I conducted some of the ‘first generation’ work in analytic philosophy on conspiracy theories (Clarke 2002; 2006; 2007),^[1] and then set the topic aside for 14 years.^[2] The current scene is quite different from the one I left. One difference is the sheer volume of analytic philosophical work now being conducted on the topic. In 2006 it was possible for David Coady (2006) to put together a slim edited volume (about 180 pages) containing contributions by all six analytic philosophers who had published work on the philosophy of conspiracy theories. If anyone were game enough to try to create an equivalently comprehensive volume, in 2023, they would have to be willing and able to edit several thousand manuscript pages. … [please read below the rest of the article].

Image credit: Cristina Ivan via Flickr / Creative Commons

Before Generalism and Particularism

A second difference is that many of the analytic philosophers working on conspiracy theories have taken to using a taxonomical distinction to situate their (and other) contributions within the burgeoning literature. This is the distinction between generalism and particularism. It seems that I am widely regarded as a generalist.^[3] The distinction was introduced by Buenting and Taylor (2010), and they list Clarke (2002) as one of three examples of generalism about conspiracy theories.^[4] Their other two examples are Keeley (1999) and Mandik (2007).

After learning that I was an exemplar generalist, in the now influential taxonomy, I was naturally curious to find out what it was about my work that made it generalist; and when I tracked down Buenting and Taylor’s (2010) paper I was in for a surprise. Buenting and Taylor describe generalism as the view that “conspiratorial thinking qua conspiratorial thinking is itself irrational” (2010, 568). This is not my view at all. In Clarke (2002) I advocate a Lakatosian approach to deciding when it is rational to believe in a particular conspiracy theory. As anyone who has read Lakatos (1970) will be aware, he was concerned to uphold the role of rationality in theory choice—to push back against Kuhn’s (1962) view that the process by which dominant theories are replaced in the natural sciences is arational. I thought I’d made it clear that I was specifying circumstances in which it is rational to believe in particular conspiracy theories. I also mentioned a specific example of a conspiracy theory I regard as rational to accept—the theory that the Nixon administration conspired to orchestrate the break in in the Democrat National Committee offices in the Watergate complex—the ‘Watergate theory’.^[5]

Another of Buenting and Taylor’s exemplars of generalism doesn’t fit their definition either. Like me Keeley specifically mentions the Watergate theory as one that it is rationally acceptable (1999, 111 and 126). He also mentions a second such theory: the ‘Iran-Contra theory’ (Keeley 1999, 111 and 126). Buenting and Taylor cite specific pages where Keeley (1999) and Clarke (2002) allegedly assert that conspiratorial thinking is irrational.^[6] However, at the points in our respective papers mentioned, Keeley is referring to a subclass of conspiracy theories—unwarranted conspiracy theories—and I am also referring to a subclass of conspiracy theories: those that are at the core of what Lakatos (1970) refers to as ‘degenerating research programs’. Neither of us asserts that all conspiratorial theorising is irrational. Buenting and Taylor (2010) are right about one of their three examples, at least, as Mandik (2007) really does regard it as irrational to accept any and every conspiracy theory.

A New Generalism

The term ‘generalist’ has come to be used more loosely than Buenting and Taylor’s (2010) definition, with their stress on the irrationality of conspiracy theorising dropped. Stokes describes generalism as the view that ‘conspiracy theories as a class of explanation are intrinsically suspect’ (2018, 25), Harris, characterizes generalism as the view that ‘conspiracy theories may be assessed as a class’ (forthcoming) and Boudry describes generalism as ‘the view that there is indeed something prima facie suspicious about CTs, properly defined, and that they suffer from common epistemic defects’ (2023, 611).

On these characterizations generalists continue to make universal generalizations about conspiracy theories, but these are defeasible generalisations. It is possible to endorse such generalisations and still think that there are some conspiracy theories that it can be rational to accept: namely those theories that have sufficient epistemic merits to overcome the general epistemic deficit or deficits that supposedly afflict all conspiracy theories. It is also possible that some of the various people who regard me as a generalist suppose that I am a generalist in this looser sense. If so, are they right? To answer this question, it can help to revisit an old debate.

The conspiracy theorist who contests an official theory offers a competing explanation of the events accounted for by that theory, often by postulating a ‘cover-up’ that has gone unseen by most and which ensures that the alleged conspirators have been able to operate largely in secret. In making the case for their favoured conspiratorial rival to the official theory, the conspiracy theorist relies heavily on what Keeley (1999, 117–118) refers to as ‘errant data’—data that the official theory fails to account for, or which contradicts the official theory. This reliance on errant data can make conspiracy theories appear epistemically unusual. But we need to be careful about concluding that it leads to any kind of epistemic defect.^[7]

As Keeley notes, errant data has also played a crucial role in the history of science, and the fate of scientific theories has sometimes turned on the ability of a new theory to explain errant data that an established theory cannot explain (Keeley 1999, 120). The Kennedy assassination conspiracy theorist’s fixation on the grassy knoll is not obviously more epistemically problematic than Einstein’s focus on the precession of the perihelion of Mercury. Keeley suggests that nevertheless there is an epistemic defect that affects many conspiracy theories in some circumstances—those circumstances in which ‘positive evidence for the conspiracy fails to obtain’ (1999, 121). It seems that conspiracy theorists who find themselves in such circumstances often seek to explain the lack of positive evidence by alleging ever more grandiose conspiracy theories involving more and more participants. Over time conspiracy theorists end up embracing a global skepticism that undermines our confidence in such mundane ordinary beliefs as the ordinary conviction that the platypus is a mammal, or so Keeley argues (1999, 121).

I responded to Keeley by arguing that he was mistaken in blaming conspiracy theories for the slide towards skepticism, when it occurs (Clarke 2002).^[8] If there is a problem, it is with the behaviour of those conspiracy theorists who remained committed to degenerating research programmes. These are sequences of theories that are successively modified, with the addition of more and more auxiliary hypotheses in response to failed predictions and retrodictions, and which become increasingly convoluted over time. So, am I a generalist about conspiracy theories in the newer loose sense? Definitely not. As we have seen, the generalist, in a loose sense, is someone who holds that conspiracy theories, as a class, are epistemically defective.^[9] I’ve pushed back against allegations of such epistemic flaws, arguing that the defects alleged by Keeley (1999) are better explained by appealing to the psychology of some conspiracy theorists—those who advocate degenerating research programs with sequences of conspiracy theories at their core.

I’m an anti-generalist. In Clarke (2002) I speculated that the tendency of some conspiracy theorists to stick with clearly degenerating research programs long past the point where most would abandon such research programs could be explained by appeal to the ‘fundamental attribution error’. In a later addendum to that paper (Clarke 2006), I stated that it has been a misstep to make this appeal, but I stuck to my guns on the issue of how we should go about explaining why conspiracy theories are accepted when available evidence does not warrant such acceptance. I continued to insist (in 2006 and now) that we should appeal to the sciences of the mind and try to locate cognitive deficiencies that affect sub-groups of conspiracy theorists, instead of looking for epistemic defects that are endemic to conspiracy theories.

Conspiracy Theorists

There is large body of work in psychology and other sciences of the mind, that can be and is appealed to, to explain why people adhere to conspiracy theories in the face of a lack of evidence for those theories.^[10] Unfortunately many of the scientists conducting this work present their ideas as insights about conspiracy theories rather than about the distorted reasoning patterns of some conspiracy theorists.^[11] But what they are actually providing are accounts of the various cognitive defects that lead many people to find particular conspiracy theories convincing despite the lack of objective evidence for those theories, along with studies of the circumstances in which such people are especially liable to find such theories attractive.

Scholars working on conspiracy theories, including many philosophers, often seem to proceed as if the distinction between the class of conspiracy theories and the theories that conspiracy theorists happen to hold is trivial. But there are important differences here. Arguments about the epistemic merits or defects of conspiracy theories are normative arguments about epistemic standards that ought to apply to the assessment of a class of theories and should be able to account for both possible and actual cases of conspiracy. Arguments that appeal to the cognition of conspiracy theorists to explain conspiracy beliefs are descriptive arguments and focus on actual cases. We need to be very careful about generalizing from the actual cases of conspiracy theorizing that are studied to conspiracy theories as a class. It may well be that in the actual world many of the less warrantable of the very many various possible conspiracy theories that could be held are currently held.^[12]

The above reasoning may seem a bit hard to follow, so let me offer an analogy: suppose we were to provide clear commonsense descriptions of the various theories of physics that have been dominant for periods of history, laying out the evidence for and against each theory and then ask research subjects, who were representative of the general public, to tell us which of these they regarded as most likely to be true. Further suppose that a clear majority of our research subjects told us they regarded Thalean, Pythagorean or Aristotelian physics as most likely to be true and were dismissive of modern theories. How should we go about explaining this curious result (assuming the methodology was sound and the research subjects had understood the task at hand)?

The right way to go would be to look for features of human cognition that operate to make Ancient Greek theories of physics seem credible to ordinary people when these theories lack credibility in the eyes of experts. The wrong way to go—and I’d be surprised if anyone working in the philosophy or sociology of physics would try do this—would be to suppose that there is some general epistemic defect in the class of physical theories waiting to be identified. But many generalists about conspiracy theories appear to proceed in this wrong way. They start with the fact that large numbers of members of the general public appear to accept conspiracy theories that are lacking in objective warrant and proceed to try to locate epistemic flaws that supposedly affect conspiracy theories as a class.

Two Problems with the Generalist-Particularist Distinction

So, if I’m not a generalist, am I a particularist, and is the problem here simply one of miscategorisation, or is the generalist-particularist distinction itself problematic? To answer these questions we need consider particularism in more detail. According to Buenting and Taylor ‘the particularist view about conspiratorial thinking denies that the rationality of conspiracy theories can be assessed without considering particular conspiracy theories’ (2010, 568-9).^[13] This still seems to be the way the term ‘particularism’ is used. Harris (forthcoming) describes particularism as the view that ‘conspiracy theories are to be assessed individually’, and Stokes describes it as the view that ‘conspiracy theories should each be taken on their own merits and not simply dismissed out-of-hand’ (2018, 25).

The problem raised by these accounts of particularism is that they are silent about less-than-universal generalisations about conspiracy theories, which, as we have established, are insufficient to count as generalist. But, as we have seen, such less-than-generalisations about conspiracy theories are advanced by philosophers, including Keeley (1999) and Clarke (2002). Keeley (1999) and Clarke (2002) both examine particular conspiracy theories, so that makes both of these papers works of particularism. However, one could apply the less-than-universal generalisations supported in either of these papers, to assess a subset of the class of all conspiracy theories, without ever examining particular conspiracy theories. Someone who did so would be neither a particularist nor a generalist.

A possible response to the previous point would be to argue that the person who endorses less-than-universal generalisations about subclasses of conspiracy theories that were created by a particularist, really ought to count as a particularist, as their views about those conspiracy theories have a particularist genealogy. However, it seems possible to establish rationally warranted less-than-universal generalisations about conspiracy theories that bear no genealogical relation to studies of any specific conspiracy theory.

Consider the case of a researcher working on shamanism who investigates the possibility (which is sometimes alleged) that some shamans can shape-shift between human and non-human form. Suppose that after extensive investigation the researcher establishes that it is impossible for any being to shape-shift between human and non-human form. Suppose further that this researcher hears mention of David Icke’s theory that the world is secretly controlled by reptilian aliens able to shape shift between human and reptilian form (Icke 1999). Surely it is rational for the researcher to infer, on their basis of their findings about the impossibility of shape-shifting, that Icke’s theory, and all other conspiracy theories that rely on the possibility of shape-shifting, are unwarranted. This is a case that is neither particularist nor generalist. The shamanism researcher does not assess the merits of any particular conspiracy theory, before rejecting some conspiracy theories, so they are not a particularist. But they also fail to draw any universal generalisations about conspiracy theories, so they are not a generalist.^[14]To return to my earlier questions, Clarke (2002) is, contra received opinion, a particularist work, but consideration of it points to a problem with the generalist-particularist distinction. The distinction fails to underpin a complete taxonomy of the possible epistemic positions that philosophers working on conspiracy theories can hold.

Taxonomies are useful things and in the absence of a better taxonomy of accounts of philosophical treatments of conspiracy theories, there is merit in seeing if we can reform the generalist-particularist distinction to address the aforementioned problem. This could perhaps be done by revising the definition of either generalism, or particularism, to ensure the cases identified that fall between generalism and particularism are allocated to one these categories. The problem could also be addressed by recasting the distinction between generalism and particularism as a spectrum rather than a dichotomy, as Boudry has recently suggested (2023, 615). Either of these approaches would be helpful, provided the revised terminology was widely adhered to and applied consistently.

But there is a further problem with the enthusiasm with which philosophers working on conspiracy theories apply the generalist-particularist distinction. It is an epistemic distinction and its application to a paper like Clarke (2002), which is not primarily driven by epistemic considerations, can lead to that work being misconstrued as a work that is primarily driven by epistemic considerations (which may help explain why it has been persistently mis-identified as a generalist work).

Many of the philosophers working on conspiracy theories develop views that appear to be driven primarily by concerns outside of epistemology (as it has traditionally been understood). Recent examples include, Bortolotti (2023), Clarke (2023), Hauswald (2023), Keeley (2023), and Lepoutre (forthcoming). Attempts to shoe-horn the ideas articulated in these papers into the framework of the generalist-particularist distinction shed little or no light on those ideas. Such attempts also risk obscuring the non-epistemic nature of the concerns underpinning the research contained with the papers in question. To head off this risk it can help to keep another distinction in mind: the distinction between work on conspiracy theories and work on conspiracy theorists.

Acknowledgments

Thanks to Maarten Boudry, Dominic Bunnell and J. C. M. Duetz for helpful comments on earlier drafts of this paper.

References

Bortolotti, Lisa. 2023. “Is it Pathological to Believe in Conspiracy Theories?” Transcultural Psychiatry https://doi.org/10.1177/13634615231187243.

Boudry, Maarten. 2023. “Why We Should Be Suspicious of Conspiracy Theories: A Novel Demarcation Problem.” Episteme 20 (3): 611­–631.

Buenting, Joel and Jason Taylor.  2010. “Conspiracy Theories and Fortuitous Data.” Philosophy of the Social Sciences 40 (4): 567–578.

Clarke, Steve. 2023. “Is There a New Conspiracism?” Social Epistemology 37 (1): 127–140.

Clarke, Steve. 2021. “Conspiracy Theories?” In Future Morality edited by David Edmonds, 139–147. Oxford: Oxford University Press.

Clarke, Steve. 2007. “Conspiracy Theories and the Internet: Controlled Demolition and Arrested Development.” Episteme 4 (2): 167–180.

Clarke, Steve. 2006. “Appealing to the Fundamental Attribution Error: Was it All a Big Mistake?” In Conspiracy Theories: The Philosophical Debate edited by David Coady, 129–132. Aldershot: Ashgate.

Clarke, Steve. 2002. “Conspiracy Theories and Conspiracy Theorizing.” Philosophy of the Social Sciences 32: 131-150.

Coady, David, ed. 2006. Conspiracy Theories: The Philosophical Debate. Aldershot: Ashgate.

Dentith, M R. X. 2023a. “The Future of the Philosophy of Conspiracy Theory: An Introduction to the Special Issue on Conspiracy Theory Theory.” Social Epistemology 37 (4): 405-412.

Dentith, M R. X. 2023b. “Some Conspiracy Theories.” Social Epistemology 37 (4): 522-534.

Douglas, Karen M. and Robbie M. Sutton. 2023. “What Are Conspiracy Theories? A Definitional Approach to Their Correlates, Consequences, and Communication.” Annual Review of Psychology 74: 271-98.

Duetz, J. C. M. 2023. “What Does it Mean for a Conspiracy Theory to be a Theory?” Social Epistemology 37 (4): 438–453.

Hagen, Kurtis 2022. “Is Conspiracy Theorizing Really Epistemically Problematic.” Episteme 19: 197–219.

Harris, Keith Raymond. 2022. “Some Problems with Particularism.” Synthese 200: 447. https://link.springer.com/article/10.1007/s11229-022-03948-9.

Hauswald, Rico. 2023. “That’s Just a Conspiracy Theory!”: Relevant Alternatives, Dismissive Conversational Exercitives, and the Problem of Premature Conclusions.” Social Epistemology 37 (4): 494-509.

Heering, David 2022. “Alethic and Narrative Models of Conspiracy Theories.” Zeitschrift für Praktische Philosophie 9 (2): 143–174.

Icke, David. 1999. The Biggest Secret. Ryde: David Icke Books.

Keeley Brian L. 2023. “Conspiracy Theories and (or as) Folk Psychology.” Social Epistemology 37 (4): 413­–422.

Keeley, Brian L. 1999. “Of Conspiracy Theories.” The Journal of Philosophy 96 (3): 109-126.

Kuhn, Thomas. 1962. The Structure of Scientific Revolutions. Chicago: University of Chicago Press.

Lakatos, Imre. 1970. “Falsification and the Methodology of Research Programmes.” In Criticism and the Growth of Knowledge edited by Imre Lakatos and Alan Musgrave, 91–195. Cambridge: Cambridge University Press.

Lepoutre, Maxime, C. 2022. “Narrative Counterspeech.” Political Studies https://doi.org/10.1177/00323217221129253.

Mandik, Pete. 2007. “Shit Happens.” Episteme, 4: 205–218.

Stokes, Patrick. 2018. “Conspiracy Theory and the Perils of Pure Particularism.” In Taking Conspiracy Theories Seriously edited by M R. X. Dentith, 25–37. London: Rowman & Littlefield.

^[1] ‘First generation’ is Dentith’s terminology (2023a, 408).

^[2] My more recent work includes Clarke (2021) and (2023).

^[3] See, for example, Duetz, (2023, 450, n. 10); Hagen, (2022, 198); Heering (2022, 147).

^[4] See Buenting and Taylor (2010 568, n. 3.).

^[5] See Clarke (2002, 136).

^[6] See Buenting and Taylor (2010, 572, n. 8.).

^[7] For further discussion of Keeley (1999) on errant data, see Clarke (2002, 139-140).

^[8] I also argued that the extent of the problem Keeley had identified was exaggerated. See Clarke (2002, 140-1).

^[9] Note that I am using the term ‘class’ here to refer to all logically possible conspiracy theories. This usage in in keeping with usual uses of the term in philosophy. It is possible that some other users of the term intend to refer only to actual conspiracy theories.

^[10] For a recent survey, see Douglas and Sutton (2023).

^[11] For an example of this phenomenon, see Douglas and Sutton (2023).

^[12] I argue similarly in Clarke (2002, 142). Note that there may be scope to reduce the influence of cognitive bias on ordinary cognition and so reduce the popularity of unwarranted conspiracy theories. For discussion, see Clarke (2021).

^[13] Buenting and Taylor (2010) defend what they take to be a specific form of particularism. This is the view that ‘an evaluation of any particular conspiracy theory must involve a weighing of evidence against its rival theory—the official story (2010, 569).’ I like this view which they don’t name—call it ‘comparativism’—as it encapsulates a key idea that seems to me right: we evaluate theories relative to their rivals. So, a theory can have all manner epistemic virtue, but that does not make it rationally acceptable. It has to exhibit more epistemic virtue than its rivals to be rationally acceptable. I do want to quibble with their characterization of comparativism, however. There will be circumstances in which there is no official theory—and in those circumstances the main rivals to a conspiracy theory will be competing non-official theories.

^[14] For other discussions of hard-to-categorise cases between particularism and generalism, see Boudry (2023), Dentith (2023) and Stokes (2018).

Categories: Articles

Tags: Brian Keeley, conspiracy theories, conspiracy theorists, generalism, Imre Lakatos, particularism, SERRC 8/29/2023, Steve Clarke