Introduction

From the interactionist perspective of language studies, the study of textual and linguistic resources is important because they constitute procedures for (inter)acting. In this study, the value of these resources is measured and evaluated in context, since it is through them and non-verbal procedures that interlocutors can act and react, repair interactional problems, jointly define the situation they are in, claim authority over domains of knowledge, endorse social roles and identity images, negotiate power (a)symmetries, etc. (Ford; Fox; Thompson, 2002a; Heritage, 2013a; Kerbrat-Orecchioni, 1992; Schegloff, 2007). Therefore, the systematic study of language and text occupies a central place in this perspective. However, we do not postulate autonomy of grammar and text in relation to the environment in which they are used and which they help to constitute. In this view, grammar and text are inseparable elements of the context.

In this vast interactionist field, different approaches have contributed to the understanding of the role that textual and linguistic resources play in interaction. In linguistic pragmatics, studies on politeness and impoliteness reveal that several linguistic phenomena, such as linguistic hedges, specific verb tenses and syntactic constructions, find explanation and motivation in face (positive and negative) negotiations (Brown; Levinson, 1987; Culpeper, 2011; 2016). In the field of Ethnomethodological Conversation Analysis and Interactional Linguistics, it has been observed that syntax plays an essential role in carrying out fundamental actions in interaction, such as projecting the end of the turn or the relevant place of transition, changing the topic of the conversation, constructing statements jointly, assuming positions of greater or lesser knowledge in relation to the interlocutor, etc. (Clayman, 2013; Ford; Fox; Thompson, 2002a; Ochs; Schegloff; Thompson, 1996; Thompson; Couper-Kuhlen, 2005). Within the scope of the textual-interactive organization group, part of Spoken Portuguese Grammar [Gramática do Português Falado], the analysis of interviews and university classes has demonstrated the interactional function of different resources, such as discursive markers, question and answer turns, the structuring of the utterance in theme and rheme, certain verbal tenses, textual formulation activities, etc. (Koch; Jubran, 2006).

Affiliated with this interactionist perspective, in recent years, we have studied the role of textual relations and their markers (connectors and syntactic constructions) in interaction. Moving away from a perspective that conceives them as simple cohesion mechanisms, we have sought to demonstrate, from a praxeological or actional conception of language, that these relations are procedures that interactants use to interact, and not simple resources that allow the speaker to compose a text that can be evaluated as coherent by the listener (Cunha, 2022a).

Focusing mainly on institutional interactions, such as electoral debates and media interviews, we have obtained evidence of the effective role that textual relations play in interaction, some of which are:

• The decision to establish a relation, such as reformulation, is sensitive to the sequential position of the turn. This decision may be motivated by the interlocutor’s actions in the previous turn and may impact the actions they will take in the following turn (Cunha, 2022a).

• Textual relations, such as causality, counter-argument and comment, allow the speaker to anticipate possible objections from the interlocutor regarding the offensive nature of their intervention to make sure the interlocutor will not interpret this intervention as an attack on their positive or negative face. (Cunha, 2020a; 2020b; 2022b).

• In conflictual interactions, e.g. debates, relations such as metadiscursive comment allow the speaker to maintain a tense line of conduct, through which they constantly assess the intensity of their attacks and defenses against the faces involved (Cunha; Braga, 2018; Cunha; Braga; Brito, 2019).

• Textual relations, such as preliminary, counter-argument and reformulation, allow the interlocutors to manage the epistemic dimension, carrying out actions as claiming epistemic authority on certain topics, ranking informations, endorsing positions of greater or lesser knowledge, etc. (Cunha, 2023a; 2023b).

Developing this perspective for the study of textual relations, we highlight their importance in the attribution of actions, that is, in the process through which interlocutors use different resources (verbal language, prosody, gestures, objects, etc.) to project and recognize actions, such as asking for information, asserting, inviting, threatening, congratulating, criticizing, etc. (Deppermann; Haugh, 2022; Levinson, 2013; Schegloff, 2007; Stivers; Rossi; Chalfoun, 2023). As we will discuss in the next item, this process of attributing actions is central to every interaction, as it is the process through which interlocutors negotiate (recognize, validate, contest, reconsider) “what is happening here?” (Goffman, 2012 [1974], p. 30) and evaluate the adequacy and inadequacy of the behaviors performed. As Stivers, Rossi and Chalfoun observe (2023, p. 1555), “successful action ascription can make the difference between strengthening and straining a relationship, playing a crucial role in the production of solidarity, conflict, and authority”.

In the literature on the subject, the fundamental role of verbal language in the action ascription is recognized. (Deppermann; Haugh, 2022; Heritage, 2012; 2013b; Levinson, 2013; Stivers; Rossi; Chalfoun, 2023). However, little attention has been paid to the role that textual relations can play in this process. Even interactionist approaches to the study of textual relations, such as Conversational (or Discourse) Pragmatics (Roulet, 2002; 2006; Roulet et al., 1985), did not study the role of these relations in the projection and recognition of actions. This paper aims to contribute to the understanding of this issue, investigating whether the articulation of one utterance to another within a turn contributes to the projection and recognition of actions.

To develop this study, we articulate theoretical and methodological contributions from two approaches: Conversational Pragmatics (Roulet, 2002; 2006; Roulet; Filliettaz; Grobet, 2001; Roulet et al., 1985), on the definition of textual relation and the role of illocutionary markers in the attribution of actions, and ethnomethodological Conversation Analysis, with regard to the formation and attribution of actions (Heritage, 2012; 2013b; Levinson, 2013; Schegloff, 2007; Stivers; Rossi; Chalfoun, 2023). These contributions will be presented in the next section. Empirical evidence of the role of these relations in the action ascription was collected in two Brazilian presidential election debates. One of them took place in 2018, in the first round of the campaign for the presidency of the Republic, while the second debate took place in 2022, in the second round of the campaign for the presidency of the Republic. More details about the data and the analysis steps will be provided in the section on methodological procedures. After considerations of a methodological nature, we will present the analysis of the role of textual relations in the action ascription by participants in these debates.

Attributing actions: projection and recognition of actions in interaction

In research on language in use and interaction, a central issue is the definition of the action or actions that interactants carry out through language (Thompson; Couper-Kuhlen, 2005). Even though they adopt different approaches, solutions, and methods; pragmatic, textual/discursive or conversational approaches share an actional or praxeological vision of language for which saying is doing (Austin, 1962). From this perspective, any attempt to establish a fixed and one-to-one relationship between form and meaning is abandoned, and all resources that interlocutors use in the complex task of intercomprehension are admitted as objects of study. These resources range from phenomena traditionally studied by semantics, such as lexical ambiguities and presupposition, to the genre of the interaction or the social roles endorsed (Kerbrat-Orecchioni, 1992; MARCUSCHI, 2007). In this sense, the definition of the action or actions that interactants perform through language, a basic condition for intercomprehension to occur, is a common problem to approaches that contemplate the indexical or contextual nature of language (Deppermann; Haugh, 2022; Labov; Fanshel, 1977; Roulet; Filliettaz; Grobet, 2001).

In this vast set of approaches, the approaches, solutions and methods are, as explained, distinct and numerous. Therefore, it is not our intention to list them. We are interested, more specifically, in mentioning those whose propositions about projection and recognition of actions can be useful in studying the role of textual relations in the attribution of actions: on the one hand, Conversational Pragmatics (Roulet, 2002; 2006; Roulet; Filliettaz; Grobet, 2001; Roulet et al., 1985) and, on the other hand, the Ethnomethodological Conversation Analysis (Heritage, 2012; 2013b; Levinson, 2013; Schegloff, 2007; Stivers; Rossi; Chalfoun, 2023).

In the 70s and 80s, different linguistic approaches analyzed the use of speech acts in effective interactions (or not created for theoretical purposes), based on Austin and Searle’s theory of speech acts (cf. Searle et al., 1992). As one of these approaches, Conversational Pragmatics, proposed by Roulet in the 1980s, developed a consistent and original articulation between this theory, the theory of implicatures (Grice, 1975), the integrated pragmatics (Anscombre; Ducrot, 1983), the theory of politeness (Brown; Levinson, 1983) and Conversation Analysis (Sacks; Schegloff; Jefferson, 1974).

In this approach, two types of textual relations (illocutionary and interactive) are proposed. Illocutionary relations allow analyzing the function of each intervention that constitutes the exchange and these relations can be initiative or reactive, depending on the place occupied by the intervention in the exchange. The approach proposes three generic categories of initiative illocutionary relations (question, request and information) and two generic categories of reactive illocutionary relations (response and ratification) (Cunha, 2021a; 2021b; Roulet, 2002; 2006). Unlike the theory of speech acts, and based on Conversation Analysis, this perspective considers that illocutionary value does not characterize isolated acts, but rather the function of an intervention in relation to the other interventions constituting the Exchange (Cunha, 2020a). Thus, it is the sequential position of the intervention that helps in attributing the action. For example, a sentence can be defined as an answer because it is uttered after a question.

Illocutionary relations can be specified by discourse markers, such as syntactic constructions, adverbial expressions and performative verbs. Thus, if an intervention is linked to the next one by the question relation, the description of the markers allows us to identify, more precisely, whether this relation is one of invitation, threat, advice, etc. Based on Grice (1975), Roulet (1980) proposes three types of illocutionary act markers, which correspond to three modes of communication of illocutionary value:¹

• denominative markers of illocutionary act (performative verbs: “I declare the session open”) – the act is communicated explicitly;

• indicative markers of illocutionary act (adverbial phrases, interjections, etc.: “Close the door, please”) – the act is communicated by conventional implicature;

• potential markers of illocutionary act (in particular, modal verbs: must, can, want, believe, etc.: “Can you close the door?”) – the act is communicated by generalized conversational implicature.²

Roulet also proposes a fourth type of marker, illocutionary orientation markers, which correspond to the three fundamental syntactic constructions (imperative, declarative, interrogative).³

In turn, interactive relations allow us to analyze the function of the constituents (acts and interventions) that form the intervention. More precisely, they define the pragmatic links (arguing, counter-arguing, commenting, etc.) that interlocutors establish between a portion of the text (act or intervention) and information expressed in previous portions of the text or even inferences (Roulet, 2022; 2006). In Table 1, we present the categories of relations proposed by Conversational Pragmatics and the markers (connectors and syntactic structures) that typically signal them in Portuguese.

Unlike most categorizations of relations (cf. Mann; Thompson, 1988), the definition of these categories of interactive textual relations is based on the study of oral interactions (Roulet et al., 1985). According to Roulet (2002, p. 149), “These categories [of interactive relations] are also defined by ‘cognitive traits of an interactionist nature’, such as preparing, narrating, supporting/sustaining, completing and reformulating/revising”. It is, therefore, an interactionist approach to the study of these relations.

As stated, Conversational Pragmatics, briefly presented, aims to analyze the effective use of textual and linguistic resources in real situations of language use, systematizing and inventorying the resources (markers of illocutionary and interactive relations) that allow scholars to define the functions that, in exchange, one intervention exerts in relation to another (illocutionary relations) and that, in that intervention, one act exerts in relation to another (interactive relations). Strongly focused on the description of the negotiation process developed by the interactants, this approach thus aims to provide categories for the scholar to analyze how, in context, the interactants participate in this process.

In the scope of the ethnomethodological Conversation Analysis, it is important to analyze the interactants’ point of view on the development of interaction, as well as the methods they use to interact, and this corresponds to the assumption of a radically emic stance (Heritage, 1984; Pike, 1967). Therefore, this perspective is interested in the formation and attribution of action because designing and recognizing actions are constitutive procedures of interaction, as a social action, and perceptible in the reaction/interpretation of each interactant to the previously produced turn. Thus, this approach is interesting in the way in which each interactant, when (re)acting to the actions of other interactants, reflexively expresses the recognition of these actions (Heritage, 1984; Drew, 2005; Schegloff, 2007).

The problem of action formation or action recognition/ascription is defined in these terms by Schegloff (2007, p. XIV),

The action ascription by interactants consists, therefore, not in a precise and static identification of the action performed, but in a “fallible, negotiated and even potentially ineffable” process (Levinson, 2013, p. 104) through which interactants negotiate this ascription. It is, therefore, an “interactional achievement” (Stivers; Rossi; Chalfoun, 2023, p. 1555). In analyzing this process, a set of resources has been considered by scholars as methods used by interlocutors to design and recognize an action (Deppermann; Haugh, 2022; Heritage, 2012; 2013b; Levinson, 2013; Stivers; Rossi; Chalfoun, 2023). For instance:

• the selection of linguistic (lexical and syntactic) units for the composition of turns;

• multimodal aspects such as prosody, but also gestures, looks, use of objects, etc.;

• the sequential position of the turn and the organization of preferences;

• institutional aspects, such as the type (genre) of interaction and social roles assumed;

• the negotiation of rights and duties over (epistemic) domains of knowledge.

In this sense, the recognition of action by interactants is based on a series of factors that encompass not only the linguistic units of turn composition and their sequential position, but also epistemic elements (how much knowledge about the subject do the interlocutors have?), proxemic (how do interactants behave in the environment and what “scenic” resources do they have access to?), institutional (what social roles do interlocutors play in the interaction?), etc. Due to this multiplicity of factors, in a medical or psychiatric consultation, a declarative sentence produced by the professional may be understood by the patient not as an assertion, but as a request for information (Heritage, 2012; 2013b; Labov; Fanshel, 1977) or, in a media interview, the interrogative sentence produced by the journalist may be understood by a politician not as a request for information, but as a criticism or a provocation (Clayman; Heritage, 2022; Cunha, 2023a; Heritage, 2002; Heritage; Roth, 1995).

In this paper, we argue that one of the resources that interlocutors use to project and recognize actions are interactive textual relations, such as argument, preliminary, reformulation, comment, etc. (table 1). Although these relations have been extensively studied in Conversational Pragmatics (Roulet, 2002; 2006; Roulet et al., 1995; Roulet; Filliettaz; Grobet, 2001), nothing has been said about their possible role in the projection and recognition of actions such as criticizing, inviting, praising, ordering, advising, etc., maybe because of understanding that only the illocutionary markers and the position of the intervention in the exchange would have this role. And despite the interest of Conversation Analysis in the procedures of action formation/ascription, this approach paid little attention to the role, in the process of action formation/ascription, played by phenomena similar to textual relations, such as explanations in dispreferred reactions (Heritage, 1984), the use of adverbial sentences (FORD, 1993), the joint construction of turns formed by complex sentences (the compound TCU format) (Lerner, 1996), clause combining with because (Couper-Kuhlen, 1996), the increments of turns with extensions of a prepositional or adverbial nature (Decat, 2011; Ford; Fox; Thompson, 2002b) or the composition of multi-unit turns formed by preface and question (Heritage, 2002; Heritage; Roth, 1995).

Combining propositions from both perspectives, our hypothesis is that, just as a turn can perform different actions, depending on the position it occupies in the sequence (Levinson, 2013), a sentence can perform different actions, depending on the units of the turn to which it articulates. This is what happens in excerpt (1), which was taken from one of the debates in our corpus and which will be analyzed in more detail in the analysis section.

Considered in itself, the sentence “por que que você pagou tão pouco aos beneficiários do bolsa família/” [why did you pay so little to the beneficiaries of the Bolsa Família /] could be interpreted by the opponent as a question (a request for information), due to the interrogative construction and rising intonation. However, linked to “já que você se julga o pai dos pobres e a tua economia +tão+ bem/” [Since you think you are the father of the poor and your economy is so good], the same sentence can be interpreted as a provocation and an ironic reaction to the statements made by Lula in the previous turn, not reproduced here, on the increase in the minimum salary during his administration, when President of the Republic (2003-2010). And this is the opponent’s interpretation, as revealed by his reaction in the second turn, when he disqualifies the previously produced turn.

This example clearly shows that interactive textual relations, far from constituting an abstract list of theoretically and artificially constructed categories (ethical perspective), are procedures that interlocutors effectively use to project and recognize actions and have consequences for the development of interaction. After all, establishing the relation signaled by the connective expression “já que”, the producer of the first round signals that he wants to provoke the opponent, who reveals, in his reaction, that he understood this intention. In this sense, the textual relation acts directly in the negotiated and emerging process of attribution of action. Before studying the attribution of actions in the selected debates, we present information about the procedures we adopted in collecting the corpus and analyzing it.

Methodological procedures

The corpus of this research is composed of two Brazilian presidential electoral debates. One of them took place on 10/04/2018, in the first round of the campaign for the presidency of the Republic, and was the last debate of that campaign, because, in the second round, there were no debates. The following candidates participated in this debate: Alvaro Dias (Podemos), Ciro Gomes (PDT), Fernando Haddad (PT), Geraldo Alckmin (PSDB), Guilherme Boulos (PSOL), Henrique Meirelles (MDB) and Marina Silva (Rede). This debate lasted 151 minutes and 33 seconds.⁴ The second selected debate took place on 10/28/2022, in the second round of the campaign for the presidency of the Republic, and was also the last debate of that campaign. The candidates who participated were Luiz Inácio Lula da Silva (PT) and Jair Bolsonaro (PL). This debate lasted 118 minutes and 32 seconds. The selected debates were promoted by the same television channel, Rede Globo, and in both the same journalist, William Bonner, played the role of mediator. In the 2018 debate, there was an audience in the studio. Formed by the two debates, the corpus of this research has a total duration of 270 minutes and 5 seconds (or 4 hours, 30 minutes and 5 seconds).

Once the corpus was selected, we followed some steps in the analysis. First, we reviewed the transcript available on the G1 website, for the 2018 debate, and on the UOL website, for the 2022 debate. This review, based on viewing the complete debates, aimed to obtain a transcription that was as close as possible to the interaction actually developed by the participants in the debates.⁵

After reviewing the transcription, the other steps consisted of analyzing the role of textual relations in the action attribution by debate participants. In the analysis, we used the categories of interactive textual relations proposed by Roulet (2002; 2006, Roulet; Filliettaz; Grobet, 2001) (Table 1).

To verify the role of textual relations in the action ascription, we analyzed the relations that the participants (candidates and mediator) established in each turn. In the 2018 debate, there are 241 turns; in the 2022 debate, there are 206 turns.⁶ In the analysis of each turn, we considered the relations that occurred both at the microlinguistic level, in which the terms of the relation are the information expressed by two acts (for example, a main sentence and an appositive or an adverbial phrase); as well as at the macrolinguistic level, in which the terms of the relations are the information expressed by an act and an intervention (unit formed by a set of acts) or by two interventions (for example, a preparatory segment formed by several assertions and a question).

In this analysis, due to the objectives of this research, we were only interested in the relations that participated in the action projection by the producer of the turn and in the action recognition by the next speaker. Therefore, while relations such as those illustrated in excerpt (1) of the previous section and in excerpt (2) below were considered, relations in which this participation, in our view, does not occur, as in excerpt (3), were not considered in this research.

In each excerpt, there are two acts linked by a relation of condition or potential argument (Roulet, 2006) signaled by “se” [if]. However, in excerpt (2) the act introduced by the connector (“se você parte da turma do temer do partido do temer/” [if you are part of the Temer’s group from the Temer’s party]) can lead the opponent to identify the interrogative sentence not as a question (request for information), but as a challenge or a provocation. Otherwise, in excerpt (3), the act introduced by “if” (“se esse pedido [de direito de resposta] for considerado procedente” [If this request [for the right to reply] is considered valid]) does not affect the instructional nature of the main sentence (“o candidato ofendido vai ter um minuto pra se defender\” [the offended candidate will have a minute to defend himself]), a nature that remains even with the suppression of the condition.

To validate or, at least, better support our analysis regarding the effective role of a textual relation in the action ascription, we used two parameters. The first of them – and the most important – was the interlocutor’s reaction. Thus, most of the relations we analyzed occur in turns followed by a reaction with which the interlocutor demonstrates his recognition of the action projected in the previous turn (or in part of it). As an example, immediately after Boulos’s question/provocation (excerpt 2), Meirelles’s response (excerpt 4) reveals his recognition that Boulos did not ask him for information, but, instead, challenged him.

However, in some instances, the speaker establishes the textual relation at the end of a sequence (for example, at the end of a part of the debate) and therefore does not receive a reaction from the interlocutor. In other cases, although the interlocutor takes a turn, he does not react to the projected action with the relation and addresses another topic. In these cases, we used the second parameter, which was the suppression of one of the terms of the relation, in order to verify, without the suppressed term, whether the action is modified or maintained. This is what we show in excerpt (5).

In the italicized excerpt from the turn produced by Ciro Gomes (l. 05-07), the act “e é bom que a gente lembre que pode ser a nossa filha a nossa mãe\” [and it’s good that we remember that these women could be our daughter and our mother] is linked to the act “sessenta mil mulheres brasileiras foram estupradas\” [sixty thousand Brazilian women were raped] as a comment. The comment “e é bom que a gente lembre que pode ser a nossa filha a nossa mãe\” allows us to analyze “sessenta mil mulheres brasileiras foram estupradas\” not as a simple assertion, but as an alert. Without the comment, the act “sessenta mil mulheres brasileiras foram estupradas\” could be interpreted as an assertion. However, Meirelles, in the next turn, deals with public security in general and does not refer to the topic of rape against women. In cases like this, we therefore used the suppression criterion.

Applying these criteria, we identified in the corpus a total of 73 textual relations that effectively participate in the action projection and recognition: 39 in the 2018 debate and 34 in the 2022 debate. Below, we present a more detailed analysis of the occurrences we identified. As it is not possible to analyze all occurrences, we will address the actions in whose projection and recognition textual relations played a central role.

Action ascription in the electoral debate

In both debates, the establishment of textual relations is a procedure with which the interlocutors project and recognize the following actions: challenge (provoke), criticize, complain, accuse, threaten, promise, self-praise, denounce, alert. The most recurring actions were challenging, criticizing and promising, actions that will be analyzed below.

Making challenges (provocations)

In the analyzed debates, it is common for a candidate to produce interrogative sentences not to ask (request information), but to challenge (provoke) the opponent.⁷ In the occurrences identified in the corpus, the recognition by the opponent (and by the analyst) that the interrogative sentence constitutes a challenge is rarely based on the sentence itself, but mainly on the textual relations that the speaker establishes between the act that the sentence encodes and other acts. The preliminary relation is the one most recurrently used to attribute the action of challenging,⁸ as in excerpt (6).

In this excerpt, the interrogative sentence produced by Boulos “eu queria saber pra você/ saneamento é um negócio ou é um direito/” [I wanted to know if for you sanitation is a business or a right?] (l. 06-07) does not contain elements, such as illocutionary markers, that allow this sentence to be interpreted as a challenge. However, the segment that precedes the sentence, especially the segment “vocês têm uma receita que é a da privatização” [you have a practice that is privatization], associates the use of private resources with the idea of “business”, and this constitutes a problem from the perspective of the candidate, affiliated with a left-wing party (PSOL), and his electorate. The opponent’s recognition that the sentence, due to its preliminary, constitutes a challenge is revealed in his choice to avoid the term “privatization” and in his delay in mentioning private investment. This mention occurs only after he criticizes the predominance of state-owned companies in the provision of basic sanitation (l. 16-18) and at the end of the turn, when he says “você pode trazer investimento privado eh pra poder investir mais” [you can bring private investment to be able to invest more] (l. 25-26). In excerpt (7), the preliminary relation also attributes the action of challenging to an interrogative sentence.

In the first turn, the sentence “por que que cê não concluiu a transposição do São Francisco/” [why didn’t you complete the transposition of (the river) São Francisco?] (l. 07-08) also does not contain elements that allow the opponent to understand this sentence as a challenge or a provocation. This understanding, which is visible in Lula’s response, occurs due to the preliminary, in which the opponent accuses him of stealing (“você levou foi gra:na pro teu bolso” [you took money into your pocket], l. 04-05). In this turn, the metadiscursive order, “responde pra gente aqui” [answer us] (l. 08), which follows the interrogative sentence as a comment, also contributes to the recognition that the action that the speaker performs in the sentence is to challenge or provoke the opponent. And it is worth highlighting that this comment aggravates the provocation, because, although in it the previous action is referred to as a question and not as a provocation, the speaker, with this comment, endorses a deontic stance of greater power and authority over the opponent (Couper-Kuhlen; Thompson, 2022; Stevanovic, 2018) and can thus impose on the opponent the duty to respond, a duty already implied by the interrogative sentence of the previous act (Roulet, 1980).

There are cases in which, in addition to the preliminary relation, another, more localized, relation contributes to attributing the action of challenging to the interrogative sentence, as shown in excerpt (8), briefly analyzed in the previous section.

In Boulos’s speech, the sentence “como o povo pode acreditar que você vai combater a corrupção” [how can the people believe that you will fight corruption] (l. 06-07) could be identified as a challenge, due to the construction “how can the people believe that you [...]”, This identification, in our opinion, would disappear if the sentence were limited to “how will you fight corruption”.⁹ However, this identification is favored by the relations of preliminary and potential argument (condition) with which Boulos expands the turn. With the preliminary, the candidate precedes the sentence with accusations against Michel Temer; with the potential argument, the candidate adds a conditional sentence whose presupposition is “you are part of Temer’s group, of Temer’s party [PMDB]”. With both relations, Boulos associates Michel Temer with corrupt practices (preparatory segment) and the interlocutor, Henrique Meirelles, with Michel Temer’s party (conditional segment). The challenge is therefore based on the causal reasoning “if x, then y”. According to this reasoning, if Meirelles belongs to Temer’s party, he therefore cannot fight corruption.

Next, the opponent shows that he understood Boulos’s interrogative sentence as a challenge and not a question. After all, the beginning of his reaction is composed, at the same time, of a defense (“eu sou um candidato que faço parte da minha história\ e a minha história é um com- é uma história de quem trabalha em primeiro lugar\” [I am a candidate who is part of my story, and my story is a story of those who work in the first place], l. 09-11) and of an attack (“eu sei que pode parecer estranho pra você. essa história de trabalhar” [I know this story of working may seem strange to you], l 11-12).

A similar procedure occurs in excerpt (9), in which Dias prepares the interrogative sentence—the challenge—with criticisms against the opponent and establishes a more local argument relation between the sentence and another act.

In the turn produced by Dias, the segment “com +tantos+ escândalos de corrupção no governo do pt o senhor diz que vai combater a corrupção sendo presidente/” [With so many corruption scandals in the PT government, do you say that you will fight corruption as president?] (l. 11-13) is prepared by assertions that allow to identify the sentence “o senhor diz que vai combater a corrupção sendo presidente/” not as a question, but as a challenge. However, in addition to this preliminary relation, the segment in which the challenge occurs is formed by the adverbial phrase “com +tantos+ escândalos de corrupção no governo do pt” and by the interrogative sentence (polar question) “o senhor diz que vai combater a corrupção sendo presidente/”. Between these segments, the speaker establishes a relation of argument, the first act being paraphrased as “since there were so many corruption scandals in the PT government”. This relation also contributes to identifying the interrogative sentence as a challenge, due to the threat that the adverbial phrase represents to the opponent’s face (Brown; Levinson, 1987). The opponent identifies the interrogative sentence as a challenge and, when taking the floor, refutes the information expressed by Dias and does not offer the negative answer, preferred after a polar question (yes/no) (Clayman; Loeb, 2018).

As in the two previous excerpts, it is common for the speaker to establish an argument relation between an interrogative sentence and another act, attributing the action of challenging to the sentence. In excerpts (10) and (11), the same candidate uses this resource, and in both cases the opponent identifies the sentence as a provocation.

In the first turn of excerpt (10), already briefly analyzed, two acts are linked by an argument relation signaled by the connector “já que” [since]. Because the dispute over epistemic domains is a characteristic of the debate (Cunha, 2023b), the sentence “por que que você pagava tão pouco pro bolsa família/” [why did you pay so little to Bolsa Família?] (l. 02-03) can be identified by the interlocutor not as a sincere request for information, despite the interrogative construction, but as a challenge or provocation. In this case, the type of activity would have ascendancy over the grammatical type of the sentence in the action ascription (Stivers; Rossi; Chalfoun, 2023). However, the statement “já que você se julga o pai dos pobres e a tua economia tão bem/” [Since you think you are the father of the poors and your economy is so good] also contributes to the identification of the interrogative sentence as a provocation. This identification is made by the opponent, who, in his reaction, disqualifies the previously produced turn.

Likewise, in excerpt (11), produced in the same debate part, some time after excerpt (10), the candidate repeats the challenge, claiming that the opponent would not have answered the question asked in excerpt (10) (“você não respondeu\” [you didn’t answer]). This time, the act that functions as an argument, introduced by the connector “se” [if], expresses a doubt about the economic performance of the opponent’s government, being paraphrased as “if it is true that your government was so good in the economic part” (cf. Cunha, 2010). Once the turn has been taken, the interlocutor attributes to the interrogative sentence the action of challenging or provoking, as revealed by his response, which, through a suggestion (“I suggest that X”) to the direction of the broadcaster responsible for the debate, disqualifies the opponent’s behavior.

Making promises

Especially in the 2018 debate, candidates repeatedly performed the procedure of presenting problems and then informing that (and how) these problems would be resolved if the candidate were elected. With this procedure, the candidate establishes an argument relation through which initial statements can be reinterpreted not as mere assertions about problems, but as campaign promises or commitments. In general, this relation is signaled by some argumentative connector, such as “portanto” [therefore] and “então” [then], which signals that the introduced information is a conclusion derived from the previous segment, the assertions. This is what happens in excerpt (12).

In the first turn, the declarative sentence “isso impede o governo de sobretaxar a terra improdutiva ou mesmo de desapropriar a partir da concessão de títulos da dívida agrária\” [this prevents the government from surtaxing unproductive land or even expropriating it, granting agrarian debt bonds] (l. 04-06) could be identified as an assertion. However, followed by “portanto nós vamos ter que enfrentar esse desafio\” [therefore we will have to face this challenge] (l. 07), this sentence can be reinterpreted as a promise or a commitment from the candidate. According to this commitment, if elected, the candidate will surtax and expropriate unproductive land. This seems to be the interpretation of the interlocutor, who, when taking the floor, criticizes governments from the opponent’s party, suggesting that promises may not be fulfilled.

The same procedure of reinterpreting assertions as promises occurs in excerpt (13). In the first turn, all occurrences of the connector “então” [then], in bold, signal argument relations through which the speaker leads the interlocutor to understand the assertions that precede each occurrence of “então” as promises or campaign commitments.

Let’s analyze in more detail the second occurrence of the connector “then” (l. 08). With this connector, the speaker articulates the following segments:

1. hoje como nós sabemos bem muitas vezes a polícia militar não troca informação com a polícia civil que não troca informações com a polícia federal e um estado não troca a informação com outro\ [today, as we know, the military police do not exchange information with the civil police, which do not exchange information with the federal police, and one state does not exchange information with another.]

2. então nós temos que ter antes de mais nada um sistema de informação unificada controlado e administrado pela polícia federal\ [then we need a single information system controlled and managed by the federal police.]

In the first segment, the candidate presents information that, in his opinion, represents a problem related to public safety. In the second, he makes a proposal (the creation of a “single information system”) that will solve the problem presented in the first segment. The opponent’s response, ironically accusing the candidate of copying his campaign proposals, shows that, for him, the information preceding the connectors are not assertions, but rather campaign promises copied from his own government program.

Although rare, the same procedure also occurred in the 2022 debate. Excerpt (14) is the only occurrence of this procedure in this debate.

In the first turn, the segment that goes from “na casa civil tinha uma sala de prefeito” [in the civil house there was a room to receive the mayors] to “era só apresentar o projeto a gente atendia\” [they just had to present the project and we would grant it] (l. 06-10) can be interpreted as a set of assertions with which Lula compares the treatment he gave to mayors and the treatment they receive from Bolsonaro. However, the excerpt “e isso nós vamos voltar a fazer com governadores e com prefeitos” [and we will do this again with governors and with mayors] (l. 10-11) allows to reinterpret the way he treated mayors (l. 06-10) as promises. Then, the opponent addresses not the opponent, but a part of the voters (“você de cidade pequena em especial” [you from a small town in particular]) to implicitly attack the promises, by suggesting that he gave better treatment to mayors than Lula, when he was president.

Making criticisms

In both debates, the establishment of textual relations is a procedure that candidates use repeatedly to project and recognize criticism. In this case, acts that, without the relations, could be identified as assertions, advice, explanation or orders, are instead identified as criticisms, because of these relations. This is what happens in excerpt (15).

In this excerpt, the speaker constructs the first turn based on the construction “If I were you, I would ask X”, which can be understood as an advice indicative marker (Cunha, 2021b) or a grammatical format with which the speaker asks the interlocutor to act in their benefit (Couper-Kuhlen; Thompson, 2022). However, the act “que você deve ter acabado” [which you probably closed down] (l. 02), linked to the term Ministry of Planning as a comment, and the act “pra você pelo menos copiar um pouco e aprender o que é investimento em infraestrutura\” [so you can at least copy a little and learn how to invest in infrastructure] (l .03-05), linked to the previous act as an argument (purpose), allow the opponent to identify the first turn not as advice, but as criticism. In the opponent’s reaction, the use of the marker “bem” [well], signaling the realization of a dispreferred turn (Heritage, 2015; Risso, 2006), and the mention of a subject not directly related to the topic of the first turn reveal that he did not identify this turn as an advice.

In the 2018 debate, the procedure of expanding the turn with comments and arguments to project a criticism, which would be projected differently if not for the relations established, was also common. Excerpts (16) and (17) were taken from this debate.

In excerpt (16), there is a pre-sequence in which a candidate chooses who to address the question to. In this pre-sequence, the interlocutors are the mediator (William Bonner) and two candidates (Marina Silva and Fernando Haddad). Then, Marina Silva precedes the question (“e a pergunta é. nós temos um país/” [and the question is: we have a country], l. 09) with a metadiscursive explanation about the opponent to whom she would like to ask the question (candidate Jair Bolsonaro), if he was present in the debate. In the preparatory segment, the candidate articulates the acts “eu iria fazer essa pergunta também para o candidato bolsonaro” [I would also ask this question to the candidate Bolsonaro] (l. 07) and “que mais uma vez amarelou” [who was once again got scared] (l. 08) through a comment relation. Then, the candidate is interrupted by applause from the audience, who, due to the segment “que mais uma vez amarelou”, analyzed the entire segment not as an explanation, but as a criticism of the absent candidate. At this point, the mediator asks the audience not to speak and asks the candidate to rephrase the question.

When reformulating the question (l. 16-18), the candidate starts the turn by explaining again who she would address the question to. In the excerpt, she establishes a textual relation of argument, a relation signaled by the connector “como” [since] (“como ele mais uma vez amarelou” [since he was once again got scared], l. 17), which also allows interactants (opponents and audience members) to understand the sentence “deu uma entrevista para a Record e não veio aqui debater conosco” [he gave an interview to Record and didn’t come here to debate with us] (l. 17-18) not as an explanation, but as a criticism.

Adopting a similar procedure, Gomes, in the first turn of excerpt (17), establishes a set of textual relations that allow identifying both the declarative sentences and the final interrogative sentence as criticisms addressed to the absent candidate (Jair Bolsonaro) and members of his team, instead of his interlocutor (Henrique Meirelles).

In the first turn, the appositive “rude as he is” [tosco como é] serves as a comment on the phrase “o general mourão”, then Bolsonaro’s vice-president, while the act “sem ter as habilidades políticas de um mentiroso que fala a mesma coisa” [without the political skills of a liar who says the same thing] is set as a relation of argument (paraphrased as “because he does not have the political skills of a liar”) for the entire previous act. These relations project the sentences about Mourão not as assertions, but as criticisms against him.

Still in this turn, after an assertion about a fight between members of Bolsonaro’s team, the speaker produces the interrogative sentence (polar question) “cê imagina que isso vai dar certo no brasil/” [do you believe this will be good for Brazil?] (l. 06-07). The informations expressed before this sentence serves as a preparation for it. With these informations, the negative answer is the preferred reaction to the polar question (Clayman; Loeb, 2018). Therefore, it is difficult to identify the interrogative sentence as a request for information. By giving the expected answer (“não não vai dar certo\. a +eleição+\” [the election will not be good], l. 09), the opponent affiliates himself with the speaker (Ostermann; Andrade; Frezza, 2016; Stivers; Rossi; Chalfoun, 2023) and thus reveals that he has understood the previous turn as a criticism (or set of criticisms) against the absent opponent and not as a simple request for information.

Final considerations

In this paper, we reveal the importance of textual relations (preliminary, argument, comment, etc.) in the action ascription (projection and recognition) in two presidential electoral debates. We show that the articulation of varied textual portions inside a turn allows the action ascription, as well as microlinguistic resources (illocutionary markers), the prosody, the epistemic dimension, the type of interaction, etc.

The analyses of the two debates revealed that, in fact, the establishment of textual relations is a procedure that allows interlocutors to project and recognize the following actions: challenge (provoke), criticize, complain, accuse, threaten, promise, self-praise, denounce, alert. In this study, we presented in more detail how the process occurs in the projection and recognition of the actions of challenging (provoking), criticizing and promising (making campaign commitments). As a result, we realize that textual relations, expanding the turn, allow us to interpret some acts as challenges (provocations), criticisms, accusations, threats, promises, etc. Without these relations, these same acts could be interpreted as assertions, requests for information and advice.

Having confirmed the phenomenon of action projection and recognition through textual relations in electoral debates, the following question arises: why does a candidate opt for the complex procedure of projecting an action by articulating textual constituents, if this procedure involves a greater expenditure of time (a precious commodity where speaking time is scarce) and the possibility of being misunderstood by the opponent and, worse, by the electorate? As in the debate most of the actions are threatening to the faces of the interactants (Cunha; Braga; Brito, 2019), carrying out procedures that make these actions deniable and less direct is an advantage (Roulet, 1980; Stivers; Rossi; Chalfoun, 2023). This is what the procedure studied in this work allows candidates to do. After all, a candidate can always claim that an act, despite the acts to which it is linked, is just information or a question and not a threat or accusation.¹⁰

We believe that the focus of this paper on the role of textual relations can contribute both to pragmatic approaches, for which it is important to systematically investigate the verbal resources on which the action ascription is based, and to conversational approaches, for which it is important to identify the procedures effectively considered by interactants to project and recognize actions. Thus, with this study, we hope to contribute to studies that, inserted in the interactionist perspective of language studies, such as Conversational Pragmatics and Ethnomethodological Conversation Analysis, deal with the procedures that interlocutors carry out to negotiate, with support from language (verbal and non-verbal), the interaction in which they participate and the type of relation that associate them to one another. We believe that systematic studies of these procedures can offer safer support for a detailed analysis of how interlocutors attribute (project and recognize) actions and frame the interaction in which they participate.

Acknowledgments

The research presented in this paper was developed during the author’s residence period at the Institute for Transdisciplinary Advanced Studies (Instituto de Estudos Avançados Transdisciplinares - IEAT), at UFMG. We thank IEAT for the favorable conditions for carrying out this research, as well as CNPq for the research productivity grant (process no. 304805/2022-0). We would also like to thank Paloma Bernardino Braga for reviewing the translation of the paper into English. Paloma Bernardino Braga is a graduate, master and doctoral student at the Federal University of Minas Gerais (UFMG).

Appendix

2018 debate

Link to access the video:

https://www.youtube.com/watch?v=epDDSEVnLmI

Link to access the transcript (G1 news website):

https://g1.globo.com/politica/eleicoes/2018/noticia/2018/10/05/veja-a-integra-do-debate-na-globo.ghtml

2022 debate

Links to access videos for each part of the debate:

1º part:

https://www.youtube.com/watch?v=EK_hxsxWF4I

2º part:

https://www.youtube.com/watch?v=-woWv61-Urk&t=32s

3º part:

https://www.youtube.com/watch?v=MVeRuwkig18&t=35s

4º part:

https://www.youtube.com/watch?v=ay1QAn1rYjw

Lula’s final considerations:

https://www.youtube.com/watch?v=DunBALbcNmQ

Bolsonaro’s final considerations:

https://www.youtube.com/watch?v=0PjhouB28gI

Links to access the transcripts of each part of the debate (UOL news website):

https://noticias.uol.com.br/eleicoes/2022/10/28/integra-debate-na-globo-primeiro-bloco.htm

https://noticias.uol.com.br/eleicoes/2022/10/28/integra-debate-na-globo-segundo-bloco.htm

https://noticias.uol.com.br/eleicoes/2022/10/28/integra-debate-na-globo-terceira-bloco.htm

https://noticias.uol.com.br/eleicoes/2022/10/29/integra-debate-na-globo-quarto-bloco.htm

References

Publication Dates

History