The anatomy of a conspiracy theory in Covid-19 political commentary

The pandemic has exacerbated moral panics about conspiracy theories. Yet de ﬁ ning what conspiracy theories are is just as fraught as ﬁ guring out what to do about them. This article provides the ﬁ rst empirical demonstration of how the categories ‘ conspiracy theory ’ and ‘ conspiracy theorist ’ are used in social interaction. We examined comments from a New Zealand politician about a Covid-19 outbreak at the start of the election period. Using conversation analysis, membership categorisation analysis, and discursive psychology, we tracked how his talk was built and interpreted by participants. The ﬁ ndings show how a conspiracy theory was made recognisable through the machinery of storytelling and how its status as a conspiracy theory was accomplished and challenged through categorisation. We argue that conceptualising conspiracy theories as social actions offers a way to move beyond de ﬁ nitional debates to examine how participants understand and use conspiracy theories in everyday life. (Conspiracy theory, social interaction, categorisation)

The pandemic has exacerbated moral panics about conspiracy theories. Yet defining what conspiracy theories are is just as fraught as figuring out what to do about them. This article provides the first empirical demonstration of how the categories 'conspiracy theory' and 'conspiracy theorist' are used in social interaction. We examined comments from a New Zealand politician about a Covid-19 outbreak at the start of the election period. Using conversation analysis, membership categorisation analysis, and discursive psychology, we tracked how his talk was built and interpreted by participants. The findings show how a conspiracy theory was made recognisable through the machinery of storytelling and how its status as a conspiracy theory was accomplished and challenged through categorisation. We argue that conceptualising conspiracy theories as social actions offers a way to move beyond definitional debates to examine how participants understand and use conspiracy theories in everyday life. (Conspiracy theory, social interaction, categorisation)

I N T R O D U C T I O N
The Covid-19 pandemic has exacerbated moral panics about the supposed increase in 'conspiracy theories' 1 and the dangerous consequences of people believing and acting upon them. For example, 'conspiracy theories' about the cause and spread of coronavirus have been linked to arson attacks on 5 G telecommunication towers (Ahmed, Vidal-Alaball, Downing, & López Seguí 2020), deliberate violations of lockdown orders (Gruzd & Mai 2020), and lower uptake of public health measures like hand-washing or mask-wearing (Imhoff & Lamberty 2020). Yet despite widespread news coverage, increased academic attention, and shared cultural understandings, we know surprisingly little about HOW people recognise and label social actions as 'conspiracy theories' in their everyday lives. This research presents a novel analysis of how talk comes to be understood as a 'conspiracy theory' in social interaction.
Debate about the nature of conspiracy theories, how they should be investigated, and what to do about them characterises both popular media and academic spheres (Grodzicka & Harambam 2021). One strand of research examines conspiracy theorists, asking what kinds of people are more likely to believe conspiracy theories and the effects of holding such beliefs (van Prooijen & Douglas 2018). This work largely operates with realist definitions of conspiracy theories as explanations for causes of social or political events involving secret plots and powerful cabals (Douglas, Uscinski, Sutton, Cichocka, Nefes, Ang, & Deravi 2019). However, realist definitions can be pathologizing, housing the assumption that believing conspiracy theories is false, irrational, and harmful (Coady 2018) and positioning those who believe conspiracy theories as paranoid radicals on the fringes of society (Boullier, Kotras, & Siles 2021).
Another strand of research examines conspiracy theories, asking how they are related to other forms of reasoning, such as scientific knowledge (Dentith 2018;Boullier et al. 2021), how they are constructed through recurring themes (Byford 2011;Boullier et al. 2021), and why they are appealing (Saglam 2020). Some of this work operates with relational definitions of conspiracy theories as ideas that challenge mainstream knowledge and officially sanctioned truths (e.g. Boullier et al. 2021), while other studies draw attention to the politics involved in defining and labelling conspiracy theories (Husting & Orr 2007). The term 'conspiracy theory' has no fixed meaning, but the label functions to 'narrow the range of acceptable opinion and restrict the terms of acceptable debate' (Coady 2018:166). There are affective and moral dimensions to this restriction, as the politics of labelling restrict what can be said, known, thought, and felt (Husting 2018).
Despite recognition of the power of labelling, there is a paucity of research that examines how the labels 'conspiracy theory' and 'conspiracy theorist' come to be applied to people and actions. Determining what counts as a conspiracy theory is a political act. But the process of recognising or labelling action is a fundamental part of everyday social interaction. Rather than using 'conspiracy theory' as an analytic category and wrestling with the meaning of conspiracy theories in abstract terms, this article examines how conspiracy theories are built through the turn-by-turn unfolding of social interaction. This approach foregrounds participants' understandings of what counts as a conspiracy theory. The analysis is informed by ethnomethodological ideas about the intelligibility and accountability of social action. Garfinkel's (1967) key insight was that shared methods for building and recognising actions provide the machinery and moral logic of social interaction and social life. Although conspiracy theories evade definitions in the academic literature, participants in interaction have shared methods for building and recognising certain actions as conspiracy theories. This article presents the first empirical documentation of those methods.
The fields of conversation analysis and discursive psychology draw on ethnomethodology to document how social actions are organised in social interaction and how psychological matters are implicated in the morality of everyday social life. Describing people and actions are fundamental building blocks for getting things done in social interaction and relating with others. Although seemingly 2 Language in Society (2022) straightforward, these descriptions carry profound consequences for how people understand themselves and others. One basic fact is that any description is selective. The same person or action can be described in almost infinite ways, but people select descriptions at particular moments to accomplish different goals in social interaction. For example, describing someone's action as intentional is a way to attribute responsibility and allocate blame. Describing the same action as unintentional can absolve responsibility and deny an accusation. In contexts like police interviews, these different descriptions can carry heavy sociopolitical and legal consequences (Auburn, Drake, & Willig 1995;Edwards 2008).
Describing people with categories or labels can activate a host of shared common-sense meanings, providing shorthand judgements about who they are and what they are like (Schegloff 2007). Certain actions are linked to categories, so that simply describing an action can invoke someone's category membership (Sacks 1972). For example, the action of voicing a conspiracy theory is closely linked to the category conspiracy theorist. Thus, describing someone's action as a conspiracy theory can invoke their identity as a conspiracy theorist. Yet these identity labels can also be resisted, contested, and negotiated in social interaction (Hester & Eglin 1997;Stokoe & Edwards 2009).
Given the link between identity and action, and the way that 'conspiracy theorist' operates as a derogatory category (Byford 2011), voicing a conspiracy theory can be a difficult social action to accomplish. Similarly, describing an extraordinary or supernatural experience can lead to judgements about the speaker and their character. Speakers manage these risks by sequentially organising their stories to first describe their innocuous interpretation before describing their realisation of the extraordinary or supernatural situation (Wooffitt 1992;Jefferson 2004). This sequential organisation allows speakers to present themselves as reasonable and ordinary rather than predisposed to seek out the unusual.
The analysis below documents how a New Zealand politician built a recognisable conspiracy theory through turns of talk in a press conference, and how he and others negotiated how to interpret, categorise, and judge his actions. Tracking the sequential organisation of the conspiracy theory and its subsequent categorisation reveals how participants manage issues of identity, intention, reasonableness, and motive.

D A T A A N D M E T H O D
We analysed a naturally occurring moment in New Zealand's media landscape where participants used the label 'conspiracy theory' to make sense of a politician's conduct. The conspiracy theory was first published in an online press release from the centre-right opposition National Party where deputy leader Gerry Brownlee called on the "Government to come clean on what they know about the state of Covid-19" (see Figure 1). The press release was issued in the same day as Director General of Health Ashley Bloomfield warned that a resurgence of Covid-19 was Language in Society (2022) 3 likely (Andelane 2020), despite the successful elimination of the virus earlier in the year (Baker, Wilson, & Anglemyer 2020). Although the press release contains rhetorical features characteristic of conspiracy theories (Byford 2011), the analytic focus is on how conspiracy theories are deployed and interpreted in the turn-by-turn unfolding of social interaction.
We tracked how the conspiracy theory was built in social interaction and interpreted over subsequent days. The first data extracts are drawn from a press conference at Parliament on 12 August 2020, where opposition leader Judith Collins called for the election to be delayed due to the recent Covid-19 outbreak, and deputy leader Gerry Brownlee was questioned on his comments in the press release. The next data extracts are drawn from an interview the following day where Collins was questioned about her deputy's conduct on state broadcaster Radio New Zealand. The final extract comes from a commercial radio interview on 14 August where Brownlee denied being a conspiracy theorist.
We accessed a public video recording of the press conference from Radio New Zealand and audio recordings from Radio New Zealand and Newstalk ZB websites. By repeatedly listening and viewing recordings, we produced conversation analytic transcripts that captured features such as overlapping talk, intonation, and speech delivery and features of embodied conduct for the video recording (Hepburn & Bolden 2017).
We used conversation analysis and discursive psychology to empirically document how the conspiracy theory was accomplished and made sense of in social interaction. Conversation analysis provides a systematic method for analysing the  Language in Society (2022) sequential organisation of naturally occurring social interaction (Sidnell 2013). We used this method to document how the conspiracy theory was built to be recognisable across turns of talk and sequences of action. The analysis also attends to when and how participants described Brownlee or his actions. We used membership categorisation analysis (Sacks 1972;Stokoe 2012) to identify how descriptions of action invoked inferences and category memberships and how these descriptions were implicated in different trajectories of action (e.g. blaming, denying, etc.). Thus, rather than a priori categorising Brownlee's talk as a conspiracy theory, the analysis documents how this category term was demonstrably relevant and procedurally consequential for participants in the interaction (Schegloff 1997). The analysis also draws on the discursive psychology concept of stake (Edwards & Potter 1992;Potter, Edwards, & Wetherell 1993) to examine how psychological matters such as motive, intention, and social responsibility were invoked by participants as they attempted to make sense of Brownlee's actions.

A N A L Y S I S
In two sections below, we document how the conspiracy theory was accomplished and interpreted by participants. The first section demonstrates how the conspiracy theory was constructed through the sequential organisation of storytelling and initially challenged by recipients. The second section shows how the action of conspiracy theorising was ascribed and resisted through practices of categorisation.

Constructing and challenging a conspiracy theory through narrative
This first section demonstrates how Brownlee used the sequential organisation and causal logic of storytelling to construct talk that was subsequently interpreted and challenged as a conspiracy theory. The first two extracts come from a press conference that was held on the day Parliament was due to dissolve for the general election. However, new cases of Covid-19 were announced the previous day, the city of Auckland entered lockdown, and the opposition party called for the election to be delayed. Extract (1) occurs nineteen minutes into the press conference. Party leader Judith Collins (JC) is at the podium, but a reporter (RP) directs the next question to deputy leader Gerry Brownlee (GB) by referencing his comments in the earlier press release. The analysis highlights how Brownlee's answer is constructed in a rhetorical style characteristic of conspiracy theories (Byford 2011). (2022 have you been wearing masks and will you be wearing masks?

Language in Society
The reporter's first turn displays an understanding that Brownlee's earlier press release was concerned with the government's knowledge. The suggestion that the government should "come forward, on what they actually knew" (lines 3-4) suggests the government is holding back information and knows more than what they have claimed. Brownlee's comments in the press release are thus animated here by the reporter, providing Brownlee with an on-the-record slot to respond to them in the interactional context of the press conference. The reporter suggests Brownlee may feel "vindicated" (line 11) by the recent outbreak, drawing a connection between Brownlee's earlier statements and the current situation.
Brownlee resists the description "vindicated" but takes the opportunity to respond to the topic of his earlier statement. He describes what he was doing as "having a look" at "some pretty obvious signs" (lines 15-16), characterising his actions in innocuous and reasonable terms. Byford (2011) identified that concerns with foreknowledge and the rhetoric of just asking questions are recognisable features of conspiracy theories. Here Brownlee questions why the government was concerned in the previous week and describes himself and others as "wondering what do the health authorities know that they're not-not fully explaining" (lines 33-34), while criticising the withholding of information by describing it as being "dribbled out" (line 24) in official briefings.
Yet the actions of questioning and criticising the government are not solely the preserve of conspiracy theorists. These actions are also a fundamental part of the role of an opposition politician (not least in an election period!). Thus, although Brownlee's turn makes inferences available that have resonances of conspiracy theory, there is no evidence at this point that participants understood his talk in that way. The next question is asked by another reporter who raises the topic of masks and the press conference moves on.
However, two minutes later, Brownlee returns to the question and expands his answer. These remarks were interpreted as a conspiracy theory in subsequent media coverage (e.g. Grieve 2020; Rindelaub 2020). The analysis attends to the sequential unfolding of his turn and his recipients' responses in the turns which follow.
(2) RP: reporter; JC: Judith Collins; GB: Gerry Brownlee (RNZ-PC 22:11-23:33) The press conference is characterised by a turn allocation scheme where participants orient to journalists' rights to ask questions and politicians' obligations to respond to them (Clayman 2013). Thus, Brownlee's request to "make a, comment" (line 1) about the earlier question treats his action as a deviation from normative speaking arrangements and marks his turn as noteworthy. He begins his comment with the assessment "I just think it was interesting," (line 2). Initial assessments tell listeners how to evaluate the talk that follows (Mandelbaum 2013). Before explaining what he found interesting, Brownlee begins an if-then construction that ends incomplete, "if you think about what was said" (lines 2-3). The second pronoun "you" works here to address both the reporter and the overhearing audience. The unstated upshot is that those listening will draw similar conclusions. Likewise, the shift from the first person singular "I" to the plural pronoun "we" in "what we've observed" (line 3) attributes the observations (and possible inferences) to a collective rather than to Brownlee alone.
Brownlee uses the sequential machinery of listing and storytelling to link six events. Starting with "one" (line 4) indicates a list of events which, by virtue of being grouped together, can be understood as members of the same category (Jefferson 1991). The first event is messaging about a possible resurgence of Covid-19. Brownlee marks the timing of this event as relevant, pausing the progressivity of talk to replace "a couple of weeks ago" with "about ten days ago in fact" (lines 5-6). Changing the unit of measurement from weeks to days increases the temporal proximity between the event and the current moment (C. Raymond & White 2017). (2022)   9 Events within the structure of a story are designed to be heard as logically (or causally) linked to one another (Mandelbaum 2013). Thus, directions to purchase masks, the prime minister's visit to a mask factory, and Dr. Bloomfield's choice to get tested are not presented as discrete events, but as connected to one another. One aspect of connection is their relative timing. Bloomfield's test is marked as noteworthy given it occurs "after a hundred and (.) three days" of no Covid-19 cases in the community (lines 16-18). Formulating the length of time in this way treats Bloomfield's action as inapposite and remarkable, inferentially raising the question: why now? At the same time, Brownlee's embodied conduct-a slight headshake, raised eyebrows, and smile (see extract (2), images 2a-2d)-provide a non-verbal assessment of the events he describes. The description of the final event-a return to lockdown-occurring "a matter of hours" later (line 21) highlights that it is the close temporal relationship between events that is the relevant-and questionableaspect of Brownlee's narrative. More explicitly than extract (1) then, Brownlee builds a narrative that suggests a hidden truth needs to be uncovered in a recognisably conspiratorial style (Byford 2011).

Language in Society
In the next turn, the reporter's question displays an understanding of the hidden truth hinted at in Brownlee's narrative. The question begins with "so" (line 24), targeting the upshot of Brownlee's turn and seeks clarification about "what are you saying there then?" (line 24). No answer is forthcoming as Brownlee leaves the podium and the reporter pursues an answer by reformulating the question (line 27). The reporter's questions contain the presupposition that Brownlee was saying something more than what is self-evident. Participants orient to some turns as containing multiple actions, which Sidnell (2017) distinguished as vehicular and derived actions. Here, the reporter treats Brownlee's turn as a vehicle for an as-yet-unspecified action and asks him to make it explicit.
Brownlee's response challenges the presuppositions of the reporter's question. He begins his response with "oh" (line 31), a practice that treats the question as unexpected and inapposite (Heritage 1998). He then describes his conduct as "just outlining facts" (line 31). This description asserts the vehicular action and challenges the presupposition that there was any derived action to be specified (cf. Sidnell 2017).
The reporter does not accept Brownlee's description of his action, pursuing an alternative explanation with "but but what do you mean by that?" (line 32). Asking 'what do you mean' can target a moral dimension of a problem with a prior turn (G. Raymond & Sidnell 2019), and here the reporter pursues the derived action-or hidden meaning-of his turn. As evidence of this derived action, the reporter formulates a possible upshot for Brownlee to confirm or deny. That the reporter can suggest he is "saying it's not a coincidence that they knew" (line 33) is evidence that these inferences were available from Brownlee's talk. Here, possible foreknowledge of Covid-19 in the community is a logic that connects the description of events as a causal narrative. Foreknowledge is a 10 Language in Society (2022) characteristic of conspiracy theories (Byford 2011) and renders Ardern's and Bloomfield's actions morally culpable.
Brownlee neither confirms nor denies this interpretation, instead repeating the evaluation "interesting" (lines 35-36). As we examine further in the final extract, Brownlee distances himself from explicitly asserting the conspiracy theory, while nonetheless pursuing it in a tacit way. Party leader Judith Collins (who has now taken the podium) challenges the basis of the reporter's questions, asserting the media's responsibility to ask questions but redirecting responsibility to answer them. We pick up on the issues of responsibility and examine Collins' response further in the extracts which follow.
Ascribing and denying 'conspiracy theory' through categorisation Brownlee's comments were widely interpreted as a conspiracy theory in print media in the days that followed the press conference (e.g. Grieve 2020; Rindelaub 2020). This section examines two radio interviews where hosts used the categories 'conspiracy theories' and 'conspiracy theorists'. The analyses track how interviewers and interviewees negotiated how to categorise and interpret Brownlee's actions.
The day after the press conference, Judith Collins was interviewed by Kim Hill (KH) on Radio New Zealand. The interview largely concerned the party's calls for the election to be delayed given the recent outbreak. Yet, a recurrent line of questioning from Hill concerned the intelligibility and accountability of Brownlee's conduct the previous day. Across the four extracts from the interview, we examine how Hill and Collins negotiated how to make sense of what had happened and who was responsible.
Extract (3) comes from the start of the broadcast.
( Hill's description of the previous day provides context to the audience and sets the scene for her discussion with Collins. Hill describes Brownlee's actions as "obliquely" (line 2) suggesting Ardern and Bloomfield "held back information" (line 4) about the outbreak. Much like the reporter's question in extract(2), Hill's description displays an understanding that Brownlee's turn made inferences available without stating them directly. By characterising the day's events as "political" (line 1) and using the categories "deputy leader" (line 1) and "National party leader" (line 5), Hill invokes Brownlee and Collins' stake as politicians. Invoking someone's stake or interest is one way to undermine their actions (Potter, Edwards, & Wetherell 1993). Thus, Hill suggests that their actions were motivated by political interest.
Hill offers Collins an opportunity to clarify "what you're accusing Doctor Bloomfield and Jacinda Ardern of?" (lines 9-10), using the second person pronoun "you" to encompass both Collins and Brownlee. But Collins resists the characterisation of her action as an accusation. She flatly denies any accusation against Bloomfield (line 11) and characterises her action towards Ardern as "simply (.) pointing" to a document on pre-election government conduct (line 12). This description of her conduct rejects the derived action of an accusation and asserts the vehicular action. Much like Brownlee's description in extract (2) ("just outlining facts"), the activity of pointing something out is unobjectionable. Thus, Collins resists the moral inferences associated with an accusation by formulating her conduct as reasonable.
In extract (4) Hill's question is evidence that Brownlee's remarks were recognisable as a conspiracy theory. The plural pronoun "we" (line 1) positions Hill within a collective who witnessed Brownlee's conduct and suggests the understanding of his remarks as conspiratorial is a shared reality. In paraphrasing Brownlee, Hill repurposes the machinery of narrative to link Ardern's visit, the advice to buy masks, and the announcement of Covid-19 in the community. Hill displays an understanding that it was the relative timing of these events that Brownlee found interesting and that his turn was purposefully built to allow his listeners to draw conspiratorial inferences.
Hill directs the question "what did he mean by that?" (lines 7-8) to Collins, displaying an understanding that she can and should answer on his behalf. Categorising Brownlee as "your deputy" (line 1) invokes Collins' membership as party leader and her category-bound responsibility to account for the actions of party members. Yet Collins challenges the presupposition that she is responsible for Brownlee's actions. Her well-prefaced turn treats the question as inapposite (Schegloff & Lerner 2009) and asserts "you should ask him!" (line 10). Collins marks that Hill's question should properly be addressed to Brownlee, invoking the commonsense position that speakers are individually responsible for their actions (cf. Sacks 1984) and that matters of meaning or intention are private internal states owned by individuals (Edwards 1997). Thus, in this question-answer sequence, Collins challenges the suggestion Brownlee was espousing a party line-for which she could be presumed to have knowledge and authority over-and instead positions him as solely responsible for his actions.
Hill picks up on the issue of responsibility in extract (5). The analysis tracks how participants use intentionality to make sense of action and attribute responsibility.  Hill's line of questioning displays an understanding that Collins and Brownlee acted intentionally to advance a conspiracy theory about government cover-up. Describing actions in different ways invokes different forms of accountability, and here Hill's description of "having collaborated" (line 3) is linked directly to a "responsibility for a mood of fear and panic" (line 2). Rather than answer the question, Collins assesses it as "a ridiculous thing to say" and begins to formulate "what we're saying" (line 8). Before Collins finishes speaking however, Hill asks a new question targeting the party's calls for transparency (lines 11-12). As in the previous extracts, asking "what does that me:ean::." (line 12) targets this action as accountable and in need of explanation (cf. G. Raymond & Sidnell 2019).
Hill provides Collins with an opportunity to confirm or deny the conspiracy theory. One way to make sense of calls for transparency is a belief that "information is being withheld from the public." (line 18). Collins does not answer the question, instead asserting that "people need to know" how the latest outbreak occurred (line 20). Hill orients to this as topic shift, re-centering the conversation on the conspiracy theory that "they know something .h we don't." (lines 24-25) and issuing another yes=no interrogative, "are you suggesting that," (line 25). Rather than a yes=no answer, Collins responds with a multi-unit turn that is recognisable as a complaint about other occasions when government information has been inaccurate.
Thus, across multiple questions, Hill provides Collin with an 'on the record' slot to clarify her (and her party's position) with relation to Brownlee's remarks. As is characteristic of political interviews (Clayman 2013), Collins avoids directly answering the question in attempts to advance her own topical agenda.
In extract (6), Hill explicitly uses the label 'conspiracy theory' Collins resists this categorisation as they negotiate how to make sense of the actions of the previous day.  Hill invites Collins to assess whether she "mispla:yed [her] ca:rds" (line 1), using the metaphor of a card game to formulate her actions as strategic choices. This action is directly tied to her stake when Hill categorises her "as the leader of the national pa: rty" (line 2), invoking her categorical identity and the associated stake in discrediting the government in an election period. Hill claims she "allowed" Brownlee to "dog whistle conspiracy theories" (lines 3-4), suggesting she had prior knowledge and gave permission for his actions. The metaphor of the dog whistle suggests Brownlee's conduct was designed to be understood as a conspiracy theory by members of some groups while appearing innocuous to others. Hill thus positions Collins as knowledge and culpable in using conspiracy theories as a political strategy.
Collins resists the yes=no formulation of the question, instead asserting that Brownlee "was simply stating um some of the: the information" (lines 5-6). This neutral description of action challenges Hill's invocation of stake by asserting the vehicular action and denying a hidden meaning or derived action. Yet Hill challenges this characterisation, pursuing clarification about what information could suggest a government cover-up.
Rather than answer the question, Collins' well-prefaced turn points out that "he hasn't actually said that has he?" (line 15). Using "actually" highlights the difference between what happened and Hill's characterisation of events. The tag question "has he?" shows that this understanding is mutually available, establishing a shared position about the events that both have witnessed. Accountability is attached to action (Sidnell 2017), so people cannot reasonably be held accountable for things they have not done. Thus, Collins suggests Hill's line of question is unreasonable as it is based on a spurious characterisation of what happened. Extract (7) comes from a radio interview the following day on commercial station Newstalk ZB. The extract comes from the beginning of the interview where the host Chris Lynch (CL) invites Brownlee to account for his actions. Both he and Brownlee use the category 'conspiracy theorists', but distance Brownlee from this categorisation. (2022 Lynch's opening question to Brownlee presents categorically organised actions and responsibilities. Letters have been sent by "scientists and health professionals" (lines 2-3), two occupational categories with expertise and authority about the pandemic. The act of sending a letter to "all politicians" (line 3) displays an understanding of politicians' category-bound responsibility to take advice from experts. The letter urges politicians to act against their category-bound stake "to put politics aside" (line 4) and avoid politically motivated actions including scaring the populace or scoring points against their opponents. By asking Brownlee if he received the letter, Lynch categorises him as a member of the group 'all politicians' and invites him to account for himself in terms of the descriptions 'scaremongering' and 'point-scoring'.

16
Language in Society (2022) Brownlee displays an understanding that his conduct in the press conference was understandable as the kind of negative political behaviour referenced in the letter. The indexical reference "the other day?" (line 9) displays an assumption that his listeners know which day and which events he is referring to. Although he takes responsibility for having "got" into a "bad spot" (line 9) the nature of the transgression is not further specified. He displays an understanding of scaremongering as an intentional action. Thus, by denying the intention "to create any fear" (line 10), he denies this description for his actions. Instead, he characterises his action as "simply ah asking questions" (line 14). The word 'simply' asserts the vehicular action and denies a derived or hidden meaning (Sidnell 2017). Asking questions is an entitlement of any reasonable person and asking questions "in the house" (line 12) is a politician's category-bound entitlement.
Lynch's second question mobilises the category 'conspiracy theorist'. However, in contrast to Hill's description of strategically dog whistling conspiracy theories (extract (6)), Lynch's description aligns with Brownlee's stance on intentionality. The suggestion that Brownlee "played into the hands" (line 22) suggests that Brownlee may have unintentionally supported conspiracy theorists without believing it himself-a key attribute that would categorise him as a conspiracy theorist.
Brownlee's answer likewise distances himself from the category of conspiracy theorist. In the course of question-answer sequences like these, journalists and politicians reflexively constitute mainstream and marginal political positions (Clayman 2016). Here, both Lynch and Brownlee constitute conspiracy theories as marginal and illegitimate. Lynch uses the definite article "the" (line 22) to categorise conspiracy theorists as a group of people located "out there" (line 23) who are OTHER to host, guest, and audience. Likewise, Brownlee's assertion "there is no credence in (.) conspiracy theories" (lines 31-32) positions this as an extreme position in the sociopolitical landscape and one he does not endorse (cf. Clayman 2016).
Although Brownlee formulates conspiracy theorists as a problem to be "defeat [ed]" (line 25), his proposed solution bears remarkable similarities to his original comments and reiterates the party's calls for greater transparency. Thus, he balances the dilemma of appealing to centrist voters by disavowing the legitimacy of conspiracy theorists, while appealing to more partisan voters by reiterating his tacit criticism of the government. In this way the issue of potentially withheld information remains live, even as the participants jointly position conspiracy theories as politically illegitimate.

D I S C U S S I O N
This article has documented how the category 'conspiracy theory' was ascribed and resisted in a naturally occurring social interaction. The case study analysed remarks made by Gerry Brownlee, deputy leader of the opposition, after an outbreak of  in New Zealand at the start of the 2020 election campaign. Brownlee's remarks were widely reported as a conspiracy theory in news media (Grieve Language in Society (2022) 2020; Rindelaub 2020). The analyses tracked how Brownlee's talk was built to be recognisable as a conspiracy theory-and how the categorisation of his actions as a conspiracy theory was ascribed and resisted in the moments and days that followed.
The findings are distinctive for documenting the practices participants use to build talk recognisable as a conspiracy theory. We found that Brownlee used the sequential machinery of storytelling to link otherwise discrete events, implying a causal connection without explicating it directly (extract (2)). The psychological matters of knowledge and motive were inferentially available to make sense of the story and to recognise Brownlee's remarks as an accusation of a government cover-up. The sequential organisation of the conspiracy theory bears similarities to the practices documented by Jefferson (2004) and Wooffitt (1992) in their studies of extraordinary experiences or supernatural encounters. Although his talk made inferences available that were voiced by recipients, Brownlee maintained plausible deniability to describe his action in innocuous ways and resist categorisation as a conspiracy theorist.
The analysis documented how participants oriented to the categories 'conspiracy theory' and 'conspiracy theorist'. These category terms were used differently in two radio interviews that followed the press conference. In the Radio New Zealand interview, Hill displayed an understanding that Collins and Brownlee knowingly invoked a conspiracy theory for political gain (extract (6)). In the Newstalk ZB interview, Lynch displayed an understanding that Brownlee had unknowingly invoked a conspiracy theory, unwittingly furthering the agenda of "conspiracy theorists" (extract (7)). Psychological research suggests that people primarily share conspiracy theories because they believe them to be true (van Prooijen & Douglas 2018). Yet the question for participants was what Brownlee had done and what he was accountable for. Hill, Collins, Lynch, and Brownlee debated the psychological matters of motive and intention, yet the issue of belief was not relevant as a way to make sense of the conspiracy theory in this case.
In evaluating Brownlee's conduct, both radio hosts invoked his category membership as deputy leader of the opposition party. Making this aspect of his identity relevant provides resources to understand what he did and why. As an opposition politician in an election period, Brownlee had a vested interest or stake in discrediting the government. Participants displayed understandings that Brownlee was acting as a politician. Understanding his conduct in this way allowed for judgements about the party's political strategy and the effects of such tactics at a time of crisis. However, in his defence, Brownlee presented his actions as the warranted activity of raising questions and holding the government to account. In a pandemic era with sweeping state powers, the role of opposition politicians is arguably more important than ever. Different identity categories (e.g. politician, conspiracy theorist) connect to different actions (e.g. asking questions, spreading conspiracy 18 Language in Society (2022) theories) and make available different moral judgments about motive and reasonableness.
In negotiating the label of conspiracy theory, participants oriented to the limits of reasonable debate. Press conferences and political interviews are spaces where participants negotiate the limits of marginal or legitimate positions (Clayman 2016). Previous research has documented how conspiracy theorising is perceived as antithetical to 'proper' democratic politics and positioned as 'beyond the limits of respectable enquiry and legitimate political dissent' (Byford 2011:23). Our analyses show that dynamic at work. Both Lynch and Brownlee described conspiracy theorists "out there" (extract (7)) at a distance from themselves and the presumed audience. Brownlee's explicit condemnation of conspiracy theorists likewise drew a line between acceptable political criticism and conspiracy theorising. Nonetheless, his actions show that this is in fact a permeable boundary that participants can invoke and negotiate at different moments across interactions with others.
The pandemic has reconfigured life in New Zealand. The March-May 2020 lockdown measures were widely supported by the public, but the August 2020 outbreak and the start of election campaigning saw increasing concern about conspiracy theories. Media reporting roundly condemned Brownlee's comments and positioned conspiracy theories as unacceptable from a leading political figure (Rindelaub 2020). When the election was held in October, Jacinda Ardern's Labour Party achieved a sweeping majority victory. The National Party was resoundingly unsuccessful, and Brownlee lost the electorate he had held for the past two decades. Although some news outlets saw the election result as a rejection of conspiracy theories (Walters 2020), it appears both conspiracy theories and coronavirus will remain fixtures of local and international life.
Conspiracy theories raise challenging questions of media responsibility, healthy democracy, and individual freedoms. We argue for further study of conspiracy theories in use. Although the data analysed here is an instance of political communication, it is not just powerful figures like politicians who use conspiracy theories. Conceptualising conspiracy theories as social action offers a way to move beyond concerns with belief to examine when and how conspiracy theories are used in social interaction, how they are understood, and what they accomplish. In this case study, the conspiracy theory allowed Brownlee to voice criticism of an enormously popular government during an election period. Yet even when the stakes are not so clear, conspiracy theories can be used to accomplish social actions such as persuading, justifying, or influencing. Further research on the interactional accomplishment of conspiracy theories can better document how they are used and understood by participants. Grounding research in participants' orientations to conspiracy theories is one way to avoid moral panics about conspiracy theories by Language in Society (2022) 19 examining how they are actually used in practice and how they matter to those who hear them.

GB
speaker identification for verbal conduct gb speaker identification for embodied conduct ?
falling intonation [ ] overlapping talk = talk continues across lines . , talk faster than surrounding speech , .
talk slower than surrounding speech (()) transcribers' comments () undetectable speech underline emphasis : sound-stretch (1.0) silences timed in seconds (.) silence less than a tenth of a second sound cut-off .h in-breath $ smiley voice ↑ raised pitch UPPER CASE louder speech°°q uieter speech italics translation of non-English words @ @ start and end of JC's embodied actions Δ * + start and end of GB's embodied actions → embodied action continues across lines # where the screenshot occurs relative to talk N O T E S 1 'Conspiracy theory' is a widely used but contested category. We use the term as a broad way of referring to the phenomenon under investigation without taking a stance on legitimacy or truthfulness.
2 Transcription conventions are given in the appendix.