1 Introduction

The development of argumentation theory has been greatly influenced by speech act theory. Several proposals in argumentation theory make use of it as a general theoretical framework. For instance, Jackson and Jacobs (1982, 1992, p. 161) characterize argumentation as “interactionally emergent structures organized around the function of managing disagreement” in which, as Jacobs (1989) points out, a variety of speech acts can be performed regarding the different purposes that arguers can have. In a similar vein, Hitchcock conceives arguments as abstract structures characterized as “premiss-illative-conclusion sequences” (2007, p. 121) in which a premise always constitutes an assertive, while the conclusion can be any kind of speech act.Footnote 1 More recently, Lewiński (2021b) has stressed the need to account for the limitations of approaches that assume an illocutionary monism and a dyadic reduction (i.e., the assumption that the interaction only involves two agents, namely, a speaker and a hearer), proposing a pluralism of speech acts which, in the case of argumentation, would be characterized as argumentative polylogues.

However, in this kind of theories no specific type of speech act characterizes an illocutionary act of arguing. In this respect, there are two theories that deserve to be highlighted. Van Eemeren and Grootendorst (1984) propose a Pragma-Dialectical approach in which the main goal of argumentation is to resolve a difference of opinion, and in which the act of arguing is conceived as an illocutionary act complex. According to them, “even the simplest argumentation for or against an expressed opinion contains […] at least two statements (cf. the datum and the warrant in Toulmin’s model)” (1984, p. 32).Footnote 2 Likewise, Bermejo-Luque (2011) proposes a linguistic-pragmatic model of argumentation in which she provides a characterization of the act of arguing as a second-order speech act complex which consists of adducing a reason (or reasons) and concluding a target-claim or conclusion. The speech act of arguing, from an illocutionary point of view, counts as an attempt by the speaker at showing that a target-claim is correct. Regardless of their differences, these proposals of arguing as an illocutionary act are of great interest because both make it much clearer that being a reason or being a target-claim depends on an illocutionary act which has its proper conditions. A mere sequence of utterances does not become a case of argumentation.

Both theories follow Searle’s (1969) account in considering that some conditions must be fulfilled for the speaker’s utterance (or set of utterances) to count as a speech act of arguing. The application of the Searlean account allows Pragma-Dialectics and the Linguistic Normative Model of Argumentation to provide a suitable definition of what it means for a speaker’s utterance to constitute a speech act of arguing. However, the Searlean characterization assumed in both is not exempt from problems. Specifically, I will argue that it involves two interrelated problems, namely, that it leaves out (i) the role played by the hearer in communication, and (ii) the normative effects that any speech act, including that of arguing, brings about. These normative effects consist in the production of changes in what Sbisà (2006), following Austin’s (1962) account, calls Deontic Modal Competence, i.e., the set of rights, obligations and entitlements that can be attributed to the participants of a communicative exchange and that can be modified and affected by the performance of speech acts. These effects concern not only the speaker’s but also the hearer’s obligations and entitlements.

In this paper, I will argue that, to account for the active role played by the hearer in the performance of illocutionary acts and these normative effects in the case of argumentation, we must distinguish between two different levels in the analysis of the speech act of arguing: one related to the speaker’s utterance, and another one related to the communicative exchange in which both the speaker and the hearer are involved. In the first level, the successful performance of the illocutionary act of arguing is associated with the fulfillment of the condition of the securing of uptake by the speaker, understood as the speaker making her utterance graspable for a potential hearer.Footnote 3 By contrast, in the second level of analysis, to determine whether the illocutionary act of arguing has been successfully carried out, it is necessary to take into account the hearer’s response. In this level, the illocutionary act of arguing refers to the speaker’s act which involves and affects the hearer in a certain way, i.e., it introduces changes in both the speaker’s and the hearer’s set of rights, obligations, and entitlements. Thus, in this second level of analysis, the successful performance of the illocutionary act of arguing would be associated with the production of changes in the deontic modal competence. I will draw from the Austinian perspective known as the interactional account of speech acts (Corredor 2021; Witek 2015) to characterize the distinction between the two levels of analysis.

The paper is structured as follows. In Sect. 2, I present the Pragma-Dialectical approach and the Linguistic Normative Model of Argumentation. In Sect. 3, I introduce the problems associated with Searle’s speech act theory (assumed by both the Pragma-Dialectical model and the Linguistic Normative Model of Argumentation), which yield an incomplete characterization of speech acts. In addition, I present the interactional approach to speech acts as an alternative to the Searlean one. Finally, I contend that, despite the modifications introduced by the Pragma-Dialectical approach to the Searlean account, it is not exempt from the problems previously outlined. In Sect. 4, I present the two levels of analysis that must be distinguished to account for the illocutionary effects consisting in the production of changes in the deontic modal competence. As I will argue, accounting for these illocutionary effects is crucial in order to provide a characterization of argumentation that does not entail the problems associated with the Searlean approach. In this section, the solution provided will be contrasted with LNMA. In Sect. 5, I will draw the main conclusions of this paper and sketch out some further implications.

2 Two Accounts of the Illocutionary Act of Arguing

Speech act theory has considerably influenced the development of argumentation theory. Pragma-Dialectics (van Eemeren & Grootendorst 1984) and the Linguistic Normative Model of Argumentation (Bermejo-Luque 2011) are two proposals that develop a systematic and full-fledged model in which they provide different characterizations of the speech act of arguing. In this section, I will firstly present the Pragma-Dialectical approach (henceforth referred to as PD). After that, I will introduce the Linguistic Normative Model of Argumentation (henceforth referred to as LNMA), which provides a solution to some problems that Bermejo-Luque observes in the PD model.

2.1 The Pragma-Dialectical Approach to Argumentation

The PD model of argumentation proposed by van Eemeren and Grootendorst (1984) submits that the main goal of argumentation is resolving a difference of opinion in a critical discussion. They adopt speech act theory and, more specifically, Searle’s (1969) account, as their theoretical framework because, according to them, it constitutes the most appropriate theoretical framework to characterize the speech act of arguing (1984, p. 23). In their approach, argumentation is conceived as a complex illocutionary act which is formed by elementary illocutions that have the illocutionary force of assertions which constitute an illocutionary set that stands in a relation of justification (or refutation) of an expressed opinion (which is not part of the speech act of arguing) (1984, pp. 34–35). In order to characterize argumentation as a speech act, they stem from Searle’s account of speech acts. However, they consider that Searle’s proposal presents some problems that must be solved for his account to be applied to the analysis and characterization of argumentation.

One of the main problems they recognize in the Searlean approach is that it only takes into account the communicative aspects of language (1984, p. 23), leaving out what they consider as their interactional aspects, which are expressed “in attempts to bring about perlocutionary effects” (1984, p. 23). In the case of argumentation, they contend that the speech act of arguing is always performed with the intention to produce two types of effects, namely, the illocutionary effect of understanding and the perlocutionary effect of acceptanceFootnote 4 (1984, pp. 25–26). Regarding the characterization of argumentation as a particular type of speech act, they advance the following hypothesis:

Our hypothesis is that in the communicative sense argumentation is a form of language use corresponding to the forms of language use characterized in the speech act theory as illocutionary acts and that as regards its interactional aspects argumentation is linked with the perlocutionary act of convincing. (1984, p. 29).

This consideration allows them to formulate the set of problems associated with the Searlean view of speech acts that must be addressed in order to characterize the speech act of arguing. As Snoeck Henkemans (2014, p. 43) points out, the speech act of argumentation, as it is characterized in Pragma-Dialectics, can be distinguished from the standard Searlean approach in three aspects. First, in Pragma-Dialectics, argumentation consists of (at least) two statements which can be expressed in more than one sentence (van Eemeren and Grootendorst 1984, p. 32; Snoeck Henkemans 2014, p. 43). Second, “argumentative utterances always have a dual illocutionary force: taken individually they are assertives, but together they form an argumentation” (Snoeck Henkemans 2014, p. 43). Finally, a speech act “can only be regarded as argumentation if it is linked to another speech act which expresses a standpoint” (2014, p. 43). The solution offered by van Eemeren and Grootendorst to these problems is based on the distinction between the illocutionary forces at the sentence level and at the textual level (1984, p. 34). Argumentation is understood as an illocutionary act complex which consists of elementary illocutions. The set of these elementary illocutions (which have the illocutionary force of assertions) is what constitutes the illocutionary act complex of arguing.

In order to characterize the speech act of arguing based on the (modified) Searlean account, van Eemeren and Grootendorst formulate the constitutive conditions that must be fulfilled for the illocutionary act of arguing to be happily performed, i.e., for it to count as a speech act of arguing (1984, p. 40). Among the conditions for the performance of the speech act, they distinguish between the recognition conditions and the correctness conditions (1984, p. 42). They establish this distinction because they consider that “[…] although an illocution may be recognized (e.g., the listener knows that the speaker intends to perform a particular illocution), it need not necessarily have to be entirely correct […].” (1984, p. 41). In this way, they claim that for the illocutionary act of arguing to be actually performed, the conditions that must be fulfilled by the speaker are the propositional content condition and the essential condition, which are formulated as follows (1984, p. 43)Footnote 5:

  1. 1.

    Propositional content condition: the constellation of statements S1, S2 (,…,Sn) consists of assertives in which propositions are expressed.

  2. 2.

    Essential condition: advancing a constellation of statements S1, S2 (,…,Sn) counts as an attempt by S [the speaker] to justify O [the expressed opinion] to L’s satisfaction, i.e. to convince L [the listener] of the acceptability of O.

In addition, for the illocutionary act of arguing to be considered as correctly performed, the following correctness conditions must be fulfilled by the speaker, which are the preparatory and the sincerity conditions (1984, p. 44):

  1. 3.

    Preparatory conditions:

    1. i.

      S believes that L does not (in advance, completely, automatically) accept the expressed opinion O.

    2. ii.

      S believes that L will accept the propositions expressed in the statements S1, S2 (,…,Sn).

    3. iii.

      S believes that L will accept the constellation of statements S1, S2 (,…,Sn) as a justification of O.

  2. 4.

    Sincerity conditions:

    1. i.

      S believes that O is acceptable.

    2. ii.

      S believes that the propositions expressed in the statements S1, S2 (,…,Sn) are acceptable.

    3. iii.

      S believes that the constellation of statements S1, S2 (,…,Sn) constitutes an acceptable justification of O.

According to van Eemeren and Grootendorst, if the recognition conditions are not fulfilled by the speaker, then the speech act of arguing has not been performed. In the case of the correctness conditions, if they have not been fulfilled, the illocutionary act of arguing would have been performed, but it would not constitute a correct illocutionary act complex of arguing.Footnote 6

2.2 The Linguistic Normative Model of Argumentation

The LNMA developed by Bermejo-Luque (2011) is framed within a pragmatic-linguistic approach to argumentation that incorporates a critical re-elaboration of Toulmin’s (1958) material conception of inference. In this model, argumentation is understood as a communicative activity consisting in an attempt by the speaker at showing that a target-claim is correct, that is, at justifying a target-claim, and it is characterized as a specific type of speech act.

Bermejo–Luque identifies a set of problems in the PD model that she addresses in her proposal. The first one has to do with van Eemeren and Grootendorst’s consideration of convincing as the intrinsic perlocutionary goal of argumentation (2011, p. 59). According to Bermejo–Luque, convincing would be one of the multiple goals that we could have when we argue, but not the only one. The second problem she identifies in the PD approach is that van Eemeren and Grootendorst exclude the claim the speaker is attempting to justify from the speech act of arguing. Instead, they treat it as a different illocutionary act “linked to the sentences uttered in argumentation” (2011, p. 59). She considers that, because of this exclusion, in the PD model the speech act of arguing would be equivalent to the speech act of adducing. The reason why Bermejo–Luque considers this a problem for Pragma-Dialectics is because, according to her, it poses serious consequences regarding the formulation of the constitutive conditions that would make a set of utterances count as an act of arguing (2011, p. 59). Finally, the third problem she recognizes in the PD model is that, although they characterize argumentation as an illocutionary act complex, they do so because they consider that the warrant is part of the act of adducing, whereas for Bermejo-Luque the speech act of arguing is a complex one because she takes the act of concluding as one of the constitutive parts of it (2011, pp. 59–60).

In order to characterize the speech act of arguing, Bermejo–Luque follows Searle (1969) in considering that there are certain conditions that must be fulfilled for the speaker’s utterance to count as a speech act of arguing. In her model, a speech act of arguing is characterized as a second-order speech act complex formed by two speech acts, namely, the speech act of adducing (a reason) and the speech act of concluding (a target-claim).Footnote 7 These are second-order speech acts because they can only be carried out by means of performing first-order speech acts; in particular, in LNMA the act of adducing and the act of concluding are carried out by means of performing two constative speech acts (R and C), which can be performed either directly or indirectly, or literally or non-literally (2011, p. 60). The constative speech acts R and C become speech acts of adducing a reason and concluding a target-claim because there is an implicit inference-claim which establishes a relationship between the content of both constative speech acts (2011, p. 60). The propositional content of this implicit inference claim is “if r [the content of R (and its pragmatic force)], then c [the content of C (and its pragmatic force)]” (2011, p. 61).

Bermejo–Luque claims that “conventionally, acts of arguing are attempts at showing a target-claim to be correct” (2011, p. 70). As she points out, we must be able to interpret the speaker’s act as an attempt at justifying a claim in order to say that the speaker is arguing. This amounts to saying that, for the speaker’s utterance to conventionally count as a speech act of arguing, the speaker must count as fulfilling certain conditions. Following Searle, Bermejo–Luque contends that “there are constitutive conditions that make certain performances acts of arguing” (2011, p. 70). These conditions would be the following (2011, pp. 71–72):

Preparatory conditions:

  1. (i)

    S believes that a claim R, having such and such pragmatic force, may be taken to be correct by L [the listener].

  2. (ii)

    It makes sense to attribute to S a conditional claim, with a certain pragmatic force, whose antecedent is “R is correct,” and whose consequent is “C is correct”.

  3. (iii)

    S takes the correctness of a claim C to be in question within the context of the speech-act.

  4. (iV)

    S takes a claim R to be a means to show a target-claim C to be correct.

Propositional content conditions:

  1. (V)

    The content of the reason is that a claim R’ is correct.

  2. (Vi)

    The content of the target-claim is that a claim C’Footnote 8 is correct.

Sincerity conditions:

  1. (Vii)

    S believes the propositional content of R in a certain way and to a certain extent, namely, the way and extent that correspond to the pragmatic force of the claim R’.

  2. (Viii)

    S believes that R being correct is a means to show that a target-claim C is correct.

  3. (iX)

    S believes the propositional content of C in a certain way and to a certain extent, namely, the way and extent that correspond to the epistemic pragmatic force of the target-claim C.

Essential conditions:

  1. (X)

    Adducing R with such and such pragmatic force is a means to show that a target-claim C is correct.

  2. (Xi)

    S aims to show that a target-claim C is correct.

Let’s illustrate this with the following example offered by Bermejo-Luque (2019, p. 664):

  1. (1)

    I promise I’ll take care, so don’t worry

Bermejo–Luque characterizes this utterance as a speech act of arguing in which two other acts are carried out: a speech act of adducing (a reason) and a speech act of concluding (a target-claim). According to her, two first-order speech acts (a promise and an advice) are reconstructed “as two indirect speech-acts of claiming connected to each other by the corresponding inference-claim” (2011, p. 60). In her account, the utterance of “I promise I will take care” constitutes a speech act of adducing in (1). This is so because, by uttering it, the speaker does not only promise that she will take care but, since she has also implicitly made the claim that, if (it is true that) she commits herself to take care, then (it is true that) the hearer does not have to worry, she is also adducing as a reason that she commits herself to take care (2011, p. 65). The utterance of “don’t worry” constitutes an act of concluding in (1) because in uttering it, the speaker is suggesting the hearer not to worry; given that she has implicitly claimed that if (it is true that) she commits herself to take care, then (it is true that) the hearer should not worry, she indirectly claims that the hearer should not worry, thus turning this act into an act of concluding (Bermejo-Luque 2011, p. 66).

As we will see in Sect. 3, despite the solutions offered by LNMA to the problems identified in the PD approach, the two models still entail a set of problems associated with the Searlean characterization of the speech acts they both assume. In the following section, I will present what I consider to be the main problems of the Searlean approach, and I will introduce the so-called interactional approach to speech acts, which is the account (or set of accounts) that have raised the specific problems that I will point out.

3 An Alternative to the Searlean Account: The Interactional Approach to Speech Acts

The Searlean characterization of the speech act of arguing provided by Pragma-Dialectics and LNMA gives raise to some important questions. Firstly, it poses the question whether it is appropriate to say that a speech act of arguing has been performed if there is no hearer that understands and recognizes the speaker’s utterance as a speech act of arguing, or if the hearer’s response is irrelevant for the performance of the speech act. In addition, if one only takes into account what a speaker does, it seems difficult to account for the changes in the set of rights, obligations, and entitlements of the participants of the communicative process that are produced by a speech act of arguing. Finally, we can also ask whether what we want to do when we argue is limited to trying to make our utterances understandable or getting our interlocutor to understand our utterances.

These questions, which were firstly formulated in relation to speech acts in general,Footnote 9 have been considered by many authors whose proposals are framed within the so-called interactional approach to speech acts (Carassa and Colombetti 2009; Clark 1996; Sbisà 2006, 2009; Witek 2015, 2019). In the next section, we will see how these considerations are also applicable to the characterization of the speech act of arguing.

3.1 Two Problems in the Searlean Account

Searle’s (1969) approach to speech acts has greatly influenced the development of speech act theory in general, as well as the study of argumentation. However, it has also given rise to important criticisms, such as the one pointed out by Clark (1996). According to Clark (1996, p. 137), in Searle’s view it is irrelevant whether the speaker’s act is received, read, or understood by a hearer.Footnote 10 Clark points out that “this view is, of course, absurd. There can be no communication without listeners taking actions too—without them understanding what speakers mean” (1996, p. 138). Clark’s criticisms have been subsequently taken up by other interactional approaches (Carassa and Colombetti 2009, p. 1840; Sbisà, 2009, p. 37). Here I will focus on two interrelated problems of the Searlean perspective that these approaches have pointed out. The first one is related to the mere passive role that such perspective attributes to the hearer, while the second one consists in that it leaves out the normative effects brought about by speech acts. Let’s now see these criticisms in more detail.

According to the speech act theory formulated by Austin (1962, pp. 115–116), the performance of a certain speech act is associated with the production of three types of illocutionary effects: (i) the securing of uptake, (ii) producing effects that change the normative facts, and (iii) the inviting of a response. Several authors have reformulated Austin’s considerations about the illocutionary effects produced by an utterance focusing on the effect (i), the securing of uptake. As Sbisà (2009, p. 35) indicates, Strawson (1964), motivated by his intention to make Austin’s proposal compatible with the Gricean theoretical framework, played a fundamental role in advancing and promoting the subsequent consideration of the securing of uptake as the central illocutionary effect associated with the successful performance of a speech act. His considerations greatly influenced further developments of the speech act theory, including Bach and Harnish’s (1979) as well as Searle’s proposals.

As Sbisà (2009, p. 37) points out, Searle agrees with Strawson in considering the securing of uptake as the only illocutionary effect that is essentially connected with the performance of an illocutionary act. In their view, this effect is conceived as the only illocutionary effect that must be overtly intended by the speaker in order to carry out a certain illocutionary act. Here is where the two interrelated problems that I mentioned above arise. The first one, illustrated by Clark’s remarks, is that this perspective assigns a mere passive role to the hearer: if the only thing needed for the successful performance of an illocutionary act is that the speaker overtly intends to secure the uptake, then the hearer’s response seems to play no role in the performance of the illocutionary act. Clark’s (1996, p. 139) own view, elaborated along Austinian terms, differs from Searle’s (1969) in two respects.Footnote 11 Firstly, Clark (1996) can be attributed a characterization of the securing of uptake that differs from the one offered by Searle and Strawson (as well as the one offered by Sluys (2018) and adopted by Bermejo-Luque; see footnote 3). Contrary to Searle’s view of the uptake, for Clark this notion refers to the hearer’s response which shows the hearer’s recognition (i.e., understanding) of the speaker’s act. Secondly, Clark assumes that the securing of uptake thus conceived (i.e., as the hearer’s response which shows their recognition of the speaker’s act), and not only the speaker’s overt intention to achieve it, is needed for the successful performance of an illocutionary act.Footnote 12

Relatedly, the second problem has to do with the fact that the Searlean perspective leaves out the second type of illocutionary effect formulated by Austin (1962, p. 116), i.e., the effect consisting in the production of changes in the normative facts (that is, in the set of rights, obligations and entitlements of the participants of the communicative process). This set of obligations, rights and entitlements is what Sbisà calls the Deontic Modal Competence (2006, p. 158). The fact that the Searlean approach disregards the illocutionary effect (ii) in the characterization of speech acts entails an important consequence, namely, that we would not be able to explain how it is possible that, when we carry out certain speech acts, the interpersonal relationship between speakers and hearers changes.Footnote 13

Let’s illustrate this with a very common example. When someone utters (2),

  1. (2)

    I promise I will be there at 8:00 pm

given the fulfillment of the conditions put forward by Searle (1969, pp. 57–61) (i.e., the propositional content condition, preparatory conditions, sincerity condition and the essential condition), we can say that the speaker has carried out an illocutionary act of promising, which is successfully performed when the speaker intends to produce a certain illocutionary effect (understood as the hearer’s understanding of the speaker’s utterance) by means of getting the hearer to recognize the speaker’s intention to produce that effect (1969, p. 60).

The characterization of the actual and successful performance of a speech act as dependent on these conditions allows Pragma-Dialectics and LNMA to account for argumentation in terms of what the speaker does, i.e., in terms of the sentence (or sentences) uttered by the speaker which (given the fulfillment of these conditions) would constitute a speech act of arguing. The hearer’s response would not be necessary for the performance (and successful performance) of the speaker’s act. The hearer’s role in communication and, more specifically, in the performance of the illocutionary act, is here reduced to just hearing (and understanding) the speaker’s utterance. In my view, this approach neglects how actual communicative processes work, where normally the speaker and hearer participate in the conversation actively, responding in a certain way,Footnote 14 and exchanging their role along the process.

Secondly, Searle’s characterization does not take into account how the performance of the illocutionary act (in this case, the illocutionary act of promising) changes the deontic modal competence of both the speaker and the hearer. In uttering (2), if the speaker gets the hearer to listen and understand their utterance, the speaker is acquiring a certain commitment, namely, that of delivering what has been promised. In this sense, the speaker’s deontic modal competence would have changed, but also the hearer’s. When a speaker promises something, in addition to the commitment they acquire, they also change the normative facts for their interlocutor: if the hearer responds (either explicitly or implicitly) by displaying how the speaker’s act has been received, then we would be able to say whether the speech act of promising has been successfully performed, i.e., whether the changes in the deontic modal competence have been produced, and thus, whether the hearer has acquired the legitimate expectation that the speaker will keep their promise. In the case of (2), for instance, if they have agreed to go to the movies, then the hearer will be entitled to expect the speaker to fulfill the promise to not be late.

Thus, we can see how both speaker’s and hearer’s interpersonal relationship has changed: by means of the speaker’s utterance and the hearer’s response they have introduced changes in the set of their rights, obligations, and entitlements, that is, in their deontic modal competence. And this is not an effect only associated with promises. It is an effect associated with the performance of any type of speech act, and which must be taken into account in order to offer a plausible explanation of how communication actually works. Also, the consideration of this second type of illocutionary effect involves taking into account the role of the hearer as not limited to merely hearing and understanding, but also as an active actor in the communicative process. The production of the normative effects associated with a speech act is not up to the speaker; the hearer’s response is what determines if the illocutionary act has been successfully performed, i.e., if the normative effects have been brought about.

3.2 The Interactional Approach to Speech Acts

Let’s now turn to the solution proposed by the interactional approach. In my view, a suitable definition of this approach is the one offered by Corredor:

What makes an approach to speech acts interactional is that it seeks to explain the illocutionary meaning of utterances by taking into account not only (nor primarily) the speaker’s communicative aims and intentions in issuing an utterance, but also the hearer’s recognition and interpretation in response to it (Corredor 2021, p. 464)

For the purpose of this paper, among the different interactional accounts of speech acts, here I will focus on some of the insights of Witek’s (2015) and Sbisà’s (2006; 2009) interactional proposals. In his paper “An interactional account of illocutionary practice” Witek (2015) proposes a neo-Austinian account which follows the Austinian perspective in considering that speech acts must be understood as context-changing social actions, as Sbisà (2009, p. 421) also suggests. Witek elaborates his own account following both Austin’s speech act theory (1962) and Millikan’s (2005) biological model of language. To do so, he draws on the distinction between the three effects that Austin (1962, pp. 115–116) associates with the performance of every speech act, i.e., (i) securing of uptake, (ii) producing effects that change the normative facts, and (iii) the inviting of a response.

The distinction between these three types of effects allows Witek (2013; 2015, p. 44) to distinguish between three approaches to the nature and structure of illocutionary acts, each one associated with each effect: intentionalist, institutionalist and interactional approaches. In the intentionalist account, the performance of a speech act depends on “uttering a sentence with the intention to produce effect (i)” (2015, p. 44). It amounts to get the hearer to “recognize the force and the meaning of the one’s utterance” (2015, p. 44). The institutionalist account takes the production of the illocutionary effects (ii) as the main function of illocutionary acts (2015, p. 44). Finally, in the interactional approach proposed by Witek, the force of an illocutionary act is defined “in terms of the effect of the (iii) type” (2015, p. 45), that is, in terms of “the response that the act conventionally invites or attempts to elicit” (2015, p. 45). The notion of convention that Witek (2015, p. 45) assumes has to do with patterns of interaction to which the speaker appeals to when uttering a sentence, which involves two elements: the speaker’s act and the hearer’s response.

On the other hand, Sbisà also offers a criticism of the Searlean approach to the role of the hearer. In “Communicating citizenship in verbal interaction: principles of a speech act-oriented analysis”, Sbisà (2006) proposes the use of certain conceptual tools based on the speech act theory for the analysis of discourse. She introduces a characterization of communication which, in contrast to what she calls the “received view”, does not understand it as a mere transmission of information. Rather, communication is characterized as a way of acting and introducing changes in the set of rights, obligations, and expectations of the participants of the communicative exchange (Sbisà 2006, pp. 151, 156). Hence, in Sbisà’s proposal, the hearer plays an essential role in communication. Specifically, Sbisà contends that “since the successful performance of the illocutionary act […] depends on intersubjective agreement as manifested in the hearer’s response, in order to determine whether some effect has actually been achieved […], the hearer’s response has to be examined too” (2006, p. 161).

Critically, Sbisà also addresses the second problem associated with the Searlean approach, i.e., that it leaves out what Austin refers to as the (ii) type of illocutionary effect. To do so, she delves into Austin’s characterization of this second type of effect. According to Sbisà (2006, p. 154), in Austin’s account the intersubjective relationship of the interlocutors is affected by the performance of the illocutionary act. This act is associated with the production of conventional effects, which are brought about only if there is intersubjective agreement on its production. The production of these conventional effects amounts to the production of changes in what Sbisà (2006, p. 158–159) refers to as the “Deontic Modal Competence”. In this regard, it is communication by means of illocutionary acts what produces changes in the deontic modal competence of the agents.

In “Uptake and conventionality in illocution”, Sbisà (2009) continues elaborating her interactional proposal, but focusing on the characterization of the notion of uptake in relation to the problem of the conventionality of illocutions. In particular, Sbisà (2009, p. 33) argues that illocutionary acts are conventional not (or not only) because the means used to perform them (such as performative formulas) are conventional, but because they bring about conventional effects. Sbisà stands against the view of uptake offered by Strawson (1964) and Searle (1969). As both Strawson and Searle do, Sbisà assumes “the indispensability (already stated by Austin) of uptake as a condition for the successful performance of illocutionary acts” (2009, p. 34). However, what she finds problematic is their consideration of the securing of uptake as the only effect that must be overtly intended.

Sbisà (2009, p. 44), delving into the characterization of Austin’s second type of illocutionary effect, claims that, in order to illustrate this effect, Austin (1962, p. 117) uses as an example the case of naming a ship. In this case, the effect of the act consists in that the ship acquires a specific name, and furthermore, that other acts like, for example, referring to the ship by using a different name, would be out of order (Sbisà 2006, p. 44). She claims that this effect is associated with the performance not only of conventional acts, as previously characterized, but with every illocutionary act. This interpretation, in her view, fits well with Austin’s account.

This example allows her to offer her own characterization of this illocutionary effect. Sbisà claims that, in the case of the effect of naming a ship, this effect consists in introducing changes in the norms. In particular, it consists in the enactment of a norm which manifests itself in the performance of subsequent acts related to the particular illocutionary act of naming a ship, and in how these subsequent acts are assessed. Sbisà (2009, p. 45) contends that this effect is conventional to the extent that “the state of affairs it brings about cannot exist without some kind of human intervention or decision”. In this way, she claims that the effect consisting in the creation of a norm depends on what the speaker does being “socially accepted as having that effect” (2009, p. 46). Sbisà concludes that “what is revealing of the conventionality of the illocutionary act (understood as the conventionality of its effects) is the need to secure uptake” (2009, p. 49), where the effect is conventional because it depends on the agreement between the members of the social community. This agreement, in turn, depends on the securing of uptake, not on its interpretation as an intention to secure it, but as an actual uptake.

3.3 Is Pragma-Dialectics a Solution?

As I pointed out in Sect. 2, the PD approach introduces a set of modifications in the Searlean account of speech acts in order to apply it to the analysis of argumentation. These changes allow van Eemeren and Grootendorst to account for what, according to them, are the interactional aspects of the language use. In the case of argumentation, these interactional aspects have to do with the perlocutionary act of convincing. Because of the changes they introduce in Searle's theory, which allow them to introduce into their analysis of argumentation what they consider to be the interactional aspects of language, one might think that the problems attributed to the Searlean approach assumed by them that have been presented throughout this section do not apply to their model, or even that their model poses a solution to these problems. After all, they explicitly introduce these interactional aspects as a constitutive part of argumentation. The question we should ask at this point is whether their sense of "interaction" is the same as the one embodied in the interactionist theories discussed in Sect. 3.2 above. As we will see, the type of interactional aspects that Pragma-Dialectics takes into account not only differ from those presented in Sect. 3.2, but also do not allow us to solve the two interrelated problems that have been exposed, which can be summarized as follows:

  1. 1.

    The listener is attributed a mere passive role in the performance of speech acts.

  2. 2.

    The normative effects that are produced through the performance of speech acts of argumentation are left out.

Taken these previous considerations into account, let’s see the reasons why Pragma-Dialectics would not solve the first problem. In characterizing the speech act of arguing, van Eemeren and Grootendorst (1984, p. 23, 25) claim that, when we perform speech acts, we do so with the intention not only to get the listener to understand our speech act, but also to get the perlocutionary effect consisting in the acceptance, i.e., in getting the listener to accept our speech act. Van Eemeren and Grootendorst distinguish between the recognition and the correctness conditions that must be fulfilled for a happy performance of a speech act of arguing. These are the conditions that “the speaker must fulfill if by uttering a complex of utterances he wishes to perform the compound illocution of argumentation […].” (1984, p. 40), and furthermore “that the listener may regard as having been fulfilled when he decides to treat a complex of utterances as argumentation.” (1984, p. 40). According to them (1984, p. 43–44), if the recognition conditions are not fulfilled by the speaker, the illocutionary act complex of argumentation has not been performed. By contrast, if the correctness conditions have not been fulfilled, the illocutionary act of arguing would have been performed, but not correctly.

In the formulation of both the recognition and the correctness conditions for the speech act of arguing, they introduce the listener in different ways. In the case of the recognition conditions, they formulate the essential condition of the act of arguing as an attempt by the speaker to convince the listener of the acceptability of the expressed opinion. On the other hand, in the case of the correctness conditions, they formulate the preparatory conditions in terms of the beliefs that the speaker attributes to the listener regarding the acceptability of the expressed opinion (first preparatory condition), the acceptability of the propositions expressed in the statements (second preparatory condition), or the acceptability of the constellation of statements as a justification of the expressed opinion (third preparatory condition).

What it seems to be here an attempt to account for the listener as an active part of the performance of the illocutionary act of arguing does not actually solve the problem formulated in Sect. 3.1. The crucial point here is that the active participation of the listener is not necessary to carry out the speech act of arguing. As they put it, what is necessary for the speaker’s utterance to count as a speech act of arguing (not necessarily as a correct one) is that the speaker fulfills the propositional content condition and the essential condition. They explicitly contend that the speech act of arguing, to be performed, does not necessarily need to be correctly performed: “the consequences [of an incorrect performance of the act] for both S and L may therefore be precisely the same as if it had been.” (1984, p. 44). If the recognition conditions (i.e., the propositional content condition and the essential condition) are fulfilled, then the listener is able to consider that the illocutionary act of arguing has been performed (either correctly or incorrectly) (1984, p. 44). Whether the illocutionary act of arguing has been correctly performed depends on factors that go beyond the conditions of its actual performance, for the performance of which the listener would continue to play a purely passive role.

Now that we have outlined a response to the question of why Pragma-Dialectics does not solve the first of the problems, let’s see why the theory also entails the second problem. Van Eemeren and Grootendorst argue that “argumentation may succeed while the attempt to convince fails.” (1984, p. 50). This is the reason why they make a distinction between the happiness conditions of the illocutionary acts of arguing and the perlocutionary act of convincing (1984, p. 50). According to them, the speech act of arguing has been happily performed if the speaker has performed it correctly and if they have achieved the illocutionary effect consisting in that the listener has understood the speaker’s act as an attempt to convince them of the acceptability of an expressed opinion (1984, p. 50). On the other hand, the perlocutionary act of convincing can be considered as happily performed if the speaker gets the hearer to accept the expressed opinion (1984, p. 50). The difference between the condition for each type of act, as they contend, can be found in the formulation of the second and third preparatory conditions they set for the act of arguing (1984, p. 50). These conditions, as they point out, are formulated from the speaker’s perspective regarding the listener’s attitude towards “the acceptability of the propositions expressed in statements S1 S2 (,…,Sn) or of the justificatory or refutatory potentiality of that constellation in respect of the expressed opinion O” (1984, p. 50). Regarding these conditions, they claim that, in the case of the perlocutionary act of convincing, it would be successfully carried out if the listener “actually subscribes to the attitude attributed to him by the speaker or at least if he wishes to tie himself down to that attitude.” (1984, p. 50).

Here it is necessary to stress that the kind of interaction that they have in mind and that they consider in order to characterize the speech act of arguing differs from the one endorsed by the interactional approaches presented above, and that I also adopt in order to offer a solution to the interrelated problems. Following Corredor (2021), I consider that the characterization of the perlocutionary act of the speech act of arguing as convincing is legitimate. However, as Corredor points out, for van Eemeren and Grootendorst:

[…] a complex speech act is an act of arguing provided that, and to the extent that, the listener grasps the attempt by the speaker to convince them […]. I do not share the underlying intuition that the communicative (illocutionary) dimension of argumentation consists of the listener’s understanding the speaker’s attempt, and that the interactional aspect is constrained to the perlocutionary effects. […] Instead, an alternative concept of interaction can be taken into account. (2021, p. 461).

In the interactional approaches to speech acts presented in Sect. 3.2, the interactional aspects of language concern the illocutionary acts, i.e., what in Pragma-Dialectics belongs to the communicative aspects. Furthermore, in contrast to the interactional approaches to speech acts, in Pragma-Dialectics the illocutionary effects associated with the performance of the illocutionary act of arguing are limited to the understanding by the hearer of the speaker’s act, leaving out the normative effects associated with the performance of any speech act.Footnote 15

4 Two Levels in the Analysis of the Illocutionary Act of Arguing

In analyzing argumentation, we might have two different interests: on the one hand, we can be interested in accounting for what the speaker does, the speaker’s utterance; on the other hand, we can be interested in accounting for the communicative exchange that takes place between a speaker and a hearer in argumentation, which is associated with the production of certain normative effects. In my view, the main virtue of LNMA and Pragma-Dialectics model is that they provide a sound account of the first, i.e., of why a speaker’s utterance may count as a speech act of arguing. In the case of the interactional account presented above, the main virtue is that it enables us to account for speech acts as communicative exchanges involving a speaker and a hearer, as well as for the normative effects that they bring about.

In this section, I will offer a solution to the two interrelated problems outlined in Sect. 3, which will allow us to retain the main virtues of LNMA and Pragma-Dialectics, while at the same time incorporating the insights of interactional approaches. To exemplify my solution, I will contrast it with LNMA, focusing on the discussion of Bermejo-Luque’s example (1). Although I will just focus on LNMA here, I consider that the solution I will describe could be applied to the Pragma-Dialectics model as well, insofar as both models entail the problems presented above.

Specifically, I will argue that, to avoid the problems of the Searlean perspective, we must distinguish two different levels in the analysis of the speech act of arguing. This distinction will allow us to account for what constitutes a successful performance of an illocutionary act of arguing within each level. The first level of analysis is associated with the speaker’s utterance, whereas the second level is associated with the communicative exchange, where both speaker and hearer play a fundamental role in the performance of the illocutionary act of arguing. In the first level, in order to know if the illocutionary act of arguing has been successfully performed, we only need to take into account the speaker’s utterance. Furthermore, in this level, the happy performance of the illocutionary act of arguing just requires the fulfillment of the condition of the securing of uptake (as it is conceived by Bermejo-Luque) by the speaker. By contrast, when we situate ourselves in the second level (i.e., the level of the communicative exchange), in order to determine whether the speech act of arguing has been successfully carried out, we need to take into account not only the speaker’s utterance, but also the hearer’s response. The hearer’s response is what shows how the hearer has received the speaker’s act and thus, if the normative effects (i.e., the changes in the deontic modal competence) have been produced. If this is the case, then we can say that, from this level of analysis, the speech act of arguing has been successfully performed.

As we will see, the distinction between these two levels allows us to show how the evaluation of the same act of arguing varies depending on the level which we pay attention to.

Let’s consider again example (1):

  1. (1)

    I promise I’ll take care, so don’t worry

Here, as we saw in 2.2., if the speaker fulfills certain conditions, then their utterance would count as an attempt at showing that a target-claim is correct, i.e., as an act of arguing. In the first level, the level of the speaker’s utterance, the only thing that is needed for the speaker’s act of arguing to be considered as successfully performed is that they count as trying to justify a target-claim. In order to count as trying to justify a target-claim, they must count as fulfilling certain conditions (2011, p. 72), which in LNMA would involve the securing of uptake (that is, they must make their utterance graspable for a potential hearer).

This example shows that, if we situate ourselves in the first level of analysis, a Searlean approach, such as the one adopted by Bermejo-Luque, is able to account for speech acts of arguing whose performance only depends on what the speaker does, i.e., the speaker’s utterances. However, as I have argued, this Searlean characterization of the speech act of arguing only in terms of what the speaker does entails the two interrelated problems presented along Sect. 3. In order to overcome these problems, my proposal is to distinguish a second level of analysis, where not only the speaker, but also the hearer play a fundamental role. In this second level, the successful performance of a speech act of arguing is associated with the production of normative effects (i.e., changes in the deontic modal competence). Consider this initial list of possible normative effectsFootnote 16:

  • The hearer’s legitimate expectation that the speaker can provide more reasons to justify the target-claim if requested to do so by the hearer.

  • The hearer’s entitlement to ask for more reasons.

  • The speaker’s commitment to the truth of the implicit inference-claim.

  • The speaker’s obligation to provide more reasons if requested to do so by the hearer.

  • The hearer's conditional obligation to accept the correctness of the target-claim unless they can produce reasons to the contrary.

Now consider this re-elaboration of example (1):

  1. (3)

    a. I promise I will take care, so don’t worry

    b. Ok. Don’t forget to take lots of pictures!

    c. Sure!

In this case, in order to know if the speech act performed by the utterance of (3a) constitutes a successful illocutionary act of arguing, we need to pay attention to the second level, which means that we need to take into account the hearer’s response (3b). It is this response what shows how the hearer has received the speaker’s utterance, and hence whether the speaker’s utterance counts as a speech act of arguing in this second sense. In this case, the hearer’s response (3b) shows that the hearer has received the act performed by the speaker as an act of arguing. When this is the case, it can be said that the normative effects have been produced and, therefore, that the speech act of arguing has been successfully carried out. The set of rights, obligations, and entitlements of both the speaker and the hearer has changed; now we can say that the hearer has a legitimate expectation that the speaker is able to provide more reasons to show the correctness of the target-claim, that the speaker is thus obliged to provide more reasons if requested to do so by the hearer, and so on.

As I argued, the distinction between the two levels of analysis is relevant because the hearer’s response is what shows us how the speaker’s act has been interpreted and, thus, whether the normative effects associated with the performance of the illocutionary act have been produced. To see more clearly the importance of distinguishing between the two levels of analysis, consider the consequences of the following variation in the hearer’s response:

  1. (4)

    a. I promise I will take care, so don’t worry

    b. Every time I’ve been in the car with you, I’ve seen you using the cell phone and exceeding the speed limit. So, please, don’t make promises that you won’t keep.

    c. Okay. Well, goodbye then.

Once again, if we situate ourselves in the first level of analysis, taking into account only (4a), it is correct to claim that (given that some conditions have been fulfilled) the illocutionary act of arguing has been successfully performed: the speaker has secured the uptake of their utterance, so it counts as an attempt at showing that a target-claim is correct. However, if we pay attention to the second level (the level of the communicative exchange) and consider the hearer’s response (4b) as well, can we still say that the speech act of arguing has been successfully performed? In this case, by means of (4b), the hearer is raising doubts about the commitments that the speaker is trying to acquire when they say “I promise I’ll take care”. The speaker has secured the uptake, which can be seen in the hearer’s response (4b), but what the hearer’s response shows is that they do not consider the speaker as a reliable one to perform the speech act of promising.

In the second level of analysis, for the illocutionary act of arguing to be hold as successfully performed, it must be considered that the illocutionary effect consisting in producing changes in the deontic modal competence has been produced. In this example, the hearer’s response shows that they recognize that the speaker is intending to carry out a promise, and that it would have been successfully performed if the speaker were reliable. However, since this is not the case, the normative effects associated with the speech act of promising have not been produced. The lack of production of the relevant normative effects would make the speaker’s act of arguing unsuccessful, at least in this second level of analysis.Footnote 17 This would mean that, for instance, the hearer no longer has the legitimate expectation that the speaker will provide more reasons to show the correctness of the target-claim, nor the speaker is obliged to do so if requested by the hearer, etc.

As I view it, my proposal has two main derived implications. The first one has to do with the possible interpretations of the evaluation of argumentation in LNMA and the interactional approach that I have presented. The second consequence is related to the different roles that can be attributed to the hearer in an argumentative exchange. Let’s first delve into the first implication, taking (4) again as an example. For Bermejo-Luque, by uttering (4b), the hearer would be recognizing the speaker’s utterance (4a) as an act of promising that has been successfully carried out; however, the hearer questions the truth of the implicit inference-claim, i.e., “if (it is true that) I commit myself to take care, then (it is true that) you should not worry”. In this case, a promise has been successfully performed, but what the speaker is promising is not a sufficient reason for the hearer not to worry; in a nutshell, (4b) would amount to something like “your promises are blown away by the wind”. By contrast, in the interactional approach that I have outlined, by uttering (4b) the hearer shows that the (alleged) promise the speaker intended to do by means of (4a) is an unsuccessful one. Consequently, the normative effects (e.g., the acquisition of certain commitments or legitimate expectations) have not been brought about. What (4b) shows is that, although the hearer does not consider the utterance of “I promise I will take care” as a happy promise, they do recognize that the speaker wants their utterance to count as a reason for the hearer to not worry. In short, (4b) would amount to something like “You can’t promise that; you are simply blabbering, so I can't just stay calm. What you're doing is trying to make a fool of me”.

The difference between the interpretation of the evaluation of argumentation from both LNMA and the interactional approach leads to the second derived implication of my proposal. Specifically, the distinction just made has consequences regarding the different roles that can be attributed to the hearer in communication and, in this case, in argumentative communication. On the one hand, the analysis of the example carried out from LNMA shows that the hearer must be characterized as an interpreter of what the speaker says. From this view, the hearer’s only task would be to interpret the speaker's utterance. On the other hand, from the interactionist perspective, the hearer is not understood as a mere interpreter of the speaker’s utterance. Rather, the hearer’s role would be that of being one of the parts of a process that involves (at least) two parts. In this sense, we can characterize the hearer as an active part of the communicative exchange, who contributes to put certain normative changes in place together with the speaker. This distinction allows us to contend that we, as hearers, have two different ways in which we can present ourselves to what we are said: it seems that, sometimes, hearers function simply as interpreters of what the speaker says or does (that is, interpreters of the speaker’s intentions); at other times, by contrast, they function as one of the parts of a process that involves more than one part. This plurality in the consideration of the hearer’s function is possible because of the plurality of levels of analysis. While on the first level the hearer’s role would be that of being an interpreter of the speaker’s act, in the second level the hearer would be conceived as one of the (at least) two parts of the communicative process.

5 Conclusions

In this paper, I have presented van Eemeren and Grootendorst’s PD model and Bermejo-Luque’s LNMA in order to show that both models provide a suitable characterization of the speech act of arguing from the point of view of what a speaker does. However, as far as they assume a Searlean perspective in the characterization of the speech act of arguing, they leave out the role of the hearer in the performance of the illocutionary act, and the normative effects produced by an illocutionary act of arguing. I contended that these problems have been raised by different authors whose proposals can be framed within the interactional approach to speech acts.

By applying some insights of the interactional approaches to the study of argumentation, I have argued that, in order to account for the role of the hearer and the normative effects that speech acts of arguing introduce in the set of rights, obligations and entitlements of the participants of the communicative exchange, it is necessary to distinguish between two different levels of analysis: the level that only involves the speaker utterance, and the level that involves both the speaker and the hearer.

I would like to conclude by saying that this analysis opens the possibility of investigating other features of the evaluation of argumentation. For instance, one can ask if it is only the speaker who is accountable for the argument put forward, or if the hearer must be taken into account as well when attributing responsibility. This can be seen clearly in cases of harmful argumentation (e.g., racist arguments) where not only the speaker, but also the hearer’s response can play a key role in sanctioning certain things as arguments worth debating or not, with important consequences for those whom such harmful speech is directed to. In this way, it would be interesting to investigate if the distinction between the two levels of analysis could allow us to account not only for the responsibility that can be attributed to the speaker regarding the correctness of the argumentation (e.g., in cases of fallacious discourse), but also for the responsibility that can be attributed to the hearer regarding the interpretation of the speaker’s act of arguing.