2. Imagination-Based Accounts

Several philosophers have recently converged on the view that conspiracy theorists hold some kind of imaginative attitude toward their theories. Ichino (Reference Ichino, Engisch and Langkau2022) argues that conspiracy narratives are fictional stories in Walton’s (Reference Walton1990) sense: like all fictional texts, they are props in a game of make-believe, a game we play by imagining the text’s contents are true. Similarly, Ganapini (Reference Ganapini2022) argues that conspiracy narratives are akin to religious myths and parables whose literal contents one takes to be false, but from which one extracts a message (as the myth of Icarus imparts a moral about hubris, a conspiracy narrative might convey a moral about how corrupt Democrats are). Munro (Reference Munro2023) argues that many conspiracy theorists are engaged in a fantasy of possessing “secret knowledge” that the general population lacks. And Levy (Reference Levy2022) argues that conspiracy theories are a form of “serious play”: people merely pretend they are true, but they mistakenly take themselves to believe them.

While the details of each account differ, all are built up from the more basic idea that conspiracy theorists merely imagine rather than genuinely believe. These philosophers appeal to similar lines of evidence to support this.

For one thing, conspiracy theorists often do not seem responsive to evidence the way someone who believes would be. No one is perfectly rational, so we do not always perfectly proportion our beliefs to our evidence. Still, our beliefs are typically revised to some extent in the face of that which we take to be evidence.Footnote ² And the evidence for many conspiracy theories seems very weak, far-fetched, or non-existent (Rosenblum & Muirhead, Reference Rosenblum and Muirhead2019). To most observers, it seems doubtful that many people would believe in response to such transparently bad evidence. But imagination offers another interpretation: since what we imagine is not sensitive to our evidence the way beliefs are (cf. Currie & Ravenscroft, Reference Currie and Ravenscroft2002, 15–16), this leaves open the possibility that conspiracy theorists are merely imagining instead.

It’s not just that many conspiracy theorists aren’t straightforwardly guided by evidence; it’s also that they are guided by other, non-evidential factors. For example, people engage with conspiracy theories because they have certain desires: to be special or unique (Douglas & Sutton, Reference Douglas and Sutton2018; Imhoff & Lamberty, Reference Imhoff and Lamberty2017; Sternisko et al., Reference Sternisko, Cichocka and Van Bavel2020) or to find a sense of community and belonging (Douglas et al., Reference Douglas, Sutton and Cichocka2017; Phadke et al., Reference Phadke, Samory and Mitra2021; Sternisko et al., Reference Sternisko, Cichocka and Van Bavel2020). It’s easy to see why people with such motivations would be attracted to a community of like-minded conspiracy theorists who share a claim to possessing special, secret knowledge that the general population lacks. But these are desires, not evidence, which supports the idea that they are not truly believing.

There’s also the fact that conspiracy theorists seem to be guided by the entertainment value of conspiracy theories instead of by evidence. Entertainment value is positively correlated with how likely people are to endorse a theory (van Prooijen et al., Reference van Prooijen, Ligthart, Rosema and Xu2022). Conspiracy theories are also often shared and developed online in a way that involves a lot of humor—trolling, memes, irony, and the like. Examples like QAnon are also highly gamified: the way people piece together clues to unlock hidden messages resembles one big puzzle-solving game (Davies, Reference Davies2022; Levy, Reference Levy2022; Dutilh Novaes, Reference Dutilh Novaes2024).

Another line of evidence concerns how attitudes towards conspiracy theories guide behaviors. When pressed about their reasoning, many conspiracy theorists do not respond the way we’d expect if they believed. As Rosemblum and Muirhead (Reference Rosenblum and Muirhead2019) note, they often instead hedge their claims, saying things like, “I’m just asking questions” or “well, it’s possible the story is true.” Furthermore, as Mercier (Reference Mercier2020) observes, conspiracy theorists rarely take action that seems proportionate to the horrific stories they claim to believe. For example, while a few people took action based on Pizzagate—one who brought a gun into the restaurant at the centre of the story (Menegus, Reference Menegus2016), another who tried to bomb a public monument (Winter, Reference Winter2019)—most are content to simply engage online with other, like-minded storytellers.

Finally, there’s the simple fact that conspiracy theories often resemble fictional stories in various ways. They feature exciting plots involving secret coverups; resistance movements; epic, global conflicts between good and evil; and whistleblowers fighting to expose the truth. This resemblance to engaging fictional narratives is taken as evidence that people treat conspiracy theories as fictions, too (see especially Ganapini, Reference Ganapini2022; Levy, Reference Levy2022).

So, there’s a lot of convergent evidence for the claim that some conspiracy theorists merely imagine or pretend rather than truly believing. However, we think it goes too far to treat cognitive attitudes like fantasy, pretense, or make-believe as central to or paradigmatic of what it is to be a conspiracy theorist.Footnote ³ This view is unsatisfying because it fails to give a unified account of the wide range of ways people engage with conspiracy theories, including the heterogeneity of cognitive attitudes amongst them.

Most importantly, there’s the fact that some conspiracy theorists simply do believe the stories they claim to believe. As Munro (Reference Munro2023) argues, it’s difficult to deny that those willing to undertake extreme actions in the name of a conspiracy theory are true believers. This includes the examples of violent Pizzagate proponents mentioned above, and we see similarly drastic actions from proponents of other theories. Consider the followers of Romana Didulo, a QAnon-inspired conspiracy influencer who claims to be the true Queen of Canada. Didulo has issued drastic decrees to her followers, some of whom take them quite seriously. Her order to arrest cops led to violent confrontations with police (Lamoureux, Reference Lamoureux2022b); her decree that water and electricity are now free in Canada led some to stop paying their bills, resulting in their utilities being shut off (Lamoureux, Reference Lamoureux2022a). These sound like actions of people who truly believe in her cause.

Of course, philosophers defending imagination-based accounts do not claim they apply to every conspiracy theorist; they acknowledge there may be some genuine believers. However, on their accounts, these true believers would be anomalous or outliers. It’s hard to determine the exact proportion of true believers versus game-players among people who have obvious motivation to obscure their commitments from distrusted outside researchers. But we can at least say with confidence that both types are likely present within any given conspiracy community. Which forces an intriguing question: what are the norms governing a community comprised of both sincere inquirers and witting fabulists? Answering that question will consume the rest of this paper. For the moment, our point is simply that imagination-based accounts cannot easily explain such dynamics. This suggests it is worth looking for a more unified phenomenon that includes both and so is more explanatorily powerful.

This becomes even clearer once we recognize that things aren’t so black-and-white as believing versus merely imagining. Instead, there are plausibly many who fall somewhere in between, believing some plot points of a conspiracy theory while pretending others are true. Rosenblum and Muirhead (Reference Rosenblum and Muirhead2019) emphasize the way conspiracy theorists often avoid framing their narratives as straightforwardly true, instead framing them as “true enough.”Footnote ⁴ Consider an incident from 2016, after Donald Trump repeated the National Enquirer’s claims of a connection between Senator Ted Cruz’s father and JFK assassin Lee Harvey Oswald; when he later faced skepticism from campaign aides, Trump reportedly responded, “Even if it isn’t totally true, there’s something there” (Rosenblum & Muirhead, Reference Rosenblum and Muirhead2019, 28). Similarly, in a 2020 Civiqs poll that asked about beliefs in QAnon, 16% of respondents said that “some parts are true” (Civiqs, 2020).

Admittedly, it’s difficult to know exactly how to interpret these sorts of claims and poll results. Again, we should not expect complete transparency from people who distrust outside inquirers. However, it seems plausible that, while many conspiracy theorists do not believe every plot point of a narrative like QAnon or Pizzagate, they at least believe some core elements. One sign of this is the way specific points of evidence appear to be fungible. For instance, Pizzagate theorists have repeatedly claimed that journalists who try to debunk their theory have been arrested on child pornography charges. Though they point to real arrests, each time it has turned out that the arrested person was not in fact the journalistic antagonist they claimed. In one case, the person arrested merely had a similar name to an anti-conspiracy writer.Footnote ⁵ This pattern suggests a fixed set of core beliefs (someone must be part of a pedophilic conspiracy) attached to a darkly playful indifference to verifying specific factual claims.

This example points to a broader phenomenon concerning the complex roles of truth and evidence in constructing conspiracy theories. No conspiracy theorist is completely indifferent to truth, as we might expect if they were merely engaged in make-believe or fantasy. It’s not as if they simply invent fictional clues and evidence for their theories. That would be a way of constructing a purely fictional, make-believe story (much as Arthur Conan Doyle has Sherlock Holmes build a theory of a murder case based on fictional evidence found at a fictional murder scene). Instead, conspiracy theorists typically appeal to a few genuine facts as “evidence” for their theories. These are often coincidences that mainstream narratives allegedly cannot explain (what Keeley, Reference Keeley1999 calls “errant data”). For example, QAnon supporters point out real coincidences between Trump’s tweets and Q’s posts on 4chan, which they use as “evidence” the two are in cahoots. And John Podesta’s leaked emails really do talk about eating pizza, which Pizzagate supporters use as “evidence” that Democrats are trafficking children in a pizza restaurant.

In Section 5, we’ll consider how to characterize the relationship between conspiracy storytelling and truth more precisely. For now, the key point is that there seems to be something more complex going on than merely constructing a fiction or work of make-believe. Instead, conspiracy theorists go to great pains to meticulously incorporate certain facts into their narratives (cf. Clarke, Reference Clarke2002). At the same time, it’s true that the resulting stories often have elements that look quite fantastical, and we agree with proponents of imagination-based accounts that many people do not believe all of these elements. What’s missing is an account that accommodates this mix of belief and mere imagining.

The limitations of existing accounts stem from starting with the individual psychologies of conspiracy theory participants. Given the evident diversity of these psychologies, it is difficult for such an account to retain a unifying explanatory structure. To go beyond these limits, we should start from the other side—from the macro scale rather than the micro. We need to ask what these groups are, via the social practices that bring them together.

To be fair, it’s not that proponents of existing accounts do not give analyses of the broader social practices in which individual conspiracy theorists participate. Levy (Reference Levy2022), for example, argues that conspiracy theorists participate in a kind of collective gameplay, while Ganapini (Reference Ganapini2022) argues that they imaginatively construct narratives that play important roles in group formation and identity. Still, these analyses arrive at views about the social function of conspiracy theories by first analyzing individuals’ attitudes.

Our analysis will move in the opposite direction. We’ll start by analyzing conspiracy theorizing as a social practice. Specifically, we’ll argue that it involves an activity of storytelling governed by certain social norms. Beginning at the social level will then allow us to explain the range of behaviors and cognitive attitudes amongst individual conspiracy theorists.

3. Norms of Storytelling

We will not give precise necessary and sufficient conditions for terms like “narrative” and “story.” Instead, we follow Currie’s (Reference Currie2010) view that these concepts involve loose clusters of properties, where texts have greater degrees of narrativity when they exhibit more of these properties. These properties include portraying the actions of one or more agents; representing concrete, spatiotemporally related events; and having a high degree of thematic unity.

Non-narrative texts can exhibit many such properties, too (e.g., mere causal explanations can also describe agents and spatiotemporally related events). However, there are certain norms to which storytellers are typically sensitive that are especially helpful for distinguishing paradigmatic narratives from non-narratives.

As we’ll conceive of them, the norms of storytelling are irreducibly social norms.Footnote ⁶ Social norms are patterns of actual social practice (as opposed to theoretical norms of idealized rationality). Here, they have to do with our normative expectations about the stories other people tell us. What features do we look for in good, successful instances of storytelling?

There are many kinds of stories. There’s the purely fictional, the purely nonfictional, and those involving a mixture of truth and fiction (e.g., historical fiction weaving real events with fictional plot points, as in Hillary Mantel’s influential Wolf Hall (Mantel, Reference Mantel2009)). Some are written by professional storytellers, while others are told during casual conversations. Given this diversity, are there nevertheless social norms that apply to storytelling in general? This section pinpoints two such norms. First, good stories involve unexpected turns of events. Yet—second norm—these events nevertheless follow familiar patterns of emotional release. At first glance, it might sound like these are in tension, since they prescribe both unexpectedness and familiarity. However, we’ll argue that, taken together, they provide the essence of a good story. We draw on two lines of argument: empirical evidence from the cognitive science of storytelling and arguments from the philosophy of narrative.

We’ll start with the unexpectedness norm in relation to narratives about events that really occurred. Analyses of conversations in naturalistic settings reveal that such narratives take up a large proportion of everyday chitchat (Dessalles, Reference Dessalles2021). Moreover, conversational participants expect storytellers to describe events which are somehow novel, abnormal, or otherwise unexpected (Agar, Reference Agar2005; Dessalles, Reference Dessalles2021; Ochs et al., Reference Ochs, Taylor, Rudolph and Smith1992; Schank, Reference Schank1979).

Reflecting on examples shows this is quite intuitive. Suppose a colleague walks into your office and causally says: “The weather was sunny and warm on my way to work this morning, as per usual this time of year; I got to my office at about the same time I arrive everyday.” Or, suppose the story was: “I caught a student plagiarizing on a paper last week. I followed university policy and reported it to the plagiarism committee. They got back to me within their usual timeframe, recommending the standard penalty for first-time offenders: a 0 for the assignment.” In both these cases, a natural response is to say, “So, what?” The way each story unfolds is quite predictable, with nothing unusual, novel, or surprising.

This normative expectation fits well with a widespread view about the evolutionary function of storytelling, or the purpose for which our capacity to share stories evolved. On this view, storytelling evolved as a means of sharing knowledge about actual, past events—about agents, their interactions, and the states of affairs they brought about (Boyd, Reference Boyd2018; Currie, Reference Currie2010; Dessalles, Reference Dessalles2021; Dor, Reference Dor2017; Dunbar, Reference Dunbar2004). It makes sense we’d prefer stories about the novel or unexpected: we want new information that adds something to our overall store of knowledge about the world. Learning that my colleague’s walk to work was unremarkable, or that the plagiarism committee behaved predictably, adds nothing useful.

So, we seem to apply an unexpectedness norm to nonfictional stories. Furthermore, it’s typically assumed that our ability to tell fictional stories evolved from the prior, more basic ability to tell stories about nonfictional events (see especially Currie, Reference Currie2010; Dor, Reference Dor2017; Boyd, Reference Boyd2018). If fictional storytelling reuses the same cognitive tools as nonfictional storytelling, it would make sense if our natural tendency to apply a norm of unexpectedness carries over.

This does seem to be the case. Imagine watching an episode of the crime series Law and Order. Suppose a dead body is discovered in the first 30 seconds, and that police soon set their eyes on the most obvious suspect: the victim’s spouse who has a clear motive, whose fingerprints are on the weapon, and who seems unable to provide a convincing alibi. From here, imagine two diverging plots. In the first, it’s quickly made clear through a flashback that the obvious suspect really is the murderer; he goes through a mundane trial and is convicted based on clear evidence. In the second, a twist is revealed halfway through the episode: police learn the obvious suspect does have an alibi he was hiding to protect the true perpetrator; they now have to find and build a case against the new suspect.

The first narrative evokes a similar feeling to hearing your colleague tell a mundane story where nothing unexpected happens: in virtue of its predictability, it violates our expectations for good storytelling. Because the second narrative contains unpredictable twists and turns, we view it as a better story.

So, we seem to apply a norm of unexpectedness to storytelling in general, whether fictional or nonfictional. However, as further examples make clear, there are constraints on this norm. The most effective way to maximize unexpectedness is to make the events of a narrative as random and disconnected as possible. Stories that are too random, though, feel unsatisfying. This goes for nonfictional narratives: it would be strange for a colleague to suddenly tell you, without additional context or explanation, “When I was 10 years old, I was bitten by a rattle snake; two years ago, the Red Sox won the World Series in a thrilling, last minute comeback; my younger brother unexpectedly passed away last week.” Each of these events is surprising, but they are too unrelated to make for a good story.Footnote ⁷ The same goes for fictional narratives: suppose the killer in our Law and Order episode turned out to be a character introduced right before the end, with no connection to the preceding events; we’d be frustrated and think the story was too unpredictable.

It seems what we want from a story is to be surprised, but only by something that ultimately fits together coherently. The philosophical literature on narratives aims in part to characterize this “fitting together.” On one prominent view, the events of a narrative merely need to be causally connected to one another (Carroll, Reference Carroll and Carroll2001). However, as David Velleman (Reference Velleman2003) points out, the mere fact that a chain of events is causally unified (“A caused B, B caused C…”) is not sufficient for explaining why certain events feel like appropriate endings of stories. A description of a causally related series of events could go on indefinitely, until we arbitrarily cut it off. But the mere presence of causal connections does not mean that choosing any arbitrary cut-off point results in a clear, overarching story arc and plot resolution.

Besides causal connections, events can also be linked through what Velleman calls a “narrative explanation.” Velleman builds his account on the work of literary theorist Frank Kermode, particularly Kermode’s highly influential book The Sense of an Ending (Kermode, Reference Kermode1967). Kermode analyzes the ways the craft of fiction, and narrative generally, relies on emotional anticipation. Velleman borrows from Kermode the central metaphor of a ticking clock. If an old-fashioned grandfather clock is ticking away in the hallway outside your office, you will acquire a tacit expectation for the sequence of sounds: Tick-tock, tick-tock. First the tick, then the tock, again and again. If you suddenly hear only the Tick, but not the Tock, your expectations will be violated in a way that draws your attention and may cause you to feel unsettled.

According to Velleman-following-Kermode, narrative structure works similarly. The emotions aroused in an audience by a story’s early events—the Ticks—set up expectations for later resolution of those emotions—the Tocks. A good story is one that matches its Ticks and Tocks appropriately. As Velleman puts it, “a description of events qualifies as a story in virtue of its power to initiate and resolve an emotional cadence in the audience” (Velleman, Reference Velleman2003, 18).

Velleman illustrates this with an example from Aristotle.Footnote ⁸ A man has killed someone named Mitys. Later, the killer is standing below a statue depicting Mitys when the statue suddenly topples over, crushing the murderer beneath his victim’s representation. Importantly, Velleman suggests, the feeling that this is a fitting ending does not require assuming the man’s murder of Mitys caused that same man’s death-by-statue, not even by some sort of ghostly revenge. It’s enough that the later event is an emotionally fitting sequel—the right Tock for this Tick—in that it resolves the emotional anticipation set up at the story’s beginning.

While causal explanations aim at modeling some real fact about the world (objective causal relations between events), narrative explanations are a form of subjective sense-making, through which we mentally subsume events under patterns of emotional arousal and resolution. The point of Kermode’s grandfather clock metaphor is precisely that clocks do not actually have Ticks and Tocks—these are explanatory categories we impose on auditory perception to organize the anticipation generated by regularity. The same, according to Kermode, is true of all narrative. To call any one event the beginning or ending of a story is arbitrary from the point of view of the universe—but not from an emotionally valenced human point of view.

Famously, the TV series The Sopranos concludes with an episode that piles emotional openings on top of one another, seemingly promising a maelstrom of affective resolution in some thematically perfect ending. Instead, the show suddenly cuts to black in media res, no resolution ever to come. This violation of narrative logic might be a commentary on the moral deserts of the central character, or a deliberate subversion of the pressures of crafting a satisfactory ending to a widely loved story. Whatever the writerly intentions, the ending is widely remembered precisely because it pointedly refuses to do what endings are supposed to do: give our emotions somewhere to go. It’s the exception that proves the rule.

Importantly, the idea is not that the events presented in a narrative are totally familiar or anticipated in advance—that would make for a boring story. Again, good stories also involve surprising or unexpected events (the first norm for storytelling identified above). But rather than being totally random, they follow coherent, familiar patterns of emotional arousal and resolution (the second norm). To bring together these two norms: good stories involve what we’ll call non-obvious Tocks. They describe a series of unpredictable events which nevertheless conclude on emotionally appropriate Tocks, i.e., endings which resolve emotions aroused by earlier Ticks.

Featuring non-obvious Tocks may not be necessary for being a narrative (recall how The Sopranos intentionally subverts these norms). When a text does conform to norms for good storytelling, though, this seems like an especially strong marker of narrativity. Superficially narrative-like activities, such as building mere causal explanations, typically do not involve the kinds of emotionally meaningful twists and plot arcs described in this section.

4. Conspiracy Stories and the Norms of Storytelling

We now argue that conspiracy theorizing is constrained by the norms described in the previous section. A good story links events in a way the audience would not have expected and yet yields a convincing emotional cadence. Popular conspiracy theories do both: they surprise the audience by revealing events omitted from the “official” account while simultaneously affirming the priority of human meaningfulness in selecting events as beginning and end. In other words, they are full of non-obvious Tocks.

Let us start with the unexpectedness norm. Suppose you want to explain what happened on 9/11. Here’s a relatively obvious explanation: a terrorist group that made public its intentions to attack the US succeeded in doing so, by attacking a building which had previously been targeted by a similar group nine years earlier. Qua narrative, this explanation is somewhat unengaging. That’s because it’s relatively unsurprising: it implicates the obvious suspects and accords well with how things already appear on the surface, given what was reported by the media and US government. (To be clear: the attack, of course, was a very surprising event. But this explanation of it is not.)Footnote ⁹

In contrast, conspiracy storytelling involves attempting to dig beneath initial, surface-level appearances (cf. Basham, Reference Basham2001). This generates a more surprising, unexpected narrative. Take the narrative that 9/11 was really carried out by the CIA, who used Al-Qaeda as a scapegoat. This more closely resembles a Law and Order episode in which the initial, obvious suspect turns out to be a red herring. That makes it more narratively engaging, relative to the more predictable official account.

Similarly, take the story of “Queen” Romana Didulo. It’s not much of a compelling narrative to say that things were exactly as they appeared in Canadian politics circa 2023: Justin Trudeau was Prime Minister, as he had been for some years; and King Charles III was Canada’s head of state, having succeeded his mother Queen Elizabeth II after her death from natural causes. Compare this with Didulo’s narrative: she is the true Queen of Canada; she was appointed by the true King of the US in gratitude for leading a secret mission to expel the Chinese from secret tunnels underneath Canada; Queen Elizabeth was really executed to make way for Didulo; and Didulo is now fighting against deep state operatives like Trudeau, who are oppressing Canadians through means like forcing them to get dangerous COVID vaccines. This again clashes with the surface appearances.

Yet the events of conspiracy stories aren’t totally random. They build and release emotional tension in ways that are familiar from other exciting stories about political intrigue and coverups. The 9/11 conspiracy story starts with a tragic event; the twist that it was really the CIA stokes righteous anger that a government would do this to their own citizens; and we start to feel a sense of resolution as we learn of heroic whistleblowers uncovering these deeds and bringing them to light, aiming to hold the government accountable. Similarly, Didulo’s story starts by evoking feelings of unease about how the deep state is increasingly controlling Canadian citizens and violating their rights to bodily autonomy; we feel a sense of relief, though, when a figure as heroic as Queen Romana swoops in to fight these injustices.

Conspiracy theorizing thus seems to conform to the social norms for storytelling we identified in the previous section. With this in mind, we propose that we can see it as a social practice of storytelling. On this account, conspiracy “theories” aren’t, strictly speaking, theories in the sense of being explanations of the sort constructed by scientists—rather, they are narratives produced via processes of participatory storytelling.

As we saw in Section 2, the communities that construct conspiracy stories are psychologically heterogenous. Some participants merely imagine that conspiracy stories are true; these participants treat the stories they tell as fictions. But others are true believers who do not regard these stories as fictional. Our storytelling account unifies this diversity of motives. To claim a group is engaged in the activity of storytelling is neutral about whether a particular story is true or believed by everyone: one and the same story could be fully believed by some, while others take it to be purely fictional or believe only parts. Despite this, the group can be unified in that they are conforming to the norms of storytelling, which apply to both fictional and nonfictional stories. The members of a conspiracy theory community are thus engaged in a unified social practice, even if there’s heterogeneity in individual cognitive attitudes.

Our view also provides a new perspective on some of the observations used to support imagination-based accounts of conspiracy theorizing. Consider the fact that conspiracy theorists seem to get so much entertainment from their stories. Initially, this might seem like evidence that they are engaged in mere imagination or pretend play. However, entertainment is not reserved for fictional stories and games of pretend. We can also be entertained by nonfictional stories (e.g., documentaries) and stories involving a mix of fiction and nonfiction (e.g., historical fiction). Our storytelling account can thus account for the entertainment value of conspiracy stories, while allowing that some conspiracy theorists believe and some merely imagine.

Something similar goes for the observation that conspiracy stories resemble fictions in the plot points and tropes they incorporate (secret plots, coverups, resistance movements, good versus evil, underdog whistleblowers, etc.). On closer reflection, these are tropes that appear in popular fictional and nonfictional stories, since they make both kinds of stories engaging. Entertaining true stories from throughout history feature such tropes, as evidenced by how many have been adapted into plays and films—from Shakespeare’s Julius Caesar to films about intrepid journalists like Spotlight or The Post. Secret conspiracies allow both fiction and nonfiction to conform to the unexpectedness norm, while the triumph of an underdog, or of good over evil, allows stories to come to emotional resolutions. We can thus explain why conspiracy stories incorporate plot points like these, while allowing that conspiracy theorists could treat them as either fictional or nonfictional.

To further clarify the shape of our account thus far, it helps to point out how it resists certain conceptual distinctions sometimes drawn in philosophical literature on conspiracy theories. Some philosophers distinguish between conspiracy theories (a type of explanation of events), conspiracy theorizing (the activity of developing conspiracy theories), and conspiracy theorists (those who engage in conspiracy theorizing) (e.g., Duetz & Dentith, Reference Duetz and Dentith2022). Under our account, we cannot define “conspiracy theory” in a way that’s conceptually prior to the relevant activity or what it means to participate in it. If conspiracy theories just are narratives produced via a particular kind of participatory storytelling, understanding the nature of conspiracy theories themselves requires understanding the nature of the activity and how people engage in it.

Now, some conspiracy stories aren’t narratively completed wholes but instead concern ongoing events, where the final “tocks” have yet to be revealed.Footnote ¹⁰ We can accommodate this by recognizing that larger, over-arching narratives often contain sub-narratives—think of an ongoing novel or TV series in which each installment ends with some degree of narrative resolution but also serves to set up the continuation of the ongoing, larger narrative. Along these lines, Romana Didulo’s overarching narrative starts with a sub-narrative about successfully expelling Chinese operatives from tunnels beneath Canada; this sets up a “sequel” in which she is appointed Queen of Canada and courageously fights against vaccine mandates.

There’s an important phenomenon our account so far has not fully explained: Section 2’s observation that conspiracy theorists—even the witting fabulist kind—try very hard to incorporate some truths into their narratives, such as (allegedly) unexplained coincidences. Explaining this will require looking more closely at the nature of conspiracy storytelling, to which we turn in the next section.

5. Alethic Games

We’ve talked about norms for storytelling that apply to both fiction and non-fiction. But much philosophical work on conspiracy theories has been preoccupied with whether conspiracy theorizing is an activity that’s necessarily inconducive to truth. Quassim Cassam (Reference Cassam2019) argues that it is, to firm opposition from philosophers like Charles Pigden (Reference Pigden1995) and MRX Dentith (Reference Dentith2023). It might seem strange that we appear to be side-stepping this point entirely.

Remember, though, that our goal is to characterize a single social practice in which both witting fabulists and true believers are engaged. We should not expect to find a single answer to how these different people regard the alethic status of the theories they construct. A more nuanced and multi-layered analysis is required. To that end, this section argues that, while true believers are guided in equal measure by a search for truth and the norms of storytelling, witting fabulists treat truth as a merely optional constraint that works in service of good storytelling. Each can thus work alongside one another in service of a single, overarching storytelling project, despite some diversity in individual aims. So, truth does play some role in conspiracy theory, but it is only a secondary role, in that it is not treated the same way by all participants. Viewed at a social level, the primary function of conspiracy communities—what unifies witting fabulists and true believers—is to conform to the norms of storytelling.

It will help to start from the fact that both true believers and witting fabulists engage in genuinely taxing intellectual labor. Communications scholars Alice Marwick and William Partin, examining hundreds of QAnon forum posts, concluded that “Q adherents demonstrated wide-ranging knowledge about current and past events, questioned the legitimacy of sources or the conclusions that other participants came to, and, above all, admonished others to ‘do the research’ or ‘think critically’” (Marwick & Partin, Reference Marwick and Partin2022, 13). In other words, this was no passing lark; QAnon participants hold each other to demanding standards of evidence production and interpretation. Of course, outsiders might regard these methods as intellectually bankrupt. But the point is only that some QAnon participants work hard at their theory.

This might suggest that most QAnon participants are true believers. But that’s too quick an inference, since intellectual work can also be a platform for playfulness. Consider the popular boardgame “Mysterium.” Players take on the roles of psychics investigating a series of murders, reliant on cryptic visual clues delivered by an all-knowing but non-speaking player (the ‘ghost’). The ‘psychic’ players work together, debating how best to assemble the clues into a coherent overall picture, while the ‘ghost’ must think carefully about how and when to deploy their limited supply of oblique hints. This is an intellectually demanding exercise, subject to conventional epistemic norms of coherence and evidence-updating. But it is also fun, and of course the participants are well aware that they are jointly constructing a fiction.

So, we cannot immediately infer from the intellectual demandingness of conspiracy theories like QAnon that all participants are in earnest. Indeed, there is strong evidence that QAnon began life as an elaborate game of play-pretend. Journalist Dale Beran’s book It Came from Something Awful (Beran, Reference Beran2019) documents the dyspeptic online discussion forums that spawned a series of off-the-wall conspiracy theories. First came Pizzagate, in which “Reality-challenged kids and adults collaborated with nihilistic trolls to play the internet like a video game.” According to Beran, “The goal was to make a role-playing game out of the internet by e-stalking anyone associated with the topic and digging through social media and public records for ‘clues’ to build the fiction” (Beran, Reference Beran2019, 219).

QAnon launched as a Pizzagate spinoff, appearing in the same places, possibly written by the same people, and carrying on the same central narrative of major Democratic politicos as hidden pedophiles. In October 2017, the first Q posts appeared on the 4chan /pol/ (‘politics’) forum, added by someone claiming to be a deep cover government agent with ‘Q-level’ security clearance. But this ‘QAnon’ was only one of several tongue-in-cheek government agents active on /pol/ at the time: there were also the likes of ‘FBI Anon’ and ‘White House Insider Anon’, daring each other to top their latest outrageous claim. One alleged, for instance, that Thomas Jefferson was secretly known to the US government to have been a foreign spy and traitor.Footnote ¹¹

For these early users, this was all an immersive game, a jointly constructed fiction. The ‘Q’ character was simply the most committed to the bit—other forum users referred to him as “that larper guy,” meaning LARP, or Live Action Role-Playing (Beran, Reference Beran2019, 221). But soon the stories spread beyond 4chan, spilling onto mainstream social media and the evening news, where an increasing proportion of the audience were true believers. A 2020 survey by Pew Research found that 47% of Americans had heard of QAnon, and 20% of these respondents thought it was “good for the country” (Michell et al., Reference Michell, Mark Jurkowitz and Shearer2020, Section 3). Eventually, some participants took the theory so seriously that it motivated them to attack the US capitol in January 2021.

The key point, for our account, is to notice that the rules for participating in QAnon remained more or less constant even as the community composition changed from exclusively knowing game-players to a growing share of true believers. The essential business of QAnon—interpreting cryptic posts by ‘Q’ in ways that expand a story about Democratic malfeasance—works the same for participants who treat the project as fictional game-playing akin to ‘Mysterium’ and for others to whom it is a deadly serious quest for truth.

To fully understand what’s going on here, we first need to say more about the sense in which truth plays different roles for different members of conspiracy storytelling communities. Start by distinguishing truth as a regulative ideal from truth as a side constraint. Among scientific communities truth is a regulative ideal, shared by all participants. Though data are never fully complete, and intelligible models must often smooth the edges of complexity (Elgin, Reference Elgin2017), the point of doing science is to approximate truth; that goal is what gives coherence to science as a distinct social community. As a regulative ideal, aiming at truth yields the norms and reasoning that guide the day-to-day conduct of science, shared (at least implicitly) by all participants.

Storytelling, in and of itself, does not aim at truth. But storytellers can take on truth as a side constraint. To see this, consider again Hilary Mantel’s trilogy of Thomas Cromwell novels, beginning from Wolf Hall. In her 2017 Reith lectures, Mantel set out how her literary goals cross paths with, but do not converge on, the aims of historical non-fiction. Most of her characters, from King Henry VIII on down, are real personages, and she took extraordinary care to avoid narrating anything that would conflict with the historical record. If the chronicles clearly say that Henry was at Windsor on June 7, then Mantel’s story places him there. Yet most dialogue and virtually all the inner life Mantel attributes to her characters are literary inventions.

If Mantel were a historian driven solely by the need to aim at truth, she would avoid these under-evidenced insertions. But her primary goal is literary, not historical. “For this reason,” Mantel explains, “some readers are deeply suspicious of historical fiction. They say that by its nature it’s misleading. But I argue that a reader knows the nature of the contract. When you choose a novel to tell you about the past, you are putting in brackets the historical accounts—which may or may not agree with each other—and actively requesting a subjective interpretation” (Mantel, Reference Lamoureux2017). Truth—fealty to the known historical record—is only a constraint in service of good storytelling. It is what makes these novels about real historical persons and events, rather than fantasy inspired by them. But it remains only a constraint, not a project-defining regulative ideal.

To see why this matters, contrast Mantel’s literary project with the genre of narrative non-fiction. Here, fealty to truth is more than just an artistic choice. One reviewer of journalist David Grann’s recent book The Wager, about a 1741 Patagonia shipwreck, describes Grann’s creative process like this: “One: unearth a tremendous story from within a forgotten haystack. Two: spend months and months and months researching it. Three: write the narrative with the artistry of a superb novelist” (McGoogan, Reference McGoogan2023). The reviewer intends this as high praise, not an impugnment of Grann’s commitment to truth. Grann’s prose serves two masters—the norms of storytelling and the demands of truth—in roughly equal proportion. This is quite unlike the historical novelist, who is first and foremost a storyteller and only secondarily—temporarily and voluntarily—takes on the demands of truth-telling.

Notice how this contrast roughly corresponds to the diverging motives of different members of conspiracy theory communities. True believers operate like writers of narrative non-fiction, aiming to seek truth and tell good stories at the same time. Witting fabulists, by contrast, are more like historical novelists. They will take truth-conducive epistemic norms as side constraints, like Mantel digging into Henry’s chronicles, but ultimately, it is only the norms of good storytelling that hold their deep affiliation.

Notice, then, that it is possible for true believers and witting fabulists to work together to construct the same conspiracy story, just as it is possible for narrative non-fiction writers and historical novelists to share research notes. At ground level, where epistemic norms are applied to granular items of data, the two may operate without any externally detectable difference (e.g., in the methods each uses to analyze evidence, such as Q’s online posts, to further fill out the overarching narrative). Indeed, the true believers may not even realize they are working alongside witting fabulists, so long as the latter aren’t too blatant in snickering and eye-rolling. The witting fabulists have just enough interest in truth (as a side constraint) that they can pass for true believers in short interactions, especially when mediated through the cloaking device of online communication.

This account provides an illuminating perspective on certain of the odder practices of conspiracy communities. Consider the case of the bamboo ballots. After Donald Trump lost the 2020 US presidential election, supporters in the Arizona legislature hired ‘auditors’ to seek evidence for his false claims of election fraud. One journalist observed auditors running ballots under microscopic cameras, apparently hoping to detect bamboo fiber in the paper (Levine, Reference Levine2021). This was inspired by a then-popular idea among online QAnon conspiracy discussions, claiming that 40,000 fake ballots had been imported from China. (In reality, the state Senate later issued a report finding no evidence of ballot fraud, while the ‘audit’ company shut down after failing to comply with court orders to turn over documents (Clark, Reference Clark2022).)

The bamboo-ballot theory illustrates the oblique role of truth in QAnon theorizing. There is a fact-y tinge to the claim; bamboo does, after all, grow in China, and not (in large quantities) in America. If the ballot paper did turn out to contain bamboo fiber, that could raise questions. In that sense, this was a reality-informed hypothesis in ways that claims about wizards and dragons are not. But of course, there is also a simple absurdity to the idea. If powerful forces really were trying to suborn an important election, would they not take the basic forensic precaution of having ballots printed with local paper stock? The expectation that the conspiracy could be unwound in such an elementary way better fits the narrative logic of CSI than serious scientific investigation.

Some QAnon adherents genuinely believed the bamboo-ballot theory, enough to motivate an actual (albeit amateurish) forensic inquiry. But it’s clear that many did not. Ken Bennett, an official responsible for overseeing the audit, told The Guardian he did not actually expect to find any bamboo. “I think that’s more of a euphemism for saying ‘we’re looking for everything related to the paper so that we can verify that the ballots are authentic,’” he said. “It’s just kind of a general expression for ‘we’re checking the paper and the folds and everything to make sure that these are authentic ballots’” (Levine, Reference Levine2021). The bamboo claim, then, made for a simple story that could bring together true believers, witting fabulists, and opportunistic chaos-mongers.

Consider also one of the details offered to support Canada’s QAnon variant. Some supporters of the self-proclaimed Queen of Canada, Romana Didulo, point out that her name is an anagram of the phrase ‘I am our Donald’, purportedly demonstrating her centrality to the QAnon universe (Lamoureux, Reference Lamoureux2021). Once again, this is a vaguely truth-related claim. ‘Romana Didulo’ is an anagram for “I am our Donald.” But what sort of evidence is that? The theory appears to suppose that when Didulo was born—in the Philippines in the 1970s—her parents looked out across the ocean to a then-28-year-old New York real estate speculator and, anticipating that their newborn would in five decades become the essential bridge in this man’s North American political triumph, named her accordingly.

It is hard to imagine that more than a handful of people take this claim seriously. But it is easy to see its appeal to bored internet denizens hunting for thematic resonance. Cryptic anagrams are a standard trope in paperback thrillers. What kind of conspiracy narrative would it be without such a clue hiding in plain sight? The flexibility of truth as a mere side constraint—anagrams and bamboo are real world things, though perhaps not the best evidence—allows for fluidity in the shared motivations of the true believers and the witting fabulists.

Notice how well this point fits our earlier observations about the norms of storytelling. Truth-constrained conspiratorial practices like anagram-hunting and bamboo-detection aim to provide the unexpected element that makes a good story. They also complete an emotional chime with earlier events, per the Kermode-Velleman account of narrative explanation. They are the non-obvious Tocks we described in the last section: bits of data that seem to come from out of the blue, yet make a curious sort of sense in narrative sequence. The statue of Mitys kills the killer of Mitys. The secret Queen of Canada is an anagram of the secret ruler of America.

It’s especially easy to observe these dynamics in recent, very public conspiracy storytelling like QAnon, but similar dynamics show up with earlier, seemingly more “traditional” conspiracy theories. Consider the way 9/11 conspiracy theorists appealed to rapper Notorious B.I.G.’s 1994 song “Juicy.” The lyrics briefly reference an earlier, 1993 bombing at the World Trade Center. But online sleuths have claimed they predicted 9/11, even suggesting the rapper was murdered by the US government for his involvement in planning the attack (Herbert, Reference Herbert2017). While this shows sensitivity to truth in some sense (he’s really from New York and his song does mention a bombing at the World Trade Center), it would strain credulity to think everyone making these outlandish claims believes them (even if a few do). It seems more likely that the 9/11 truther movement ranges from full-on true believers to those for whom truth is purely a side constraint. These sorts of coincidences boost the theory’s narrative power, suggesting narrative norms are really what binds people together.

6. Identity-Generating Fandom

Our goal so far has been to characterize the shared social practice of conspiracy storytelling, the unifying aim that allows true believers and witting fabulists to work together. In this section, we further elaborate on the glue that holds these different kinds of conspiracy theorists together, despite their diverse psychological profiles. What, exactly, motivates such a diverse group to converge on a single storytelling project?

Our central claim in this section is that conspiracy theory communities operate like fandoms. A fandom is a social group brought together by shared enthusiasm for a cultural product like a TV show or novel (see Jenkins, Reference Jenkins1992). Within a fandom, there is a body of information widely accepted as settled (‘canon’) and another body of claims subject to debate and theorizing. A fandom also has a set of distinctive norms governing how members will treat the subject matter (reverentially, subversively, etc.) and how they will treat each other. Importantly, there are fandoms organized around non-fiction as well as fictional products. ‘True crime’ fandom communities gather online to debate clues regarding real killings (see, e.g., Jones, Reference Jones and Larke-Walsh2023) and, in very rare cases, even turn up actual evidence.

To fully explore this point, we will borrow a concept from philosopher Christine Korsgaard. In Korsgaard’s framework, a practical identity is a way of conceiving of oneself that generates values and reasons for action; it is “a description under which you value yourself, a description under which you find your life to be worth living and your actions to be worth undertaking” (Korsgaard, Reference Korsgaard1996, 101). Some practical identities, like profession or parenthood, provide reasons to shape an entire life course. Others are more contextual and limited; being a moderately intense football fan, for instance, tells you how to schedule your weekend, but perhaps not how to live the next four decades.

Fandoms structure practical identities for their participants. Fans of the space opera Firefly have normative reason to protest the show’s abrupt cancellation and to campaign for the narrative closure offered by a long-promised movie. Fans of K-pop music acquire normative reasons to promote their idols’ songs in online polls, and also sometimes—when the fan community converges on it—to pursue seemingly extraneous aims, like interfering with police attempts to track Black Lives Matter protesters (Reddy, Reference Reddy2020). Individuals do these things in part because it’s just what other members of the fandom do.

We suggest that conspiracy theory communities function as identity-generating fandoms where the focal object is the conspiracy story. Remember that our goal is to explain how participants with a range of individual motives—some true believers, others witting fabulists—can converge on a shared set of practices. We suggest that the mutual fandom relationship provides exactly this, since it yields a set of identity-constituting norms that do not depend on whether the focal object of fanhood is fiction or non-fiction. Put another way: fandom is a way of operationalizing the norms of collective storytelling. As a shared project, those norms encourage a division of narrative labor. To explore this point, we will now distinguish three roles within conspiracy story fandoms. We will call them writers, characters, and players.

Writers are the originators of a conspiracy theory. They establish the shared information pool—“canon”—and set the community’s initial norms of inquiry. They are recognized by other participants as authoritative, at least to some extent. In some conspiracy communities, they may continue to exercise leadership, but in others, they cease to play an active role once the canon is established. In the QAnon story, the pseudonymous individuals behind the original 4chan accounts—QAnon, FBI Anon, etc.—are the writers. In this case, the writers knowingly constructed a fiction, but writers of other conspiracy theories may be entirely sincere.

Next are the characters: people who did not create the theory or establish its canon, but who can directly exert causal pressure on the theory through their appearance within the narrative itself. In QAnon, Donald Trump is a character; the narrative identifies him as the heroic leader of the resistance. Romana Didulo, the “Queen of Canada,” has made herself into another character. As these examples show, the motivations of characters may vary. Trump is clearly cognizant of his influence on QAnon; he sometimes calls out to Q-sign bearers at rallies or drops unsubtle allusions to their favored themes.Footnote ¹² But presumably Trump is aware that much of the QAnon lore is fantasy; he knows, for instance, that he was not still secretly running the American government in 2021, as many QAnon participants claimed.

Writers create the canon; characters can change it through actions that compel updates to the narrative. The final category, the players, are mostly passive with respect to canon. They cannot directly update the central shared pool of information. But—like members of other fandom communities—they can offer theories, interpretations, and out-of-canon secondary narratives about the characters. Players are by far the most numerous group; for large communities like QAnon, they may number in the millions. Among them are many true believers, the sort of people who show up at pizza parlors with assault rifles or storm the US Capitol.

This taxonomy—writers, characters, and players—allows us to see how the pieces of a conspiracy theory community link together. These are different practical identities orbiting the same narrative. Being a writer or character provides a different set of normative reasons than being a player, but all share in the same joint project. Together they are building a community centred on a pool of canonical information, arranged in compelling fashion per the norms of storytelling, which ultimately allows players to present themselves—sincerely or not—as crusading citizen journalists.

The idea of fandom also helps explain why witting fabulists are hesitant to openly admit they aren’t true believers. Many fandoms expect participants to keep up a façade that the object of their fandom is of grave importance, even for members who are self-consciously playacting. Some fans of a football team care deeply whether their team makes the playoffs, but others are just there to watch the game. Even the latter, though, could be accused of letting down their side if they made their casual allegiance too obvious. A good fan participates in even the silliest chants, perhaps with a healthy dose of internalized irony. The same is true for conspiracy fans: even—perhaps especially—witting fabulists should not admit anything less than total sincerity to outsiders.

We think this section’s analysis generalizes across conspiracy storytelling in the internet age, where online message boards, social networking, and YouTube provide platforms for fandoms to coalesce. Online fandom dynamics emerged as far back as 9/11 conspiracy theories. Figures such as Thierry Meyssan (author of 9/11: The Big Lie) and Dylan Avery (director of Loose Change) were the writers; various public figures have made themselves characters by amplifying and adding to the narrative, including fringe US political candidates such as Jeff Boss (Morin, Reference Morin2013) and former Iranian president Mahmoud Ahmadinejad (Siegel, Reference Siegel2011); and players further fueled the narrative through engagement on blogs, Facebook pages, and the like.

As per our introduction, our primary focus is on modern conspiracy theories constructed and disseminated via online communities. We lack space to consider in detail whether our account extends to pre-internet conspiracy theorizing, though we think at least some examples still fit well. Take the theory that Paul McCartney died and was replaced by an imposter, which first gained traction as a legend shared amongst students on college campuses (Tofel, Reference Tofel2021); it’s not hard to imagine that young fans collectively hunting for clues in Beatles lyrics and album covers were motivated primarily by what a juicy story this was, after which more and more true believers were born as the story spread to mainstream radio. If so, this dynamic was likely enabled by community gathering places—physical college campuses—in which a fandom could emerge, analogous to the role of the internet in present-day fandoms.

7. Implications

We conclude with some brief implications of our account, which we suggest may offer useful starting points for confronting deleterious conspiracy practices. We continue to remain neutral on whether all conspiracy theories are socially destructive or epistemologically pathological. But even philosophers who reserve judgment on conspiracy theories broadly (e.g., Dentith, Reference Dentith2023; Pigden, Reference Pigden1995) acknowledge that some are harmful. And everyone agrees that we need better tools for helping to draw people out of the most pernicious communities, which at extremes can resemble cults. Numerous recent headlines warn of people who have ‘lost’ family members to QAnon, for instance (e.g., Watt, Reference Watt2020).

Our account’s emphasis on the narrative element of conspiracy theory practice can help in two ways. First to theorists: we are now in a position to diagnose one way conspiracy theories might become pernicious. This problem occurs when people fail to distinguish between causal and narrative explanations. Velleman’s earlier work on narrative identified this as a danger of historiography generally. Sometimes, Velleman says, “the audience of narrative history is subject to a projective error. Having made a subjective sense of historical events, by arriving at a stable attitude toward them, the audience is liable to feel that it has made an objective sense of them, by understanding how they came about” (Velleman, Reference Velleman2003, 20).

Crucially, Velleman is not denying that narrative explanation has sense-making value, either in history or in human life. Rather, he is warning against the danger of unwittingly sliding between the two modes of explanation. We suggest that this may be one risk of pernicious conspiracy theory communities. Some members may initially join as witting fabulists. But as the narrative they construct becomes more immersive and the social community continually affirms a paranoid mindset, it may be that some fabulists gradually become true believers, all without reflectively recognizing this shift in their epistemic stance.Footnote ¹³

This possibility merits further investigation. But if it’s on the right track, we also have the beginnings of a second set of implications, this time yielding practical advice and perhaps even policy guidance. Many people seek advice for how to draw a loved one out of a conspiracy theory community when their participation has become harmful to family, career, or even health. Ultimately, social scientists and mental health clinicians are best placed to address this problem, but our account may offer them some starting points for a new set of tactics. Rather than confronting conspiracy practitioners about the epistemic deficiencies of their theories, it may be more productive to focus on redirecting the narrative impulse.

Part of what conspiracy theory communities offer, we have argued, is the thrill of constructing a narrative in concert with like-minded people—that is, a fandom. Perhaps some participants in pernicious conspiracy narratives can be guided to less dangerous fandoms that satisfy the same impulses, such as those around long-running fictional thriller TV programs, or historical true crime. Perhaps this approach will not work with the truest of true believers. But for people who drift into conspiratorial practice accidentally amid a long night of boredom-relieving internet exploration, this sort of intervention might be effective.

We leave the viability of this proposal for assessment by those with better empirical and clinical training. Our aim here has only been to expand the theoretical space for understanding the shared social motivations of conspiracy practice. We hope that our narrative account plays some small role in allowing us all to live happily ever after.