential process where the speaker provides, by what she says, a basis for the hearer to infer what the speaker intends to be doing. Still following Grice, part of the speaker’s intention is that the hearer identify the very act the speaker intends to be performing, and successful communication requires fulfilment of that intention (p. 4) As they say, “In general, the inference the hearer makes and takes himself to be intended to make is based not just on what the speaker says but also on mutual contextual beliefs” (p. 5) But in addition to mutual contextual beliefs, “there are two general mutual beliefs that the hearer relies on to make his inference. They are shared not just between S and H but among members of the linguistic community at large” (p. 7) These general mutual beliefs are called the linguistic presumption and the communicative presumption. Furthermore, they assume that cooperative conversations are governed by certain maxims (of quality, quantity, relevance, and manner) that they prefer to call conversational presumptions (pp. 62–63)

Bach and Harnish’s approach to communication is intentionalist. According to them, “to communicate is indeed to express a thought or, more generally, an attitude, be it a belief an intention, a desire, or even a feeling” (p. 15).¹² Thus the intended illocutionary effect is that the hearer should recognize that reflexive intention of the speaker. An illocutionary act is communicatively successful if the speaker’s illocutionary intention is recognized by the hearer. Now, observe that this concept of presumption accomplishes a cognitive role, in the cognitive context of mutual beliefs. Even if success in communication depends on the speaker’s complying with these presumptions, they argue, “When a person fails to fulfil one of them, H will take S as having spoken contextually inappropriately until or unless H finds a suitable explanation to the contrary.” (p. 63) In this sense, I would say that conversational presumptions as endorsed by the hearer accomplish the sort of shift in the burden of proof that is characteristic of presumptions in general. They also allow the exchange to go ahead, since only under the presumptions can the inferential process take place. To that extent, the notion of presumption in play in Bach and Harnish (1979) seems to feature the two main functions above mentioned.

Nevertheless, I think that this Gricean framework is subject to some objections. My first objection to Bach and Harnish’s view is the following. In agreement with the Austinian account, I take it that the norms of linguistic appropriateness in general should be formulated in terms of the objective context, rather that the cognitive (subjective) one. Gauker (1998: 153) has observed that we refer to the cognitive context of an utterance if our aim is to explain and interpret the speaker’s behaviour by attributing certain beliefs and intentions to her; but if our aim is to evaluate the speaker’s act, we have no alternative but to refer to the objective context of her utterance, namely, the set of worldly states relative to which the appropriateness of the utterance is to be evaluated.

The Austinian view on speech acts here assumed contrasts with the standard Gricean view, for which the force of an illocution is determined by the (reflexive, complex) communicative intention with which the corresponding utterance is made. In contrast, the Austinian approach contends that it is not the cognitive context of mental states and mutual beliefs what has to be taken into account in order to determine what counts as a felicitous speech act, but the objective, external contexts of social facts and interpersonal relationships. It is therefore assumed that there are normative facts (commitments, entitlements, and the like) constituting the objective context of speech actions as such.

My second objection has to do with the cognitive level in which the above mentioned presumptions are said to be at work. Remember that, according to the Gricean framework, the inferential processes that allow communication to proceed do not need (and are usually not taken to be) to be conscious, but can and do take place on a subpersonal level. In the current debate in argumentation theory, the notion of presumption which is at issue is an explicit or explicitable one, taking the form of an utterance on the level of the communicative dialogue. To that extent, it seems that Bach and Harnish’s communicative presumptions should be seen as a type of presupposition, in line with Grice’s original work.

6 Conclusion

We have put forward a tentative account of presumptions as a type of verdictive, within the framework of an Austinian approach to speech acts. In order to account for the peculiar shifting in the burden of proof that presumptions accomplish, we have proposed a reformulation of their felicity conditions, contending that presumptions (whenever they can be considered to be correctly, felicitously performed) assign to the addressee an obligation to justify, thus shifting the burden of proof that is constitutive of verdictives. To show how this can be effected, we have made use of the notion of presupposition accommodation. This pragmatic device helps to explain the process in virtue of which the speaker is endowed with the authority or authoritativeness that is required in the context of interaction.