Formulations, Politeness and Facework in Courtroom Interaction: A Study Case

This article aims to analyse how formulations, politeness, and facework are employed by speakers to achieve specific purposes in courtroom interaction. The theoretical framework is based on the concepts of conversation analysis and forensic linguistics, and the corpus consists of the trial of Jodi Ann Arias, which received widespread media attention in the United States. The excerpts under analysis were taken from the final day of the prosecutor’s cross-examination and from a reporter’s transcript of proceedings, both available on the Internet. It is important to highlight that these devices contribute to construct, enhance, threaten, or weaken people’s image involved in this kind of context.

"the sequence will then take the form of Q1-Q2-A2-A1, with the middle pair (Q2-A2) being called an insertion sequence". In this structure, the pair Q2A2 is inserted into the pair Q1A1, so the embedded pair in the sequence is an insertion sequence. The author affirms that "although that appears to be a question (Q2) in response to a question (Q1), the assumption is that once the second part (A2) of the insertion sequence is provided, the second part (A1) of the initial question (Q1) will follow". As Coulthard and Brazil (1992, p. 52) observe: In other words, during the inserted sequence the original question retains its transition relevance, and if the second speaker does not then produce an answer it is noticeably absent in exactly the same way as it would be if there were no intervening sequence, and the questioner can complain about the lack of an answer in exactly the same way. Thus the argument is that adjacency pairs are normative structures, the second part ought to occur, and for this reason the other sequences can be regarded as being inserted between the first pair part that has occurred and the second pair part that is anticipated.
Because the first question is immediately followed by another question, instead of the direct answer, it appears that the interlocutor wants to gain more time to delay the answer. We illustrate these cases as following: Q1: Do you want the early flight? Q2: What time does it arrive? A2: Nine forty-five. A1: Yeah -that's great. (Yule, 1996, p. 78) In the example above, the delay in the answer indicates that the first part of the pair does not necessarily receive the second part that the speaker expects. Yule (1996, p. 78) observes that "delay in response symbolically marks potential unavailability of the immediate (i.e., normally automatic) expected answer. Delay represents distance between what is expected and what is provided. Delay is always interpreted as meaningful".

Adjacency Pairs: Two-Part and Three-Part Structures
Adjacency pairs are the basis for the following principle: every move in a conversation is essentially a response to a previous one and an anticipation of what is to come. When speakers formulate their contributions, they often show their understanding of the previous speech and reveal their expectations about the next one. In line with Goffman (1981, p. 5): Whenever people talk there are very likely to be questions and answers. These utterances are realized at different points in "sequence time." Notwithstanding the content of their questions, questioners are oriented to what lies just ahead, and depend on what is to come; answerers are oriented to what has just been said, and look backward, not forward. Observe that although a question anticipates an answer, is designed to receive it, seems dependent on doing so, an answer seems even more dependent, making less sense alone than does the utterance that called it forth. Whatever answers do, they must do this with something already begun.
As the scholar points out: "in questions and answers we have one example, perhaps the canonical one, of what Harvey Sacks has called a 'first pair part' and a 'second pair part', that is, a couplet, a minimal dialogic unit". It is a round two utterances long, each one of the same type and temporally following directly on the other; in sum, an example of an "adjacency pair". The first pair-part sets up a "conditional relevance", on anything that occurs in the slot which follows. "Whatever comes to be said there will be inspected to see how it might serve as an answer, and if nothing is said, then the resulting silence will be taken as notable--a rejoinder in its own right, a silence to be heard" (Goffman, 1981, p. 6). Wooffitt (2005, p. 32) claims that an adjacency pair is a sequence of two utterances which are adjacent, produced by different speakers, ordered as a first part and second part and typed, so that a first part requires a particular second, or range of second parts (Heritage, 1984a, p. 246). An invitation, then, would be the first part of an invitation -response pair, a question the first part of a questionanswer pair, and a greeting the first part of a greeting-greeting pair.
Even though two-part adjacency pairs are very common, there is a lot of casual conversation that this set of categories does not cover, for instance: classroom discourse, interrogations, service encounters and many other kinds of everyday talk which have a third move in their exchanges 4 . These types of conversational routines offer "an opportunity for participants to check that they are agreed on the function of the previous pair, to comment on the exchange as it stands, to react to the response in the context of the initiation" (Sinclair, 1992, p. 85 Despite recognizing that a two-part structure, particularly the adjacency pair "question/answer", is very frequent in conversation, a potential three-part structure also occurs. The Birmingham Discourse Analysis model suggests a potential three-part structure rather than a two-part one. Coulthard and Brazil (1992, p. 66) argue that "all eliciting 4 Concerning the terms utterance and exchange, according to Coulthard and Brazil (1992, p. 2), "utterance was defined as everything said by one speaker before another began to speak, and exchange as two or more utterances". 5 (...) "the structure of exchanges was expressed in terms of three elements I(nitiation), R(esponse) and F(eedback)" (Coulthard & Brazil, 1992, p. 69). exchanges have the potential of a three-part structure, while accepting that a two-part realization may, and in the case of polar responses often does, occur". As observed in the general practitioner consultation below, "three-part exchanges are in fact by no means The discussion does not rely on the quantity of talk is made up of two-part structure or threepart structure. In accordance with Sinclair (1992, p. 85), "the problem is not going to be resolved by a majority vote-by counting up whether the greater quantity of talk is two-part or three-part in its exchange structure. We must seek an explanation of the variability of the exchange".

Formulations
The term formulation was first introduced by Garfinkel and Sacks (1970) as practices that interactants might use to show that they understand parts of interaction. These parts can be taken by them as opportunities to formulate the conversation, that is, to describe the conversation, "to explain it, or characterize it, or explicate, or translate, or summarize, or furnish the gist of it, or take note of its accordance with rules, or remark on its departure from rules" (Garfinkel & Sacks, 1970, p. 350). A very good example of this is "Pains-in your chest then" in: Doctor, I: And what's been the matter recently Patient, R: Well I've had pains around the heart Doctor, I: Pains-in your chest then Patient, R: Yes (Coulthard and Brazil 1992, p. 66) As Heritage and Watson (1979, p. 128-129) explain, "the uses of formulations are multiplex […] they may be used to address an immense variety of matters, these matters being, in their most specific terms, heavily embedded in the specific stretches of talk in which they occur".
According to them, formulations manifest three essential properties: "preservation, deletion and transformation" of a prior utterance or utterances. Based on these properties described by the above-mentioned authors, Holt and Johnson (2010, p. 28) highlight: (1) the preservation of some (selected) aspects of the sense and reference of the news materials delivered in the content of the formulating utterance (2) the transformation of the syntactic and semantic framework within which the news was originally delivered (3) the deletion of some aspect or aspects of the news delivered. Heritage and Watson (1977, p. 2-3) Deppermann (2011, p. 122) argues that "same-speaker formulations, which rework a prior version of reference or description, are used to explicate, clarify, specify, generalize, etc. meanings; other-speaker formulations are displays of understanding (and can clarify, explicate, etc. as well)". Nevertheless, in neither case formulations are a transparent device for making meanings clear. "Formulations have both an explicative and a transformative aspect": While each formulation preserves parts of the meaning established before, it deletes other parts and adds new aspects of meaning (Heritage and Watson, 1979, p. 129). In addition, there is a reflexive relationship between first and second versions: Formulations make meanings explicit which were implicit in the prior version (or which are at least treated as if they had been implicated in it), but their local meaning rests itself in part on the versions they are meant to replace. When studying semantic properties of formulations, it is important to attend both to the sequential organization and the precise linguistic design of the formulation. Deppermann (2011, p. 122) Drew (2003, p. 296) asserts that formulations "serve to perform specific interactional tasks which vary according to the setting". They are a generic mechanism in the interaction, however, the ways through which they are presented are related to the activities they manage in specific situations. These devices are therefore multifunctional as they play many different interactional roles in many different contexts.

Facework and Politeness
According to Goffman (1974, p. 9), "the term face is the positive social value a person effectively claims for himself by the line others assume he has taken during a particular contact". Politeness is a social phenomenon which aims at maintaining harmony in interaction to establish cordiality in relationships among participants. Schneider (1998, p. 51) states that politeness is usually expressed in grammatical structures by means of "pragmatic conventions recognized by a given society as having some intentional illocutionary force (e.g., please..., could you... etc.)". Goffman (1974, p. 9) postulates that "face is an image of self, delineated in terms of approved social attributes, therefore, it is an image that others may share, for example, when a person makes a good showing for his/her profession or religion by making a good showing for himself/herself". In the process of interaction, speakers usually tend to construct a positive image of themselves and their place in society by adopting two possible attitudes: a defensive attitude, in which they need to preserve their own face, and a protective attitude, in which they must preserve their interlocutor's face (Burgo & Neto, 2016, p. 885).
Every person lives in a world of social encounters that involves contacts with others. In each of these contacts, people tend to externalize their opinions, judgments and points of view that lead them to act out what is called a line, that is, "a pattern of verbal and nonverbal acts by which he expresses his view of the situation and through this his evaluation of the participants, especially himself" (Goffman, 1974, p. 9). The individual hence expresses an evaluation of participants and, above all, of himself, so that his line presents a view of self. In addition, this line has to be in accordance with the way he wants to be seen.
Regardless of whether an individual intends to take a line, he will think that he has effectively done so (cf. Goffman, 1974). "The other participants will assume that he has more or less wilfully taken a stand, so that if he is to deal with their response to him, he must take into consideration the impression they have possibly formed of him" (Goffman, 1974, p. 9).
An individual protects face when his line is consistent with that expected by the other participants; in other words, when he shows an image which is based on interactants' assumptions and judgments, approved by the line adopted by the speaker. Face is not housed within or on the surface of its owner, but it is widespread in the flow of events of the encounters, and it occurs when participants are willing to decode the judgments expressed in these events.
Facework is defined by the author as everything a person does so that his actions do not make anyone lose face, including his own. Unlike politeness, facework is composed of practices that aim to establish, enhance, threaten or weaken the interlocutor's image. Politeness, SciELO Preprints -This document is a preprint and its current status is available at: https://doi.org/10.1590/1678-460x202243363 however, concerns behaviours related to people's need to be approved and to ensure their autonomy. Originally based on the notion of face formulated by Goffman (1974), Brown and Levinson (1987) expanded the concept of politeness, pointing out two aspects of the socially constructed self-image: negative face and positive face. a) Negative face: "the want of every 'competent adult member' that his actions be unimpeded by others"; b) Positive face: "the want of every member that his wants be desirable to at least some others" (Brown & Levinson, 1987, p. 62).
During a social encounter, every speaker intends to maintain face so the interaction develops without big problems; the individual focuses both on preserving his/her own face and on respecting the other's. To preserve negative face, it is important to protect the private territory, intimacy. On the other hand, to preserve positive face, interlocutors need to be recognized and valued by others. This may cause a conflict, since their intention to preserve negative face, for example, can come up against the intention of others to preserve positive face. As argued by Belchí (1994, p. 332): [...] the defence of individual A's own territory comes into conflict with individual's B desire of starting relationships. Based on this principle, it is easy to come to a conclusion that any interaction is essentially confrontational, even though, fortunately, it does not occur in the same degree (our translation) 6 .
Interaction essentially establishes conflictual situations, so the indispensable condition for its harmonious development is that the participants, according to Belchí (1994, p. 343), "preserve interlocutor's face in order to maintain a balance between the desire to approach and the respect of other's intimacy" (our translation) 7 . Although sometimes it is not possible to avoid face-threatening acts, certain strategies can be used to minimize potential threats (Cruz, 2020, p. 42). In some of Rosulek's studies (2010, p. 227), "lawyers use the voice of the law or judge to mitigate their potentially face-threatening acts. [...] the lawyer uses the authority of the judge to tell the jurors that they cannot use the defendant's prior bad acts as evidence in this case". In the example: "As the Court has instructed you, you should not simply conclude the defendant's a bad person and therefore 6 [...] la defensa del propio territorio por parte del individuo A entra en conflicto con el deseo de entablar relaciones del individuo B. Partiendo de esta base es fácil llegar a la conclusión de que toda interacción es esencialmente conflictiva, aunque afortunadamente no lo es siempre en el mismo grado. 7 preserven la cara de su interlocutor manteniendo un equilibrio entre el deseo de acercamiento y el respeto a la intimidad del otro. should be convicted of this", the author states that "the judge has more authority to make requests and demands of the jurors". If the lawyer had made a bald-faced demand or accused the jurors, he "might have damaged his relationship with them. Thus, the lawyer uses the authority of the judge to instruct them, instead of using his own voice". Tkacuková (2010, p. 343) mentions an example that involves politeness strategies during a cross-examination.
According to her analysis: Several researchers (e.g., Cashion, 1985;Harris, 2003) have focused on the connection between politeness and power and have come to the same conclusion. Surprisingly, those participants who hold the most powerful roles tend to use politeness strategies abundantly to redress the face-threatening acts that they have to perform. But these participants can also choose not to use politeness strategies in case they need to explicitly show their power. According to Harris (2003, p. 33), it is negative politeness features that are especially relevant to institutional settings (i.e., those strategies that prevent imposing on the speaker or impeding them). The counsels in Cashion's study (1985, p. 13) were found to exhibit mostly negative politeness features whereas positive politeness strategies (strategies that attend to the speaker's needs to be liked) occurred only rarely. In his cross-examination, Rampton employs a negative politeness strategy 'correct me if I'm wrong' to avoid imposing his knowledge on the witness.
Politeness ensures the protection of both speaker and listener's face, whether performed by positive or by negative politeness strategies.

Methodology
The corpus of this research consists of Jodi Ann Arias's trial, which gained massive attention from the media. On May 8, 2013, she was convicted of first-degree murder of her exboyfriend Travis Victor Alexander in the United States. On June 4, 2008, Alexander was shot in the head, repeatedly stabbed, and had his throat cut. His body was discovered by friends in the shower of his home about five days later. They called 911 and when questioned by the police, they said Arias should be interrogated. On July 9, 2008, Arias was indicted on firstdegree murder charges in Alexander's death. She was arrested on July 15, 2008, and To select the corpus, we have gathered all the material available on the Internet regarding Arias's trial: we have watched all the videos of her trial available on YouTube, as well as the transcripts released to the media on the Internet. It is important to stress that according to the multi-platform news and information media company USA TODAY, during Arias's sentencing retrial, she testified without the presence of the media, and the Court of Appeals ordered her testimony transcripts to be released 8 .
This work follows an empirical-inductive approach, as the texts were obtained from concrete situations of usage and the analysis was carried out by examining all the occurrences for our qualitative interpretations. The data are presented in a format to illustrate how the language was spoken in real-life interactions which exhibited marks of local verbal planning (verbal planning and speaking occurred at the same time). By studying the cases, we have considered possible meanings and interpretations that led us to the theoretical framework employed in this research. With the aim of delimiting our corpus, we have chosen the types of formulations, politeness, and facework that could provide more interesting and noteworthy elements for analysis, that is, devices that were commonly used by speakers to achieve their purposes. Therefore, we have established the recurrence of those devices as a criterion for data selection. Based on that, we proceeded to the organization of the excerpts, which followed the chronological order of the events: the video containing the final day of the prosecutor's cross-examination (Day 25) in Maricopa County Superior Court in Phoenix, Arizona, on February 21, 2013 9 and a reporter's transcript of proceedings (testimony of Jodi Ann Arias) on November 3, 2014 10 . We highlight that the excerpts taken from the video were transcribed according to Leite (2017, p. 21-22).

Results and discussions
Our analysis was organized into eight excerpts that showed significant elements for discussion such as: the adjacency pair "question/answer", discourse markers, so-prefaced questions, paraphrases, positive politeness, solidarity politeness, negative politeness, deference politeness, properties of formulations (preservation, transformation, and deletion), tag questions, facework strategies, verbal planning markers, mitigation devices, and hedges.
In the excerpts, the sentences that will be analysed are given in italics. The first three excerpts were taken from the final day of Jodi Ann Arias's cross-examination (Day 25): Juan Martinez 8 https://www.usatoday.com/story/news/nation/2015/01/13/secret-jodi-arias-transcripts-released/21698747 9 https://www.youtube.com/watch?v=AJ4Svvm7reY 10 https://pt.scribd.com/document/252552705/Jodi-Arias-transcript-from-Nov-3-2014. was the deputy county attorney, Kirk Nurmi and Jennifer Willmott worked for the defendant, and Sherry K. Stephens was the judge. The remaining excerpts were obtained from the reporter's transcript of proceedings. We observe that the first pair-part of the "question/answer" adjacency pair is prefaced by the discourse marker so. It is relevant to note that discourse markers, in this work, are seen as syntactically independent elements from verbs that can be formed by one or more lexical items or expressions. They corroborate the monitoring of conversation as well as the organization of spoken text. In addition, they are also multifunctional mechanisms because they can operate as textual organizers and/or articulators, illocutionary force indicating devices, verbal planners and mitigators, among other functions. Schiffrin (2001, p. 57) postulates that "discourse markers could be considered as a set of linguistic expressions comprised of members of word classes as varied as conjunctions (e.g., and, but, or), interjections (oh), adverbs (now, then), and lexicalized phrases (y'know, I mean)".
Semantically, discourse markers exhibit "semantic aptness," or appropriateness for the type of discourse marker that they become; more importantly, their semantic development provides evidence for unidirectionality, for referential (propositional) meaning being the source for pragmatic (textual and interpersonal) meanings (Brinton, 2001, p. 149). In the example above, the discourse marker so is used as topic developer. It occurs at the beginning of the turn and appears in a direct question which contributes to encourage the interlocutor to confirm the speaker's inference.
After reading the defendant's statements under direct examination, Mr. Martinez asks her to confirm his comprehension about what really had happened. Although in everyday conversation, questions and answers are widely used by participants, the typical adjacency pair "question/answer" can vary in forensic contexts. Gibbons (2008, p. 119) states that "in courtroom examination however the atypical social relations mean that the underlying assumption, subject to certain exceptions, is that the lawyers and the judge ask the questions, and the witness replies". In these settings, "witnesses are usually not permitted to reply off topic or to refuse to reply (unless a reply would be self-incriminating)". We have to take for granted that many cross-examining lawyers do not ask questions to which they do not know the answer.
At a first glance, we might think that he seeks to summarize the events in order to check if he understood them well ("So... after two o'CLOCK... that... you... took everything and put it down in the car correct?"). However, by asking her to answer if this information was correct, he asks her to make the decision to confirm it, and uses the interactional discourse marker correct, which suggests that it is not a mere attempt to understand what she said as this type of marker has the same function as a tag question. These markers are employed to seek discursive approval in argumentative and interactional contexts. It appears as an exposure of information that the lawyer built upon his own interpretation rather than a confirmationseeking request since this strategy is obviously the best for him to base a subsequent question on. By means of using so and correct in the same question, he tries to lead her to confirm that what he was saying was an acceptable version of the truth, leading the audience to assume that his interpretation is valid by obtaining confirmation of a version of events that he had in mind. Besides, as it was already mentioned, he uses a so-prefaced question which plays, according to Holt and Johnson (2010, p. 26), an important role "in evaluatory summary and are effective in challenging and transforming the interviewee's account, to such an extent that they may be forced to reformulate it in a way that is evidentially more significant". Holt and Johnson (2010, p. 25) stress that researchers exploring legal talk (and other institutional environments such as news interviews) have noticed patterns in the design of questions that are associated with particular actions in these environments: and-and so-prefaced questions (Johnson, 2002), formulations (Heritage & Watson, 1979;Heritage, 1985) and reported speech (Philips, 1986;Matoesian, 2000;Galatolo, 2007). These devices, recurrently employed in questions (and sometimes in other parts of trial discourse such as the use of reported speech in summing up), are often central to one of the main aims of legal discourse: to establish the culpability of one or more parties involved.
The discourse marker correct contributes to make the interviewee agree with the interviewer's version of events, as the authors argue: "the attorney's questions are designed to elicit 'yes'/'no' confirmatory answers" (Holt & Johnson, 2010, p. 34). In this case, Arias does not offer an explicit affirmative answer, as expected by Martinez ("Yes, that is correct"), but she does not entirely exclude his version either. When she says, "that makes sense", the interviewer achieves his aim since she does not disconfirm his interpretation.
When Martinez says "pardon?", it does not seem to be just a hearing problem (because she said in a very clear and loud voice), but a persuasive strategy as he might intend that the witness repeats what she has just explained to make her to say something that could be used as incriminating evidence. Then she reformulates her utterance by changing having sex into otherwise engaged, which is more polite and could also be more acceptable to listeners. This device helps her construct a positive image as it would be convenient to show the audience (particularly the jury) her civilized and ladylike side. By means of this paraphrase, she repairs what might be seen as a negative point and highlights what she desires to be displayed: to be recognized as a respectful person. It is what Scollon and Scollon (1983) denominate "deference politeness" inasmuch as she chose a more polite and respectful expression than the one she had used previously.
We point out that, in this research, the words "positive" and "negative" do not mean "good" or "bad"; they are just used to indicate two opposite poles. In other words, positive politeness is related to the interlocutor's desire to belong to a group, being thought of as a member of it, whereas negative politeness concerns the interlocutor's respect, distance, and deference. In this vein, Scollon and Scollon (1983) call positive politeness "solidarity politeness" and negative politeness "deference politeness". Bazergan (2017, p. 40) postulates that "lying behind solidarity politeness is the assumption that there is little distance (-D) 11 between the participants and that there is also at most a slight power (-P) difference between them".
Concerning deference politeness, "participants are considered to be equals or near equals but treat each other at a distance" (Scollon & Scollon, 1995, p. 44). Therefore, deference politeness draws attention to the participants' distance and solidarity politeness is more focused on the common grounds of the participants' relations. According to Coulthard and Johnson (2007, p. 29), "in real courtrooms, when questioning witnesses, many of the lawyer's turns are 'face threatening acts', which challenge the witness". In agreement with the authors, "the lawyer's role in the courtroom (particularly in cross-examination) allows him to interfere with and put pressure on the witness in pursuance of the interactional goals of the Crown or State, by whom he is given his power". This device makes her version weaker because of the inconsistent way she describes the time of actions. She says that it was after two o'clock when she was putting the luggage in the car, and when Martinez asks her again: "And it was surely after two you... did what you were doing?", she is totally vague. We can observe her hesitation by the use of the verbal planning markers Uh and well, accompanied by pauses and discourse markers which show uncertainty: maybe maybe not, I guess, and I don't know. On the other hand, it is interesting to see that the prosecutor's wording is vague as well: is "did what you were doing" having sex or putting the luggage in the car?. When questions are not clear, answers also tend to be unclear, and consequently this can make listeners confused because of the impreciseness of the utterances. We highlight the use of the formulation prefaced by the discourse markers and and so in "And so... when you're saying you're not sure". Martinez prefaces his formulation with a conclusion: "and so you're saying", but he also preserves the sense of what Arias said, according to what he understood, basically repeating that she is not sure. Then he confronts her answer by asking her "you were a little bit more sure under direct examination weren't you?". This formulation deletes the part of her utterance that she is not sure, transforming it into other information which provides another way to interpret that: if she was a little bit more sure before, why is she not sure now? So, he keeps on asking her to get precise information that could be useful for his arguments, as we can see through the segment "But you were more... sure about the time... do you remember back to you?".
By drawing attention to the use of tag questions, Gibbons (2008, p. 121) points out that "perhaps the most widely discussed type of courtroom question is the tag question". In agreement with him: The first part of such a question takes the form of a statement, in which the lawyer can include his/her version of events (the information). The second part consists of a tag, which exerts various forms of interactive pressure upon the witness (the social). This form of courtroom question is therefore a paradigm example of linguistic form matching pragmatic function, and in consequence it is not surprising that many questions in cross examination take the form of tags, and that there are many types of tagssome of them unusual in everyday discourse.
In Excerpt 3, when Mr. Martinez says, "you were a little bit more sure under direct examination weren't you?", he enhances the notable contrast between Arias's previous statement and now: she was sure before, but now she is not sure about the exact time she had put her items in the car. Knowing the time when she did that is extremely important to the construction of his accusation, we observe therefore a degree of coerciveness in his questioning. In general, most closed questions such as yes/no questions are more coercive because they offer a limited possible set of answers. Although tag questions invite the interlocutors to say yes or no, they are more persuasive since they can make them (the interlocutors) feel more vulnerable because of the preceding utterance. Tkacuková (2005, p. 335) asserts that: In comparison to yes/no questions, tag questions and declarative questions are strongly biased towards a confirmative answer and so they are more coercive (Huddleston & Pullum 2002, p. 881-894). Apart from coercing witnesses into type-conforming replies, these types of questions offer one more obvious advantage to counsels: they are perceived as statements and as such help to change the questions into evidence (Hobbs, 2003b, p. 486-87). Tag questions and declarative questions thus become a powerful tool which enables counsels to give evidence on behalf of witnesses and reduce witnesses to the role of minimal responders. But it is tag questions that are considered by some scholars (Woodbury, 1984, p. 205;Berk-Seligson, 1999, p. 36) to be the most coercive type of questions as they have an additional pragmatic meaning: they imply that the person knows that the questioner is right. Gibbons (2003, p. 101), however, warns of relying heavily on the form of questions since intonation, or tone of voice, is equally important for defining the degree of coerciveness.
Pragmatically speaking, like tag questions, we consider questions that end with the discourse markers correct and right as devices which have the same illocutionary force. Other examples that illustrate this are found in Excerpts 1 and 2, respectively, when Mr. Martinez says "So... after two o'CLOCK... that... you... took everything and put it down in the car correct?" and "it was at two specifically it was around two o'clock when you actually put the luggage in the car right?". Relying on Gibbons (2003, p. 95), Holt and Johnson (2010, p. 25) show two objectives that questions can have in legal settings: "1. eliciting information and 2. obtaining confirmation of a version of events that the questioner has in mind". Our examples under analysis seem to constitute devices with the latter objective.
The following excerpts were obtained from the reporter's transcript of proceedings. It is important to highlight that "Q" represents the questioner (Ms. Willmott), and "A" stands for Arias. Mr. Martinez: well, she can certainly ask her, describe the banquet for us, who was there and that sort of thing, but you know, she's -- The Court: Well, technically she was not answering the question. She was answering who was in attendance. Why don't you just restate the question that you technically asked her. So she can answer it. So just restate it.
Q: What type of people -Jodi, what type of people attended this banquet?
A. Again they're executive directors and one invite, so typically a spouse or friend.

In Excerpt 4, through the formulation tell us what happened generally about what happened
at the executive director banquet, we see that there is a transformation of the semantic framework because the original question tell us what happened at this executive director banquet? does not imply an answer in general terms. Ms. Willmott then reformulates her question in her line of argumentation including this new information which had not been asked before, considering that it was assumed by interlocutors. It is important to stress that this is just one small example of how courtroom questioning has different rules. In ordinary conversation, Arias' reply would have been totally unremarkable. As Roca-Cuberes (2014, p.

310-313) points out, institutional talk can be analysed in different settings:
It has been observed that the participants in formal settings such as courtrooms, classrooms, news interviews, etc., tend to methodically organize their conduct so as to exhibit and produce the 'institutional' quality of their encounter (Drew and Heritage, 1992, p. 26). Furthermore, their conduct is constrained by a diminution of the array of options and possibilities for action that usually take place in ordinary conversation. […] In comparing the use of formulations in institutional settings and ordinary conversation, Paul Drew notes that formulations are rarely found in ordinary conversation (Drew, 2003). According to Drew, this is so because "we do not need to arrive at compromises after long negotiations in mundane conversation" (Drew, 2003, p. 306).

Using the expression I don't believe it's non-responsive, Ms Willmott tries to explain why
Arias had to describe and give some background information to answer her question. Markers such as I think, I believe, and I assume, for example, are classified as opinion markers, commonly used with the first person singular. Although the epistemic value of opinion verbs usually expresses certainty and conviction, they sometimes may denote uncertainty or inaccuracy. In this case, the opinion marker is used as a facework device since it contributes to limit or neutralize possible contrary reactions or harmful interpretations from interlocutors.
So, we observe an intention to guide listeners (mainly the Judge) on how to interpret the speech. It is a modalization marker as it mitigates the illocutionary force of the utterance.
When using I don't believe, we can protect our positive face by preserving our opinion from those who may not have the same opinion.
In the example above, there are two markers well which are used to help speakers gain some time to elaborate their speeches. We also observe the marker you know employed for purposes of social interaction. These devices are widely used in spoken language and are responsible for the development of the conversational turn. By means of the marker you know, speakers seek interlocutors' anticipatory discursive approval in order to prevent themselves from the possibility of possible disagreements. It implicitly induces the listener to identify with the speaker's speech or dissuade those who have a different opinion. It is a measure taken by the speaker to prevent something unpleasant or inconvenient from happening.
Combined with the time planner marker well, the Judge uses the hedge technically in "Well, technically", which is a mitigation sign. It is an element that functions as a hedge as it promotes the deletion of marks of enunciation to minimize potential risks to the speakers' face. It softens the danger embedded in the speaker's statement and guides the listener to a convenient interpretation of their statements.
The term "hedge" adopted in this research was coined by Brown and Levinson (1987, p. 145): "the particle, word or phrase that modifies the degree of membership of a predicate or noun phrase in the set; it says of membership that it is partial, or true only in certain respects, or that it is more true and complete than might be expected". In addition, we can see another mitigation device why don't you restate in "why don't you just restate the question that you technically asked her". It reduces the force of an imperative verb such as restate and makes the request more polite, protecting therefore the speaker's positive face.
SciELO Preprints -This document is a preprint and its current status is available at: https://doi.org/10.1590/1678-460x202243363 Ms. Willmott: Judge, I asked to approach because it's going to take a very long time if we keep getting these particular objections. The idea is what she was doing. In order to describe what she was doing, she starts to describe who was there so she can tell us who she was talking to. If she just starts talking -saying I was talking to Michelle, that's another objection for foundation for whose Michelle. So it's just a general descriptive question. I don't think it's non-responsive. She needs to answer these questions. She needs to describe who is there and give the general details in order to answer the question.
Mr. Martinez: She was asked what she did not who was there. If she wants to ask who was there, she can ask her that, I mean -

Ms. Willmott: It's the same --
The Court: Well, I understand your position. But when he makes an objection, if you could restate the question and you don't even need to come back here. I think you can clearly ask the questions you're asking. The problem is she's answering the question slightly different from what you're asking. So technically the objection is well founded. So just restate your question. I don't think the State is objecting to her answers to any of the questions. It's just that she's not being responsive to the question you're asking.
Ms. Willmott: I guess just for the record, Judge, I just see it as an objection that's to interrupt the proceedings because the actual material of the information that she's saying is not objectionable, but rather it goes to the question. I ask to approach to make just to make that record, Judge, because this is going to take very long if she doesn't answer exactly my question each time.
The Court: The problem is if she doesn't answer the question then we don't know what she's about to say. So it could be objectionable. So the question frames her answers, I think that it's appropriate. The State is able to object. Just restate your question.
Mr. Martinez: Quite frankly this witness is hard to hear. She's going really quickly. So adding to what you said, when she starts answering something that she is not asked, it's difficult to follow. I mean, that's part of the problem.
Regarding the use of I think and I don't think in Excerpt 5, we highlight that they have the same functions already observed in Excerpt 4. This is a mitigation device as it helps the speaker reduce their responsibility for upon what they say. The role of this kind of marker is to mitigate, to protect the speaker's face as it reflects, in a subtle way, an opinion semantically full of ideological values and different goals in the interaction. Markers such as I think (that), I believe (that), I know, and it seems to me, for example, apart from presenting an opinion, have a mitigation value because they not only reveal the presence of the interlocutor, but also contribute to reduce the speakers' responsibility for the opinion they state. When the Judge says "so the question frames her answers, I think that it's appropriate", for instance, she points out the reason for sustaining the objection. Since the question is very objective "what do you do", not "who was there", the defendant has to answer it in a very objective way as well otherwise it could be objected to. Even though it is clear that the interrogative pronoun what used in the question does not accept other answers, she uses the opinion marker I think to affirm that the objection is appropriate. It is a strategy that contributes to soften her responsibility on what she says and, at the same time, to preserve her positive face as her attitude may be seem by listeners as a quite fair judgement.
Another element from informal conversation is the marker I mean. In the case above, this device is employed by the speaker to guide the listener to a desired interpretation. This metalanguage form of modalization approaches the discourse practice and the representation of speech. The speaker draws attention to the way things are being said, directing or redirecting the listener to a particular intended meaning. Through I mean, speakers try to lead the listener's attention to something that is more convenient for them. Regarding the markers well and you know, we point out the same functions mentioned in Excerpt 4, as well as the hedge technically.
Excerpt 6: Q. After leaving that convention, did you have any type of impressions about --that you learned from convention?
A. Yeah. It's --I had never been involved in a network marketing company before. So that was my first experience with that. It's very inspiring. I had relatives that had tried to do things like that before with other companies. They were not successful.
Mr. Martinez: Objection, non-responsive as to her impressions. A. This was --I was at a Super Saturday in my region in Anaheim and we were texting back and forth. It was just getting more flirty and more flirty and we kept trying to -Mr. Martinez: Objection, non-responsive. When?
The Court: Restate your question.

Q. When is it that he sent --approximately when is it that he sent you a picture of his genitalia?
A. It would have been November 11th, 2006. In Excerpt 7, the first version of Ms. Willmott's question is "And this is when?", referring to Alexander sending a photograph of his genitalia to Arias. In the second version "When is it that he sent --approximately when is it that he sent you a picture of his genitalia?", we can see that one part of the meaning established before is still preserved, but now transformed by resuming her previous question. In other words, in the second version, she preserved the exact word used in the first version (when), added some new information, and transformed it into a more objective question. The real intention is to know when Mr. Alexander sent a photograph of his genitalia to Arias, but due to Mr. Martinez's objection, Ms. Willmott restates her question in a more accurate way. An interesting point is the use of approximately in the middle of the sentence. To avoid further objections, she prefers to introduce a general term that could protect Arias's face in case her answer was vague or not precise. A. His view is that vaginal sex is like --that's the holy grail, you don't go there until you're married.
In Excerpt 8, another objection is raised by Mr. Martinez. He claims again that the defendant's answer is not responsive to the question "What is it that you're learning?".
However, in this case, we do not observe any preservation property in Ms. Willmott's second version of the question, but rather deletion and transformation characteristics: "What was Mr.
Alexander's description of premarital, LDS of you want premarital sex". The message in the original question is totally deleted and transformed into a new one, so much more specific.
After several unsuccessful previous attempts against prosecutor's objections, Ms. Willmott chooses to restate her question quite directly (yes/no question), thus preventing the defendant's answer from a new objection.
Our analysis shows that formulations are very important tools in courtroom interaction since the questioners can selectively preserve, omit, alter, or even add to what the interlocutors have said. It is also relevant to consider that, by means of a textual-interactive perspective to the spoken language, it is possible to observe the process of text construction as well as the strategies that are employed to achieve the interests and purposes of the participants in this kind of forensic context.

Conclusion
According to the findings, all properties of formulations, combined or not, were used: preservation, deletion, and transformation; although the majority of the cases had a predominantly transformative function. The cross-examining lawyer preserved some of the defendant's words exactly, for example: that makes sense in "when you say that makes sense although it makes sense" (here, the words were preserved deictically in the second occurrence as that had a grammatical substitute it). In other moments, he omitted some aspects of the sense, but he mostly used altered forms in his formulations, that is, by transforming what the defendant had said, he offered his own version of what she had said and invited the listeners to agree with the interpretation that he had in mind.
It is important to call attention to the use of so-prefaced questions and the markers correct and right functioning as tag questions in some prosecutor's formulations. So-prefaced questions were very useful to challenge and transform the defendant's account by putting pressure on her to reformulate it in a more convenient way for the accusation. In relation of the discourse markers and tag questions above, we observed a degree of coerciveness in the crossexamining lawyer's questioning. These types of questions are more coercive as they offer a limited possible set of answers and, pragmatically speaking, they imply that the interlocutor knows that the questioner is right.
As observed in our corpus, mitigation devices contributed to preserve the self-image constructed by speakers as well as to produce certain meaning effects on listeners. The hedge technically and markers such as I think, I don't think, and I don't believe collaborated to neutralize possible contrary reactions or negative interpretations from interlocutors; they lessen the impact of speakers' utterances and decisions by distancing their image from divergent judgments or opinions, therefore protecting their positive face. When the Judge sustained the objections claimed by the deputy county attorney, for instance, she chose to employ politeness strategies with the defense lawyer. To reduce the force of an imperative form, she used "why don't you restate" in: "why don't you just restate the question that you technically asked her", which functions as a polite request that plays down its face threat.
Both in ordinary conversation and in forensic settings, the particular circumstances in which questions and answers are developed make facework a constant need in verbal interactions.
Since it is not possible to predict what interlocutors think or act, speakers employ mechanisms to ensure the defense of what they do not want to see exhibited and highlight what they want to be shown. According to Flower (2018, p. 31), facework strategies have multiple uses intended to save, maintain, give, and threaten face.
Politeness strategies are also valuable mechanisms that speakers use to soften the illocutionary force of what they say. They are sociointeractional strategies which are associated with people's necessity of gaining approval from others as well as the need to safeguard their autonomy. Among these strategies, we point out the use metadiscursive expressions, evasive forms, paraphrases, and markers that indicate higher or lower degrees of speaker's involvement, among other linguistic phenomena.
Regarding the monitoring of information flow, we consider the concept of involvement elaborated by Chafe (1982, p. 47): "involvement includes monitoring by the speaker of the communication channel which exists with the listener, and attempts to make sure that the channel is functioning well". The speaker can employ certain strategies that help the reaffirmation of what is being said by calling listeners' attention to a certain point of the subject in progress. Therefore, well, you know, and I mean, among other markers, are very appropriate for this kind of function.
Based on a textual-interactive approach to the spoken language, this study lies at the interface between conversation analysis and forensic linguistics. In the same vein, we hope, above all, to encourage the development of further analyses and reflections that may contribute to the studies concerning spoken texts in forensic contexts.