Taking offence at the (un)said: Towards a more radical contextualist approach

: Many researchers in impoliteness studies have set themselves the task of determining, amongst other things, (i) what linguistic or non-linguistic phenomena can cause offence, and (ii) why people take offence. However, the reality of interaction clearly shows that, on many occasions, there appears to be a marked dissonance between the speaker and hearer in their evaluations of offensive language, even in locally situated interaction. More research is therefore needed to account for and explain why and how the hearer assigns a particularly offensive meaning to an utterance during the course of an interaction. With this aim, and by drawing on insight from what is referred to as “radical contextualism”, in this study we discuss the possibility of looking at how interactants can arrive at their own (subjective) evaluations of impoliteness in ways that do not match up with the alleged intentions of the so-called offender. Drawing on a number of exchanges that involve such instances of taking offence, we will argue that the taking of offence should best be viewed as a process over which the hearer has a more active control. Accordingly, the paper contributes to current attempts at explaining the variability involved in the taking of offence.


Introduction
The study of impoliteness beyond what could be called the 'intention paradigm' has recently gained considerable momentum. These studies are triggered by the discovery that (i) people can and do take offence without necessarily being the target of markedly or grossly offensive language or conduct, and (ii) people do not readily take offence at words or conduct that are generally associated with impoliteness (Haugh 2010;Culpeper 2011;Kádár and Haugh 2013;Tayebi 2016Tayebi , 2018. However, on many occasions during the course of an interaction, it is clearly evident that there is dissonance, or rather lack of agreement or harmony, between the speaker and the hearer in how they evaluate certain language (or conduct) as offensive (Tayebi 2016;Parvaresh and Tayebi 2018;Haugh and Chang 2019). This is particularly important because, in reality, impoliteness is not always resolved.
While in more recent studies (see, e. g., Culpeper 2011; Kádár and Haugh 2013;Haugh 2015a;Spencer-Oatey and Kádár 2016;Tayebi 2016;Langlotz and Locher 2017;Ogiermann and Blitvich 2019;Parvaresh 2019;Kádár 2020, to name but a few) impoliteness is shown to be a negative evaluative meaning (Eelen 2001) which is arrived at by the participants in the interaction. A pertinent question which has not yet been properly answered would be what counts as a legitimate inference for which the hearer might be able to hold the speaker accountable (Haugh 2017)? And, even more importantly, to what extent are such inferences amenable to analysis? One way forward in the way of providing some answers to these questions would be to rely less on speaker-intended inferences and more on "assumptions which […] are not speaker-intended" (Ariel 2019: 104). That would require us as analysists to pay more attention to the hearer's role in the process of taking offence and to put under scrutiny those occasions where some sort of disagreement arises, between the speaker and the hearer, over what was meant. In these situations, the speaker's accountability dissolves and "it is the hearer who pursues further assumptions in his/her search of an interpretation" (Tayebi 2018: 103).
Influenced by some tenets of what is generally referred to as radical contextualism, this paper therefore seeks to explore the possibility of providing a more accurate analysis of those instances of taking offence which involve a lack of agreement between the speaker and the hearer. The explanations provided in this paper attest to the potential avenues that can be explored by accounting for the many nuanced layers of inference that are prevalent in designating a language as offensive.
The structure of the paper is as follows. In Section 2, an account of radical contextualism will be provided, which will provide the groundwork for the study. In Section 3, we will discuss the data used and the procedures followed. Section 4 will be devoted to a thorough analysis and discussion of a number of revealing examples. Finally, Section 5 provides a summary of the findings and some concluding remarks.

Why radical, not moderate, contextualism?
In general, pragmatics appears to be based on the tacit assumption that a difference exists between what an expression is saying and what its use in a specific context actually means (O'Keeffe et al. 2019;Zhang and Parvaresh 2019). Over the years, this has given rise to the distinction between saying and meaning (Bach 2012). In this respect, and as far as research on impoliteness is concerned, a rather dominant view seems to be that the hearer arrives at an understanding for an utterance in a controlled, bottom-up manner, but is assisted by a more top-down process which involves pragmatic inference. As perceptively noted by Recanati (2010), generally speaking, this view means that pragmatic inference pretty much determines the content (i. e., "what is said"1), but does so by being triggered by the utterance itself and/or some aspects of context. For example, consider the word "bitch", which out of context is usually viewed as an impolite expression but, in some contexts, for example in an intimate relationship, could be an expression of endearment (Culpeper 2011). On paper, calling one's partner or wife a "bitch" would mean that she is rather unpleasant. In the context of an intimate and evidently unproblematic and smooth relationship, particularly one in which both the speaker and hearer are assumed to be cooperative and rational (Grice 1975), this semantically determined meaning of "bitch", i. e., one which is impolite and offensive, does not sound plausible. Therefore, a pragmatic inference is needed, by the hearer, to provide a more "contextual" value to enable the implicature (Grice 1975) or the implied meaning of the word "bitch" in the described context to be determined. Accordingly, such an implicature would encompass a range of phenomena (e. g., the intention of the speaker, their beliefs, their relational histories, emotions, etc.) which the speaker plausibly intends the hearer to recognize, and which is compatible with expressions of endearment or even encouragement. Accordingly, the hearer's inferential processes determine the content in an interventionist way and, in this view, pragmatic processes of meaning making are therefore "subordinated" to the meaning of utterances (Recanati 2010: 4). As such, a number of pragmatic processes (e. g., inferring a positive meaning from the word bitch) are activated, but this process is predominantly guided by the linguistic material (Recanati 2010: 3-5;cf. Dänzer, in press).
The aforementioned view appears to encounter a form of circularity (Levinson 2000) as it assumes an unduly heavy interdependence between saying the word "bitch" and implicating endearment or encouragement. In other words, in this account, the endearment implicatures appear to be dependent on the hear-er's prior determination of what the word "bitch" implicates based on what is said, i. e., "bitch". At the same time, what is said, i. e., "bitch", depends on what is meant, or rather the implicature, e. g., "you are dear to me" (see, Levinson 2000). As such, "what is said [i. e., "bitch"] seems both to determine and to be determined by implicature" (Levinson 2000: 187). This is indeed a circular problem, which, if it remains unaccounted for, prevents us from providing a convincing explanation as to why, on some occasions, the hearer's inferences seem to deviate from, or at least be incompatible with, the speaker's anticipated interpretations, particularly in those contexts in which such deviations could cause offence on the part of the hearer. As Ariel (2019: 111) notes, such inferences: are conclusions an interlocutor is in principle reasonably (though not necessarily) justified in drawing about the reality behind the speaker's utterance, but he can only do so at his own responsibility, since such assumptions do not correspond to a speaker communicated message. (Ariel 2019: 111, italics original) One plausible solution to account for, or even resolve, both the problem of circularity as well as the so-called disagreement in perceptions of offence referred to above, is to provide an account in which we posit the presence of a number of "free" pragmatic processes that are not activated by any expression a priori, but take place entirely "pragmatically" with the aim of making sense of "what is said" (Recanati 2010: 4; see also Bezuidenhout 2002;Marchesan and Zapero 2018)2. To clarify, as far as our above example is concerned, the fact that the hearer is likely to interpret the word "bitch" as an expression of, say, endearment, is not because the utterance and the circumstances under which it has been used have dictated this in a bottom-up fashion, but rather because the content of "what is said" is determined by the hearer in a top-down manner3 right from the outset. The latter account has its roots in radical contextualism, a view which assumes that information that is not encoded linguistically pretty much determines the content of "what is said" (Recanati 2010: 4-5). A central tenet underlying such an account is that "we must go beyond linguistic meaning without being linguistically instructed to do so" (Recanati 2004: 139;see also, Carston 2004see also, Carston , 2012. Extrapolating from this, in what follows we explain how the examples of taking offence discussed in this paper should best be viewed as being caused by "a pragmatically controlled pragmatic process" (Recanati 2004: 136, italics original;cf. Haugh 2002), which, theoretically speaking, rules out the possibility of linguistically motivated inferences followed by contextual adjustment, in support of more top-down "pervasive contextual adjustment" (Huang 2017: 983) of meaning by the hearer.

Data and procedure
The data used in this study is a corpus of movie and TV scripts and also reality TV shows, which have been either retrieved from the Internet or transcribed by the authors. Special care has been exercised to incorporate only those excerpts in which both researchers could observe, given the many verbal and non-verbal clues, the affective stance of feeling offended on the part of one of the interactants, at what is referred to, explicitly or otherwise, as some sort of, for example, transgression on the part of other participants. The emotional stance of feeling offended includes a wide range of feelings ranging from displeasure, annoyance and anger to feeling hurt (Culpeper 2011;Haugh 2015b;Márquez-Reiter and Kádár 2015). Given the focus of the present study, the initial corpus was subjected to a thematic analysis by the researchers and was qualitatively analyzed with a view to finding examples of taking offence in which there was a mismatch between the interpretations of the speaker and hearer. As for the authenticity of the data, we believe, in line with Mishan (2004: 225), that movie scripts are close to authentic data in that they are "written to be spoken". And, in reality TV shows, although the participants are under constant video surveillance, the highly dynamic nature of the programme as well as the unexpected turn of events, which is an inherent feature of these programmes, turns the interactions that take place during these TV shows into authentic ones. Given the scope of the current special issue, namely the alternative approaches to impoliteness, and given the limitations of space, only a few revealing excerpts from the corpus will be analyzed here.

Analysis
As far as the literature on the taking of offence is concerned, the inferential processes which enables the hearer to arrive at an offensive reading of one's utterance are thought to be post-propositional as they appear to be based on the premise that the speaker has said something which has turned out to be (or rather has been perceived as) offensive. Accordingly, to arrive at an offensive reading of the speaker's utterance, the hearer needs to determine "what the speaker means" by assigning values (e. g., time and place) to the utterances, and then complementing that information through pragmatic inference assisted by the so-called conditions, including the relational histories (cf. Locher and Watts 2008), under which these utterances/expressions have been used. To fully understand how the aforementioned conceptualization pans out, in the following discussion we first analyze an example from a reality TV show with the aim of further explaining the strengths and (possible) weaknesses of the above conceptualization with a view to paving the way for a more refined account inspired by the tenets delineated in Section 2. This will then be followed by an analysis of two more examples4.

Moderate contextual evaluations of offence vs. the more radical contextual account
In this section, we will look at an instance of conflictual interaction from two different perspectives: first, a more moderate contextual account, which is then followed by another account, with a view to highlighting the suitability of the latter. The example is taken from a 2019 episode of the British TV programme, Love Island. Love Island, a popular British TV programme, features a group of young contestants, also known as Islanders, who live for a short period of time in a villa in Mallorca with no contact with the outside world, but who are under constant video surveillance. In order for the Islanders to remain at the villa, i. e., not to 4 One of the reviewers of the manuscript rightly noted that (im)politeness scholars "don't necessarily base their analyses on the assumption that interlocutors need to reject the literal meaning of an utterance to infer an implicature when interpreting (im)politeness" and that they "ultimately emphasise the importance of context -often only available to the offended party". This is no doubt an accurate description of (im)politeness studies these days. However, what we intend to do in this paper is to provide an account in which taking offence could be viewed as the offendees' (rather pervasive) attempts at going beyond (con)textual meaning at their own responsibility.
be "dumped" from the Island, they must do their best to couple up with another Islander, an endeavour which they hope will result in a romantic relationship. Each episode is full of unexpected surprises and challenges for the Islanders.
During one of the episodes, a contestant, Maura, receives a text message from the producers advising her that she has earned a night in the villa's only private bedroom, referred to as the Hideaway, to spend some private time with a person of her choice. Shouting, "Someone's going to get frisky", Maura picks one of the boys, Tom, with whom she is now coupled-up, to join her in the "Hideaway".
While Maura is preparing to go to the Hideaway, Tom is seen chatting with a few of the other male contestants. They appear to be engaging in banter about what is going to happen in the Hideaway and whether or not Tom is open to the idea of having sex: To:::m! 9. Anton: Someone is calling you mate! ((Maura walks towards Tom)) 10. Tom: ((Tom turns around, notices Maura and starts walking towards Maura)) Yeah, I'm not overthinking it but it'd be interesting, it'd be interesting to see if she's all mouth or not! ((Maura looks perplexed and boys look curiously excited)) 11. Michael: I think she heard him! ((all boys laugh)) 12. Maura: ((looks curious)) What did you just say to them? 13. In the above conversation, the participants seem to be somewhat excited and curious about what is going to happen in the Hideaway and whether or not Tom is open to the idea of having sex (with Maura). In turn 5, Tom boasts about his previous relationships claiming that he has slept "with girls after much less time", a claim which may suggest that he thinks he can do the same in the Hideaway. Michael's remark in turn 6 attests to this interpretation. At this point, Maura calls out Tom's name and walks towards him and the group of contestants. Tom turns around, sees Maura, and begins walking towards her while at the same time saying to the boys, "I'm not overthinking it but it'd be interesting to see if she's all mouth or not!" (turn 10). As Tom moves away from the boys towards Maura, Michael guesses that Maura has probably heard what they were discussing, or at least has heard part of the conversation. Throughout the question and answer exchange that follows between Tom and Maura, Tom appears to be somewhat puzzled by Maura's request for information about his comments to the other Islanders. As this exchange shows, Tom firmly believes that what he has said is not inappropriate. He even tries to laugh off the seriousness of the situation (turn 19), but this just makes Maura even more angry.
Technically speaking, taking the role of all the participants in this exchange into account, in the first 11 turns, Tom and the other lads involved in the banter are "ratified" participants in the sense that they are actually engaged in the conversation in question, but Maura is an unratified (eavesdropping) participant with whom Tom does not wish, at least not immediately, to communicate (Goffman 1981;Goodwin 1981). Maura is the unratified participant who, given the fact that she is farther away, is unable to decode or understand what Tom is saying in turn 10. In one account, Maura's taking of offence and anger appears to have been caused by the mere overhearing of, or rather confirmation of, the fact that Tom has said "it'd be interesting to see if she's all mouth or not".
Linguistically speaking, the idiomatic expression "to be all mouth", used to describe people who have a tendency to say "things just to impress people" and do not really intend to do the things they say (Macmillan Dictionary 2019), is, of course, not a polite way to describe people (see Allan and Burridge 2006). However, in the context of the banter with the other lads -a context from which Maura is absent and is therefore unable to hear exactly what is being said -"being all mouth" serves as Tom's contribution to the ongoing banter; possibly even an attempt to express his excitement at the thought of spending some private time with Maura5 (cf. Sinkeviciute 2019). However, as soon as Maura is made aware of this, either by Tom directly telling her or by overhearing it, the entire interaction takes a different turn. In this context, "it'd be interesting to see if she's all mouth or not", uttered in a situation involving other boys, enables Maura to attribute a mental state, from a third person perspective, to the behaviour observed (Ciaramidaro et al. 2007). In light of some of the mainstream approaches to (im)politeness, one could argue that, for Maura to take offence and become angry at what Tom has said, she would need, first and foremost, to work out the following: (a) the time and place of Tom's utterance and the people involved (e. g., Maura); (b) other contextual information (e. g., the fact that Maura had previously, quite openly, talked about her desire to have sex on the island, that Tom and Maura had already shared a few kisses, or even that in the banter talk in question she had been treated like an outsider).
Activating such relevant contextual information, triggered by the linguistic material, seems to be the most important top-down pragmatic process involved. While the above understanding of taking offence is in itself rewarding, it does not appear to adequately explain the complex nature of the taking of offence in this example. This makes more sense if one takes into account some crucial background information, especially the fact that in the follow-up interviews, Tom told journalists (see Gallagher and Downham 2019) that Maura had already demonstrated the habit of "putt[ing] herself onto Tommy [one of the other boys on the Island]", who had even "dodged a couple of kisses from her" despite her persistent endeavour to win him over. Tom also said that he "was really surprised at [Maura's] reaction" as she seemed very open to the idea of sex and explicitly talked about it. Consequently, while "being all mouth" could trigger the above processes involved in arriving at what Tom meant, these cannot be the main contributing factors. This motivates our alternative account according to which most of the inferences the hearer, or rather the offence taker, makes are intuitively derived from a number of prior expectations (or rather intuitions) independent of the linguistic material. It is these competing expectations (e. g., the fact that Maura had ended up with Tom whom she did not particularly fancy coupling up with) which more or less guide Maura towards arriving at a so-called more markedly offensive inference than the one envisaged by Tom. In light of the latter account, Tom's comment, "if she's all mouth or not", must be analyzed by conforming to Maura's expectations. Accordingly, Maura's role changes from one in which she deduces Tom's intention to one in which she has more conscious "control" over how she interprets Tom's utterance, thus making assumptions which may (or may not) be necessarily speaker-intended. In this context, to be interpreted as an offensive remark, "what is said" by Tom needs to be "intuitively [and consciously] accessible" (Recanati 2004: 20) to Maura in such a way as to enable her to keep considerations of contextual factors to a maximum.

Radical contextualism and taking offence
Having specified the need for a more radical contextualist account of taking offence, we are now in a position to apply our understanding to the discussion of two additional examples that involve taking offence. These examples serve to further demonstrate how the alternative account of taking offence can be used to explain both general cases of impoliteness evaluation -presumably those that do not involve any mismatch between various inferences -and cases in which there appears to be a mismatch between the perceptions of the speaker and hearer, which can lead to different forms of conflictual interaction (cf. Tayebi 2018).

Vicky Cristina Barcelona
The below example is taken from the 2008 film, Vicky Cristina Barcelona. The film revolves around two American women, Vicky and Cristina, who have travelled to Barcelona for a vacation. Here, they meet a handsome and apparently well-known local artist, Juan Antonio, who is strongly attracted to both women. The below conversation takes place immediately after a scene during which Juan Antonio asks both Vicky and Cristina, whom he has just met, to go to Oviedo with him for the weekend for something to eat, to drink good wine and, possibly, to make love. Vicky is not interested in this forthright sexual adventure because she is committed to her fiancé, Doug. Cristina, on the other hand, appears to be captivated by Juan Antonio's free spirit. The following conversation takes place between Vicky and Cristina after Juan Antonio has left the women to consider his offer: [Vicky Cristina Barcelona. Directed by Woody Allen, screenplay by Woody Allen, Mediapro and Wild Bunch, 2008. Netflix.] As the above exchange shows, the conversation revolves around whether or not Vicky and Cristina should accept Juan Antonio's invitation to travel to Oviedo and spend time with him. Vicky appears to be resolutely against the idea and is doing everything in her power to discourage Cristina from accepting the invitation. When, in turn 12, her fiancé, Doug, rings her, she tells him that Cristina is "making a potentially fatal mistake". However, Cristina does not change her mind, and in what appears to be her last attempt to justify her interest in accepting the invitation, she tells Vicky that she likes Juan Antonio because "he's not one of these cookie-cutter molds" as he is "creative" and "artistic" (turn 13). Vicky retorts, in turn 14, by asking Cristina the rhetorical question of whether she thinks Doug, Vicky's fiancé, is like a cookie-cutter mold. Cristina is greatly shocked by this response and her state of shock is evidenced in turn 15, where she not only tries to express her surprise by repeating Doug's name -possibly in an attempt to ensure that she has correctly understood Vicky -but also by firmly denying, in the form of a rhetorical question, that she was talking about Doug.
Evidently, Vicky has taken offence because she believes that Cristina, by mentioning that Juan Antonio is not a cookie-cutter mold, is insinuating that Vicky's finance, Doug, is exactly that. Doug is not, in any plausible way, the topic of the conversation. His only appearance is indirect and is in the form of a brief phone call that he makes to Vicky during her ongoing conversation with Cristina. So, the question is how and why does Vicky think that Christina is referring to her fiancé? Cristina's surprise reaction in turn 15 attests to the presence of a deviant meaning which is not in line with what the speaker is meaning.
Of course, for Vicky to infer the supposedly "deviant" interpretation of what is being meant by "he's not one of these cookie-cutter molds", she would need to bring a range of variant meanings (presumably her uncertainty about her choice of fiancé, her deep-rooted dissatisfaction with Doug, any previous discourses with her friends, including Cristina, about Doug, etc.) to the inferential process, as input, for such a pragmatic adjustment to take place. Accordingly, Vicky's understanding of what Cristina is saying comprises a set of truth evaluable expectations which are arrived at, not by virtue of the conventional meaning of Cristi-na's words, but rather by adjusting context-dependent information in ways that would personally benefit her6.
As demonstrated in the above examples, a moderate contextualist account would be rather insufficient (Parvaresh and Tayebi 2018) to establish the instances of taking offence that are experienced and/or portrayed in these situations. Taking on the role of an interpreter, the hearer in this context, develops not just his/her understanding of the interaction in question, but engenders new meanings and/or discourses within the course of an interaction. This is not only evident in those contexts in which the hearer pragmatically adjusts the meaning of what he/she hears, but also in contexts "in which there is insufficient linguistic material" (Parvaresh and Tayebi 2018: 94). By way of illustration, let us now focus on another exchange.

He's Just Not That Into You
The following excerpt has been taken from He's Just Not That Into You, a 2009 film directed by Ken Kwapis. This film involves several people, including couples such as Beth and Neil, and tells the story of many romantic adventures and the struggles experienced by different couples. Beth and Neil have lived together for many years and appear to have a relatively normal life, but they are unmarried as Neil has not yet proposed to Beth: [Beth enters the living room while Neil is hammering a nail into the wall, on which he intends to hang a painting] [He's Just Not That Into You. Directed by Ken Kwapis, screenplay by Abby Kohn and Marc Silverstein, New Line Cinema and Flower Films, 2009. Netflix.] As the excerpt reveals, when Beth arrives home, Neil is busy hanging a painting on the wall. Having greeted Beth by saying "hey", Neil asks whether the painting appears to "look straight" (turn 1). Instead of directly answering the question, Beth asks Neil why he is hanging the painting on the wall (turn 2), an action which could be the early signs of her dissatisfaction with Neil. In turn 3, he replies that she had asked him to hang this painting "about three weeks ago". Again, rather surprisingly, Beth does not answer him directly, but instead moves away from the wall on which the painting is being hung. Beth's lack of response and interest astonishes Neil and, in turn 5, he tries to establish why Beth is behaving in this way because, based on the contextual information that is available to him, Neil assumes that Beth no longer wants the painting to be hung on the wall. Beth reassures him that she loves the painting where it is, but also asks Neil to "just stop" (turn 6). This rather astonishes Neil but, at the same time, from his own perspective which is based on his background and the contextual information available to him, he believes that Beth's lack of interest must be related to the painting in question, and nothing else. Accordingly, he suggests that Beth might not like the painting itself as it has, for example, an image of a "deflated boob" which she might quite understandably find "depressing" (turn 7). In the same turn, Neil offers to take down the painting if Beth does not like it. Beth's reply in turn 8 is an emphatic no, which she immediately follows by her asking him "to stop doing anything nice". Prior to turn 8, Neil's understanding of Beth revolves around the painting, either its location or the painting itself, and hence Beth's request is totally unexpected. This is evidenced by Neil's response in turn 9, "This feels like a trick", when he explicitly expresses his shock at what Beth has said. In response, Beth first categorically denies that what she has just said is a "trick" (turn 10), something which Neil does not appear to agree with (turn 11). Indeed, Neil's rhetorical question in turn 11 has the perlocutionary effect (Austin 1962) of encouraging Beth to be open about what is upsetting her. Accordingly, in turn 12, Beth tells Neil that she needs him "to stop being nice to me unless you're gonna marry me after". This is so unexpected and confusing, and falls outside the immediate contextual information that is available to Neil, that he is unable to reply but instead chuckles (turn 13).
Given the ongoing contextual information that is available to Neil, what Beth is saying is totally unexpected or, rather, is remotely available to him. To Neil, being (implicitly) told by his partner that hanging the painting on the wall is a/ an (unnecessary) demonstration of "being nice" is a bit too shocking. From Neil's perspective, what makes Beth's interpretation of his attempts to hang the painting on the wall "being nice" even more unexpected, is Beth's explicit request for Neil, in the same turn, to "marry" her after being nice to her, i. e., after hanging the painting on the wall.
However, viewing the situation through Beth's eyes, the above events can be interpreted rather differently. In the context of Beth and Neil's long romantic relationship and despite Beth's personal wish, Neil has never asked her to marry him. Beth thus considers, in a seemingly deliberate way, that Neil's action of hanging a painting on the wall implies that he is just being nice to her without any long-term commitment. In actual fact, it is against the backdrop of this information, which is not readily available to Neil but is actively brought in by Beth in a radically top-down manner, that such an inference might become defeasible in the sense of being occasioned one-sidedly by Beth. Indeed, it is the active presence of such "free" pragmatic processes (see Recanati 2010Recanati , 2012, although not necessarily triggered by any (linguistic) material that may become available throughout the course of the immediate interaction, which modulates the dynamics of the interaction with a view to deducing an "inference of [im]polite [or rather offensive] intentions", without being "inextricably attached" to the (con)textual material in question (Brown 2001: 11623). Arguably, what Beth thinks or wants to be implicated is done "by virtue of an inference" (Recanati 2004: 6) squeezed out of this interaction.
We now return to the remainder of the conversation. Neil's chuckle, which might indicate his lack of (at least immediate) interest in the idea of marriage, forces Beth to retort in turn 14 by asking a rhetorical question, in which she expresses her feelings of frustration and/or anger. Neil, who by this time is even more confused by the turn of events, does not reply, forcing Beth to repeat the rhetorical question (turn 16), and this time Neil replies by saying that "it's not funny" (turn 17). Evidently, Beth is not merely insisting that her partner puts on record that what she has just said is not funny. She adopts this strategy to ensure that she clarifies the point she is trying to make, with a view to persuading Neil to marry her (cf. Zhang and Parvaresh 2019). With this aim in mind, in turn 18, she explains why she has made this request, with her explanation being based on the fact that they had been "together for over seven years", thereby implying that it was time they got married. In the same turn, Beth even goes so far as to claim that if Neil does not marry her, then their relationship is over. In response, Neil claims that it is not an either/or situation and suggests a third possibility, namely that he does not believe in the institution of marriage (turn 19), implying that they could live happily together indefinitely, without marrying. This greatly angers Beth, and she describes his comments as "bullshit" (turn 20).
As the above conversational exchange demonstrates, "what is said/done" involves inferences whose content is creatively adjusted by the interactants. To extrapolate from this position, we could claim that Beth's feeling of "being in a relationship for seven years and not being asked to marry her partner" impacts on the entire conversation, in a radically top-down manner, helping her to actively adjust a situation by using her state of, e. g., "feeling bad", as an excuse to get something done, i. e., to encourage her partner to marry her.

Concluding remarks
From its earliest beginnings, impoliteness literature has paid particular attention to the role of the "speaker" and the strategies that the speaker can use to cause offence (e. g., Culpeper 1996) as well as the speaker's intention and role in purposefully issuing a face-threatening act (e. g., Bousfield 2008). While research on the relational (e. g., Locher and Watts 2008), social (e. g., Kádár 2017; Horgan 2019, 2020), cultural (Sharifian and Tayebi 2017), interactional (e. g., Haugh 2015a, 2017) and contextual (e. g., Terkourafi 2001;Culpeper et al. 2003) dimensions of language has provided us with a wealth of insight into the topic, there are many occasions in which taking of offence takes place without any agreement whatsoever between the so-called offendee and the offender. Recognition of the fact that impoliteness is an evaluative moment (Eelen 2001;Kádár and Haugh 2013) and the many research studies that have investigated this issue are important steps towards explaining the notion of taking offence, but it appears that a knowledge gap still exists regarding the role of the hearer in this process, particularly when one takes into account the fact that such evaluations are often "fundamentally subjective" (Ruhi 2008: 289). It is exactly these fundamentally subjective evaluations, which not only have a tendency to vary across groups of speakers (Parvaresh 2019;Parvaresh and Tayebi 2018), but also across identifiable individuals in locally situated interaction (cf. Mitchell and Haugh 2019), which arguably makes this line of research a particularly thorny one. As Haugh (2017: 282) by drawing on Saul (2002) notes, the distinction between "what the hearer thinks the speaker implicated" and "what the speaker thinks he or she implicated" needs to be investigated more systematically. Indeed, it has been demonstrated that multiple inferences can be made from "what is said (or not), how it is said, […] when it is said (or not)" (Haugh 2017: 286), and even from what is unsaid (Parvaresh and Tayebi 2018).
The current study is an attempt to provide an account of the notion of taking offence in contexts that involve identifiable individuals in more locally situated interactions. Motivated by some of the tenets of radical contextualism, in this study we have attempted to explain how and why the hearer actually engages in such an evaluative process. More precisely, we have tried to elucidate the (pragmatic) processes which appear to be at work when the hearer either assigns a particularly impolite or offensive meaning to an utterance during the course of an interaction, or adjusts the entire interaction in ways that are not necessarily envisaged by the speaker. As the discussed examples demonstrate, the hearer can go very far in assigning an offensive meaning, or intention, to what he/she has heard, or rather to the interaction in question. The hearer seems to achieve this by engaging in a purely top-down process of meaning making. This enables the hearer to cause a certain offensive meaning to prevail in ways which are not related to immediate (con)textual information. In this respect, recent findings in pragmatics have also demonstrated that "the gap between propositions and natural-language sentences is much broader than is standardly assumed" (Wilson and Carston 2019: 36).
The explanations provided here contribute to present-day attempts to explain the variability involved in the taking of offence. However, given the fact that the data used in this study, especially those taken from films, might be slightly different from purely naturally occurring conversations, more research is certainly needed to further explore the claims made in this study. Future studies could therefore investigate in some more detail, and by using more natural data, the variability involved in taking offence and whether and how the claims put forward in this paper could explain this phenomenon. Another potential area for research would be one in which researchers explore so-called inherently impolite language alongside and with reference to the hearer's role in evaluations of offence.
A fall in tone. , Continuing intonation. ! An animated tone. ?
A rising tone. (.) A half-a-second pause.
(1) A pause in seconds. (( )) Additional information added during transcription. // A current speaker's utterance is overlapped by that of another speaker on the next line. = No pause between turns.