How should philosophers define conspiracy theory? In our paper, "What is a Conspiracy Theory?" (2021), we argued that the best definition is an evaluative one—one which characterizes conspiracy theories in terms of epistemic deficiency, appropriately spelled out. We argued for this on the basis of empirical studies we conducted on the ordinary meaning of conspiracy theory, which show that the predominant usage is epistemically evaluative. In that paper we also criticized the project of so-called particularists—philosophers who have advocated for a neutral definition of conspiracy theory as any theory involving a conspiracy.^[1] … [please read below the rest of the article].

We interpreted particularism as a conceptual engineering program aimed at improving the ordinary concept of conspiracy theory both in order to foster better scholarly investigations of conspiracy theories, and to minimize its (alleged) negative social and political impact. We argued on the basis of our empirical studies that the particularist’s engineering program doesn’t seem to be the best way to achieve the goals of theoretical fruitfulness, political accountability, and justice.^[2]

In a recent paper, Matthew Shields (2023) has raised some issues with our discussion. He argues that we have misconstrued the particularist position in a way that led us to overlook the problems of the evaluative engineering we propose. Shields’ paper contains a nuanced and thorough critique of our arguments. His core claim is that the particularist project does not amount to an attempt to engineer conspiracy theory. Rather, the particularist should be understood as aiming to diagnose conspiracy theory_e—the evaluative concept of conspiracy theory, which in our paper we showed to be prevalent among laypeople—as a site of conceptual domination.

Contrary to a common assumption, Shields (2021) argues that speakers engaged in conceptual disputes are not always driven by inquiry-related aims. In some cases, authoritative speakers exploit their institutional authority to push a particular view of a concept for goals that are unrelated to and sometimes actively interfere with inquiry (p. 15046). This is what Shields calls conceptual domination.

Based on his account of conceptual domination, Shields argues that particularists should be understood as arguing that conspiracy theory_e is a site of conceptual domination, where authoritative speakers—most notably, some academics and institutional figures—have tried to push a certain view of the concept conspiracy theory_e which serves to marginalize and stigmatize those outside of dominant institutions. Hence, he argues, our criticism of the particularist project as an inadequate engineering project falls flat. Particularists, on this view, are not even in the game of conceptual engineering—rather, they are arguing that conspiracy theory_e is a stigmatizing concept and should be abandoned. This understanding of particularism also casts doubt on our own engineering proposal. If conspiracy theory_e is indeed a site of conceptual domination, then any attempt to engineer this concept will have to carefully disentangle itself from its history and context of conceptual domination, making sure that the engineered concept will not contribute to further injustice and marginalization. And Shields is (rightly) pessimistic about the prospects of re-engineering the concept under these circumstances.

We find this interpretation of the particularist project very compelling, and we agree that we should have approached their project with a more critical perspective. In our paper, we focused on the particularists’ claims about the risk of dismissing warranted conspiracy accusations that arises from the ambiguity between a descriptive and evaluative sense in the ordinary expression ‘conspiracy theory’, e.g., the claims made by Basham and Dentith (2016). As a result, we understood their research project as an instance of conceptual engineering aimed at eliminating the ambiguity in the expression, and the related problems it causes for political accountability. However, we are now convinced that the most charitable understanding of particularism is as arguing that the concept conspiracy theory_e is a site of conceptual domination and should thus be abandoned.

While we agree that this is a better understanding of the view we critiqued in our paper, we doubt that this reading of particularism is in fact a more tenable position. In fact, we remain unconvinced by Shields’ arguments for why conspiracy theory_e is a site of conceptual domination. First, the empirical evidence that he relies on seems like a cherry-picked collection of data, and overall rather weak. Moreover, the alleged discrediting and stigmatizing effect of the concept conspiracy theory_e is compatible with an alternative and more plausible explanation which does not speak against our evaluative engineering proposal.

Does Conspiracy Theory_e Have a Key Stigmatizing Function?

What evidence does Shields present to support his assertion that "a key function of this concept […] is to stigmatize and marginalize views that challenge dominant institutions, figures, and beliefs, regardless of whether such views posit a conspiracy" (Shields 2023, 467)? To bolster his claim, Shields appears to heavily lean on the analysis conducted by Husting and Orr (2007) and Orr and Husting (2019) in their examination of New York Times articles. He quotes their work favorably on three occasions (Shields, 468), such as their statement: "the phrase is a mechanism of exclusion that symbolically banishes questions, claims, and concerns so labeled from the public sphere as unwarranted—or worse" (2007, 133).

The first crucial observation to make here, however, is that in all three instances, Husting and Orr are discussing the term ‘conspiracy theorist’, not ‘conspiracy theory’. Admittedly, Husting and Orr (2007) do present a negative perspective on the term ‘conspiracy theory’ too, but they also acknowledge that "Conspiracy theory might be used variously, for example, to conceal, defend, label, or paraphrase" (2007, 133, their italics). While it’s possible that conspiracy theory has a similar stigmatizing function as conspiracy theorist, the difference between the two may not be a trivial one (see e.g., Reuter and Baumgartner (forthcoming) for a corpus analysis on the evaluative nature of both terms). Labeling a person a ‘conspiracy theorist’ may convey information about a person’s character and dispositions which conspiracy theory may not equally carry.^[3]

Moreover, while Husting and Orr (2007) examine over 100 articles from the New York Times, and nearly 500 articles from the same source in 2019, their approach does not involve an empirical analysis of the results. Instead, they present a qualitative selection of text passages that support their core arguments. Given the lack of quantitative analysis, it is advisable not to place undue emphasis on their results. In fact, we encourage readers to explore one of the primary publicly accessible corpora themselves to form their own conclusions.^[4] Upon scrutinizing corpora such as COCA, NOW, and iWEB for instances of ‘conspiracy theory’, there is limited evidence to suggest that ‘conspiracy theory’ is widely employed for purposes of stigmatization. While it is true that most uses communicate a negative attitude towards conspiracy theories, it seems unjustified to jump to the conclusion that stigmatizing and marginalizing is a key function of conspiracy theory_e.

In our paper, we had raised doubts that conspiracy theory_e could play a key stigmatizing function, given what we know about how attributions of ‘conspiracy theory’ function in ordinary language. In a passage that Shields cites, we claimed that:

The evaluative concept conspiracy theory is prevalent in ordinary thought and language, and attributions of ‘conspiracy theory’ seem to be driven by an assessment of the target theory […]. Thus, the function that this concept serves in academic practices and discourses cannot be silencing warranted conspiracy accusations (2021, 2055).

Shields interprets our passage here as claiming that the label cannot play a stigmatizing function because ordinary speakers lack an intention to do so—rather, their intention is to assess the target theory. But this is not what we meant to say (and it is obviously false that labels cannot stigmatize without a speaker’s intention). Our point here—and we should have stated it in less ambiguous language—is that, since whether people label something ‘conspiracy theory’ seems to crucially depend on their epistemic assessment of the content of the theory (see our Study 2a and 2b), it seems hard to see how this label could successfully work to silence anything, with or without speaker intention.

In our paper, we showed that the most common adjectives before ‘conspiracy theory’ are epistemic terms like ‘baseless’, ‘debunked’, and ‘unfounded’. Labeling something a ‘conspiracy theory’ thus seems to hinge on an assessment of its epistemic status. Relatedly, as further studies have shown (Douglas et al. 2021), people don’t seem to believe a theory less if it’s labeled ‘conspiracy theory’ but rather the other way round: they use the label ‘conspiracy theory’ when they don’t find a theory believable. The content of the theory and its perceived plausibility, regardless of whether the label ‘conspiracy theory’ is used, seems to determine whether a theory is dismissed or not. And so it seems hard to see how calling something a ‘conspiracy theory’ itself could have any effect on how we treat both the theory in question and the people who endorse it.

Ultimately, we are unconvinced that the evidence presented by Shields shows that conspiracy theory_e has a central stigmatizing function—though we are currently working on a more thorough empirical investigation of this claim. Even granting that the label ‘conspiracy theory’ routinely has the effect of excluding and stigmatizing certain views, we are even more doubtful that this in turn supports the idea that conspiracy theory_e is a site of conceptual domination.

Is Conspiracy Theory_e a Site of Conceptual Domination?

Many concepts, descriptive or evaluative, can work to marginalize disenfranchised groups. In some cases, the marginalizing effect seems ethically unproblematic, and warranted by the meaning of the label used. Take the concept neo-Nazi. Calling someone ‘neo-Nazi’ has the effect of excluding their opinion from the realm of political opinions worth considering. When correctly applied, the discrediting effect of the concept neo-Nazi doesn’t seem unjust: after all, we should not consider neo-Nazism to be a political opinion worth listening to, and believing someone to be a neo-Nazi is a good reason to demote their credibility on political issues. So, the discrediting effect that the concept neo-Nazi can have against people who are outside of the relevant halls of power doesn’t seem to be ethically problematic, and doesn’t seem to speak to whether the concept neo-Nazi ought to be abandoned.

There are also plenty of cases where a concept’s marginalizing effect can be unjust, but this is not due to a problem within the concept itself. Take the case of liar.^[5] A liar is someone whose word cannot be trusted, and who should be given less credibility. So, calling someone a ‘liar’ can lead to silencing and discrediting them. And sometimes, power imbalances can lead powerful people to get away with misapplying the label ‘liar’ unjustly to discredit honest testifiers. Bosses can silence women who make allegations of sexual harassment in the workplace by calling them ‘liars’, and powerful politicians can discredit less powerful opponents by calling them ‘liars’. These are cases of marginalization that derive from (mis)using the label ‘liar’.

But the marginalization of people outside of dominant power positions that the concept liar brings about can hardly be conceived as a problem within the concept. Liar seems to be a perfectly fine concept—it is its misapplication by those who exploit their institutional authority and social power that generates unjust stigmatization and marginalization of disenfranchised groups. So, the injustice here seems to lie in the power imbalances that exist in the world, not in our concepts.

Evidence that conspiracy theory_e works to marginalize people outside of dominant institutions by itself doesn’t yet tell us that the concept is a site of conceptual domination. Conspiracy theory_e could work like neo-Nazi to discredit those who are worthy of being discredited—because they endorse absurdly implausible, and sometimes harmful theories. Or, like liar, conspiracy theory_e could sometimes stigmatize and marginalize certain groups in virtue of the preexisting power imbalances among speakers which let the powerful get away with incorrectly applying the label. In neither case the stigmatizing effect of conspiracy theory_e would speak to whether conspiracy theory_e is a problematic concept that ought to be abandoned.

So, what reason do we have for thinking that conspiracy theory_e is a site of conceptual domination and as such ought to be abandoned? Here, Shields seems to crucially rely on the observation that conspiracy theories fabricated by dominant institutions are not typically treated as central, paradigmatic cases by scholars investigating conspiracy theories. And this is so, he claims, even though dominant institutions’ conspiracy theories are more harmful and often display the same or worse epistemic flaws as theories like Pizzagate, 9/11 as inside job, or Flat earth, that are typically taken to be paradigmatic.^[6] He claims:

Part of the conceptual domination involving conspiracy theory_e, then, is that it works to uniquely stigmatize those outside of the relevant dominant institutional contexts and shield those same contexts from criticism. Such individuals may be genuine dissidents or they may well have egregiously mistaken or false beliefs, but are nonetheless singled out for censure because they are not part of the relevant dominant institutional context. […] Conspiracy theories fabricated by dominant institutions with the same, or worse, epistemic flaws and with far more devastating implications are typically absent in [psychologists’ and social scientists’] surveys, as is specific testing of those who belong to such institutions (2023, 470).

The key stigmatizing function, he continues, is operative not only in the work of these academics and institutional figures, but is also inherited by ordinary speakers:

Whether they realize or not, when ordinary speakers treat cases such as flat earthers, proponents of Sandy Hook as false flag, 9/11 as inside job, moon landing as faked, or JFK assassination claims as apt for categorization as instances of conspiracy theory_e, as in fact paradigmatic instances of conspiracy theory_e, this stigmatizing function is indeed in play. Here, ordinary speakers follow the practice of formally authoritative speakers: we are singling out cases of individuals outside of dominant institutions for special censure, even when the relevant epistemic flaws apply at least as directly and arguably much more urgently to views fabricated and endorsed by those within dominant institutions (2023, 472).

So, for Shields, the fact that conspiracy theory_e is dominant in ordinary language, like our paper showed, does not speak against the concept being a site of conceptual domination. It just goes to show that the conceptual domination has been successful and pervasive, and ordinary speakers have inherited the view of the concept which authoritative speakers have been pushing for their goals.

We find Shields’ line of reasoning unconvincing. First, ordinary speakers don’t seem to "[single] out cases [of conspiracy theories] outside of dominant institutions for special censure". We conducted an analysis using one of the largest open-access corpora available, known as NOW, to investigate who is typically accused of spreading conspiracy theories. To do this, we entered the search term ‘spreads conspiracy theories’,^[7] which yielded 39 results. Upon examining the actual text excerpts, we found that the individuals and entities described or accused of spreading conspiracy theories are a far-right website, Kennedy, a Central Bank Governor, a celebrity chef, a media organization, a Republican politician, an internet provocateur, Republicans, a major political party, YouTube, Trump (several times), a far-right channel, a media show, an organization, the current president, QAnon, the president, a network, an agency head, another Republican politician, Mr. Jones, Infowars, and the head of domestic intelligence. With the potential exception of the "internet provocateur", we primarily identify influential individuals and powerful organizations who are alleged to disseminate conspiracy theories.

As this admittedly cursory analysis suggests, conspiracy theories are not considered to be spread exclusively (or even typically) by people outside of dominant institutions. And this may initially also speak against the assumption that people don’t treat dominant institutions’ conspiracy theories as paradigmatic cases. We will investigate this assumption more systematically, but on the face of it, conspiracy theory_e as it operates in ordinary language seems to treat the conspiracy theories spread by dominant institutions as central cases.

One may take this fact to speak in support of the case for conceptual domination with conspiracy theory_e. Perhaps ordinary speakers have not yet inherited the view of the evaluative concept that authoritative speakers have been promoting. So one may think that some academics and institutional figures have been pushing a view of the concept whose goal is to marginalize those outside the relevant halls of power and shield dominant institutions from criticism, but ordinary speakers entertain a concept which is evaluative but not yet contaminated by their efforts. But this doesn’t seem to be well supported by the evidence we have either.

First, as Shields acknowledges, scholars like Sunstein and Vermeule (2009), as well as Cassam (2019), do include in their view of the concept examples of conspiracy theories spread by dominant institutions. In fact, in the two quotes cited by Shields, both Sunstein and Vermeule, and Cassam mention the Bush administration’s theory about a conspiracy of Iraq and Al Quaeda in the 9/11 attacks as an example of a conspiracy theory. Shields complains that they fail to treat examples like these as central in their analyses. But these scholars have not been promoting the view that theories like these are not conspiracy theories—which would make this case more similar to the conceptual dispute around the concept torture which Shields discusses as a paradigmatic case of conceptual domination. For Shields, in order to determine whether a concept is a case of conceptual domination,

We can ask whether the speakers [engaged in a conceptual dispute]:

(1) Are engaged in markedly deficient practices of inquiry;
(2) Exploit their formal or informal institutional authority to settle the conceptual dispute;
(3) Are unwilling to defend their view of the concept to relevant experts;
(4) Fail to be transparent about the interests informing or potentially informing their view (2023, 466).

While it’s true that both the work of Sunstein and Vermeule, and that of Cassam may not focus centrally on conspiracy theories spread by dominant institutions, it is unclear to us that they are engaging in markedly deficient practices of inquiry, using their authority to settle the conceptual dispute, that they are unwilling to defend their view of the concept to experts, and that they are failing to be transparent about their interests. The view of the concept they seem to endorse is one on which conspiracy theories like those promoted by the Bush administration are genuine conspiracy theories. And so it seems like a stretch to say that they are pushing a view of the concept which aims to exclude conspiracy theories fabricated by dominant institutions, and single out non-dominant institutions’ conspiracy theories for special censure.

Now, one could say that the problem is not that they don’t take these theories to be genuine conspiracy theories, but rather that, despite what they claim, they fail to treat them as central, just like social scientists and psychologists do. And this is enough to reveal that these speakers are in fact engaged in conceptual domination. But if all it takes for an authoritative speaker to be engaged in conceptual domination is that they operate with a view of a concept which includes certain instances but fails to treat them as paradigmatic, then the threshold for what behavior counts as conceptual domination seems to become rather low. Concepts can be structured around prototypical instances, and in many cases which instances we treat as paradigmatic can be shaped by our social stereotypes and the unjust power structures of our societies.

When we think of an immigrant we hardly ever picture a white European academic working in a foreign country; when we think of a boss we hardly ever think about a black woman; and when we think of a criminal we hardly ever picture a wealthy politician. Now, if researchers and authoritative speakers were to exclusively focus on (stereotypical) paradigmatic instances, we would likely think it problematic, as authoritative speakers would be perpetuating unjust social stereotypes. But claiming that they would be engaged in conceptual domination, that these concepts should be abandoned and that any prospects for engineering them is suspect, seems unwarranted. So, the mere fact that some academics in their work treat non-dominant institutions’ conspiracy theories as paradigmatic doesn’t seem to establish that these speakers are engaged in conceptual domination and that the concept ought to be abandoned.

Perhaps Shields here would want to say that despite our intuitions, the bar for conceptual domination is in fact low. Since the effect of marginalizing and stigmatizing is present, we should not dismiss the conceptual domination hypothesis even when the authoritative speakers involved don’t explicitly advocate for a view of the concept which excludes dominant institutions’ conspiracy theories. But if this is the case, then the prospects for engineering concepts that are sites of conceptual domination in this weaker sense, including conspiracy theory_e, seem much brighter (and very different from the examples he provides, including welfare queen and great replacement).

We take the concern of conceptual domination seriously, and we are working on follow-up studies to further investigate this hypothesis. However, the evidence provided so far seems to us to be better explained by an alternative account. Conspiracy theory_e may have a discrediting function which is warranted by what a conspiracy theory_e is. Theories with the epistemic flaws that conspiracy theory_e refers to are worthy of not being discussed, and those who endorse them are worthy of being challenged. And sometimes people in power can misapply the label ‘conspiracy theory’ and, due to the unjust ways in which power structures affect credibility attributions, they can get away with it and silence dissenting voices from disenfranchised groups.

At the same time, these unjust social structures may make it the case that when we think of conspiracy theories, those spread by non-dominant groups come to mind as paradigmatic. While this points to several ways in which we as academics could do better in not perpetuating social and political injustice, it also seems perfectly compatible with advocating for an evaluative engineering of conspiracy theory_e.

Conclusion

We are convinced that conceptual domination is a genuine phenomenon, and conceptual engineers should be careful when proposing ameliorations of concepts based on the assumption that speakers engaged in conceptual disputes are always driven by inquiry related aims. Yet we remain unconvinced that conspiracy theory_e is one such case. Clearly, not every case of discrediting is unjust and unwarranted. And even within the unjust cases, sometimes unjust power structures can make it the case that perfectly fine concepts can be used to discredit and marginalize. But this can speak to a problematic feature of the world, rather than of our concepts. So, we should be careful in checking that a certain theory to which the label ‘conspiracy theory’ is applied is in fact a conspiracy theory, especially when the accusation comes from people in power. And we should make sure that dominant institutions’ conspiracy theories are called ‘conspiracy theories’ and treated as central cases in academic analyses. But does this tell us anything about the prospects for reengineering the concept conspiracy theory_e? We don’t think so.

References

Basham, Lee and M. R. X. Dentith. 2016. "Social Sciences Conspiracy-Theory Panic: Now They Want to Cure Everyone." Social Epistemology Review and Reply Collective 5 (10): 12–19.

Boudry, Maarten and M. Giulia Napolitano. 2023. "Why We Should Stop Talking about Generalism and Particularism: Moving the Debate on Conspiracy Theories Forward." Social Epistemology Review and Reply Collective 12 (9): 22–26.

Cassam, Quassim. 2019. Conspiracy Theories. Cambridge, UK: Polity Press.

Douglas, Karen M., Jan-Willem van Prooijen, and Robbie M Sutton. 2022. "Is the Label ‘Conspiracy Theory’ a Cause or a Consequence of Disbelief in Alternative Narratives?" British Journal of Psychology 113 (3): 575–590.

Husting, Ginna and Martin Orr. 2007. "Dangerous Machinery: ‘Conspiracy Theorist’ as a Transpersonal Strategy of Exclusion." Symbolic Interaction 30 (2): 127–150.

Napolitano, M. Giulia. 2021. "Conspiracy Theories and Evidential Self-insulation." In The Epistemology of Fake News edited by Sven Bernecker, Amy K. Flowerree, and Thomas Grundmann, 82–105. Oxford University Press.

Napolitano, M. Giulia and Kevin Reuter, K. 2021. What Is a Conspiracy Theory? Erkenntnis 88: 2035–2062.

Orr, Martin and Ginna Husting. 2019. "Media Marginalization of Racial Minorities: ‘Conspiracy Theorists’ in U.S. Ghettos and on the ‘Arab Street’." In Conspiracy Theories & the People Who Believe Them edited by Joseph E. Uscinski, 82–93, New York: Oxford University Press.

Reuter, Kevin and Lucien Baumgartner. forthcoming. "Corpus Analysis — Building and Using Corpora: A Case Study on the Use of ‘Conspiracy Theory’." In Experimental Philosophy for Beginners edited by Stephan Kornmesser, Alexander Max Bauer, Mark Alfano, Aurélien Allard, Lucien Baumgartner, Florian Cova, Paul Engelhardt, Eugen Fischer, Henrike Meyer, Kevin Reuter, Justin Sytsma, Kyle Thompson, and Marc Wyszynski. New York: Springer Nature.

Reuter, Kevin and Lucien Baumgartner. 2023. "Conspiracy Theories are not Theories: Time to Rename Conspiracy Theories." (May, draft): 1–7. https://philarchive.org/rec/REUCTA-2.

Ritchie, Katherine. 2021. "Essentializing Language and the Prospects for Ameliorative Projects." Ethics 131 (3): 460–488.

Shields, Matthew. 2023. "Conceptual Engineering, Conceptual Domination, and the Case of Conspiracy Theories." Social Epistemology 37 (4): 464–480.

Shields, Matthew. 2022. "Rethinking Conspiracy Theories." Synthese 200 (331): 1–29.

Shields, Matthew. 2021. "Conceptual Domination." Synthese 199 (5–6): 15043–15067.

Sunstein, Cass and Adrian Vermeule. 2009. "Conspiracy Theories: Causes and Cures." The Journal of Political Philosophy 17 (2): 202–227.

---

^[1] This is at least one way to characterize particularism, even though, as one of us argued in a recent contribution to this publication, the labels ‘particularism’ and ‘generalism’ appear to track two orthogonal philosophical disagreements—a semantic one, and an epistemological one (see Boudry and Napolitano 2023).

^[2] Our paper provided general guidelines for engineering conspiracy theory, leaving the exact nature of the epistemic deficiency of conspiracy theories to be specified by further philosophical argument (see, e.g., Napolitano 2021).

^[3] There is also evidence that predicate nominals label and categorize people in a way that typically elicits essentializing inferences (Ritchie 2021).

^[4] If you are interested to check for yourself, go to https://www.english-corpora.org/now/, register, and then enter ‘conspiracy theory’ into the search field. There are currently over 35.000 hits for ‘conspiracy theory’.

^[5] Our examples here concern labels that are attributed to people, for ease of exposition, but a similar case could be made for different labels. The labels ‘lie’, and ‘neo-Nazism’ would be more similar to ‘conspiracy theory’, rather than ‘conspiracy theorist’.

^[6] See also Shields (2022).

^[7] Reuter and Baumgarter (ms) present the results of a corpus analysis on verbs occurring in front of ‘conspiracy theory’ and find spreading terms to be highly prevalent.

---

---

Categories: Critical Replies

Tags: conceptual domination, conceptual engineering, conspiracy theories, generalism, generalism vs. particularism, Kevin Reuter, M. Giulia Napolitano, Matthew Shields, particularism, SERRC 11/21/2023

Leave a Reply

Pages

• About
• Article Submissions
• Blogroll
  • Conferences
  • Special Issues
• Book Review Guidelines
• Book Reviews
• Book Series
• Book Symposiums
• Collective Vision
• In and Beyond the Era of COVID-19
• Knowledge for Breakfast: A Podcast
• Links
• SERRC Issues
• Site Bibliography
• Who We Are

Archives

Archives

Email Address

Join 8,257 other subscribers