1. Introduction
In 2020, during the height of the COVID-19 pandemic, German virologist and coronavirus expert Christian Drosten complained on a podcast about scientists from unrelated fields who pose as experts on coronaviruses:
I am a virologist and would never comment on a bacteriological topic. And that’s almost the same for the normal viewer, viruses and bacteria, but not for a scientist. It goes even further than that. Even within virology, I wouldn’t dare to comment so broadly and so strongly on any virus other than the one I’m working on. One cannot know the literature and expertise in this area unless one is an absolute specialist. […] And what I hear, partly from supposed experts, who are certainly also experts in their own field of research or were while they were still working, lacks any basis. These are platitudes that do not go beyond a superficial knowledge of student textbooks. And with this knowledge base, they trumpet videos to the public and bolster dangerous conspiracy theorists, some of whom also have political agendas. That is irresponsible. (Hennig and Drosten Reference Hennig and Drosten2020 [own translation])
What Dorsten describes here has been termed “epistemic trespassing” by Nathan Ballantyne (Reference Ballantyne2019). This term describes situations in which individuals who have expertise in one domain also present themselves as experts in another field in which they actually lack the required expertise. Mikkel Gerken defines “epistemic trespassing testimony” in the following way:
-
S’s testimony that p is expert trespassing testimony iff
-
(i) S is an expert in D1, where D1 is a domain of expertise.
-
(ii) S is not an expert in D2, where D2 is a purported domain of expertise.
-
(iii) p∉D1.
-
(iv) p ∈ D2. (Gerken Reference Gerken2018: 303)
-
While there is a lively debate among philosophers of science when “boundary crossing” (Watson Reference Watson2022b) is warranted in science and may even be epistemically fruitful (Gerken Reference Gerken2023; Pavličić et al. Reference Pavličić, Dimitrijević, Vučković, Đorđević, Nedeljković and Tešić2024), epistemic trespassing can also have detrimental effects on public trust in science.Footnote 1 Scientists, whether consciously or unconsciously, may present themselves as experts on a topic in which they actually have no expertise. When such epistemic trespassers publicly and vocally challenge genuine experts on research results of societal relevance, they risk generating confusion or fostering distrust in science among the general public. In the most severe cases, epistemic trespassers can even derail effective policy for decades, as the case of the climate change denialism demonstrates. Oreskes and Conway (Reference Oreskes and Conway2010) argue in their landmark book on the topic that some of the most prominent scientists in the debate presented themselves as experts on climate change despite never having worked as climate scientists. The problem for laypeople is that epistemic trespassers often come across as experts because they seem to fulfill specific criteria that we typically look for in experts, such as being employed at a scientific research institute or having academic degrees, like a PhD, in some area of science.
More generally speaking, we might say that epistemic trespassing constitutes for non-experts a special version of the recognition problem (Watson Reference Watson2022a). In many domains and specifically in science, the only direct, and therefore foolproof, way to judge the degree of someone else’s expertise is to be an expert in the same domain yourself. In such a situation, one can easily ask the purported expert questions pertaining to their general understanding of the subject matter and evaluate the reasoning behind specific arguments. One’s own expertise in a given domain gives one the ability to assess others’ expertise, or lack thereof, in that domain. Novices do not have the same options available. Instead, they can only resort to auxiliary criteria to evaluate the expertise of purported experts. What such criteria should be and to what extent they can be useful to laypeople has been the topic of much debate among philosophers (Anderson Reference Anderson2011; Collins and Evans Reference Collins and Evans2007; Goldman Reference Goldman2001). Some prominent metacriteria that are often discussed include credentials, peer recognition, track record, and dialectical superiority. However, we argue that the concept of such metacriteria is underdeveloped, as it is unclear what specifically they are indicators for. We argue that they are actually not indicators of expertise but indicators of experience and lack of epistemic integrity.
To that end, we will proceed as follows: first, we will take a closer look at some of the most prominent philosophical definitions of expertise, in particular Alvin I. Goldman’s truth-linked approach (Reference Goldman1991, Reference Goldman2001, Reference Goldman2018) and the currently widely discussed understanding approach (Croce Reference Croce2019; Croce and Baghramian Reference Croce and Baghramian2024; Grundmann Reference Grundmann, Lackey and McGlynn2025; Scholz Reference Scholz2018). However, we argue that these accounts of expertise do not further our understanding of how non-experts can assess who qualifies as an expert (section 2). Goldman (Reference Goldman2018) argues that how to define expertise and how to identify experts are two separate questions. We argue that there is an approach to defining expertise that has been underappreciated by philosophers but is useful for understanding what non-experts should be on the lookout for in experts. We propose to shift the focus in the philosophical debate somewhat: Instead of asking what cognitive abilities define an expert, we ask how someone becomes an expert (section 2.1). Based on Collins and Evans’ (Reference Collins and Evans2007) analysis of the significance of specialist tacit knowledge to expertise, we identify a necessary condition for expertise: experience. That is, to become an expert, one must necessarily have experience practicing in the specific domain. We argue that common indicators for expertise are actually indicators of experience. We then go through some of the most widely discussed indicators for expertise – academic credentials, dialectical superiority, and track record (section 2.2). By arguing that these are in fact indicators of experience rather than expertise, we highlight that there is something else about experts that non-experts can assess. Moreover, behaviors such as failing to disclose conflicts of interest or fabricating data are sometimes treated as indicators of a lack of expertise. However, we argue that these are actually indicators of the (lack of) epistemic integrity of the purported experts (section 3). We note that epistemic integrity is a precondition for the functioning of the scientific process. We therefore argue that indicators of a lack of epistemic integrity enable laypeople to assess whether an expert assessment – in particular, based on indicators of experience – is possible at all. By taking a detour into the debate about public trust in science, we further conclude that indicators of experience and indicators of lack of epistemic integrity together are indicators of epistemic trustworthiness (section 4). We also claim that this detailed analysis of what non-experts actually can tell about experts helps us further define epistemic trespassing (section 5). In particular, we argue that understanding the significance of being part of a specific scientific community in which one acquires the necessary specialist tacit knowledge shows what epistemic trespassers are actually epistemically at fault for. They purport to be experts without having any experience of working on that specific topic.
2. Defining expertise
Philosophers’ attempts to define expertise go back as far as Plato’s Charmides.Footnote 2 In more recent times, a particularly well-discussed approach is Alvin I. Goldman’s truth-linked account, or as he also calls it “veritistic approach” (Reference Goldman1991, Reference Goldman2001, Reference Goldman2018). While Goldman has developed different variations of the truth-linked account of expertise, the most prominent version (Croce and Baghramian Reference Croce and Baghramian2024; Watson Reference Watson2021) states that S is an expert in a domain D under the following conditions: (1) “S has more true beliefs (or high credences) in propositions concerning D than most people do, and fewer false beliefs”; (2) “the absolute number of true beliefs S has about propositions in D is very substantial” (Goldman Reference Goldman2018: 5); (3) that S has “a capacity or disposition to deploy or exploit this fund of information to form beliefs in true answers to new questions that may be posed in the domain” (Goldman Reference Goldman2001: 91).
This truth-linked account of expertise has been criticized for several reasons. One can, for example, imagine a situation where someone holds a significantly large number of true beliefs about a subject purely by chance or due to being told so by an actual expert (Goldman Reference Goldman2018; Grundmann Reference Grundmann, Lackey and McGlynn2025). Further, there might be instances in which experts, by virtue of being experts and tackling unsolved questions, have both more false beliefs and more true beliefs than non-experts (Scholz Reference Scholz2009). A currently popular approach in the debate that overcomes this problem centers understanding as a necessary criterion for expertise (Croce Reference Croce2019; Scholz Reference Scholz2009, Reference Scholz2018; see also Elgin Reference Elgin2017). For example, Scholz argues that one condition of expertise is that “A has a considerably better understanding of domain D than the vast majority of people do” (Reference Scholz2009: 193). The advantage of this approach is that a broad understanding of a subject matter does not happen accidentally and is not something one can acquire in passing.
Both of these approaches – the truth-based and the understanding-focused account – express something significant and true about expertise. It is undeniable that expertise requires holding (a significant amount of) true beliefs and having some degree of understanding of the subject matter. However, we will not discuss the advantages and disadvantages of these accounts of expertise in more detail here, as it does not bring us any closer to finding criteria for non-experts to assess the expertise of purported experts. The problem is that we need to be experts ourselves to judge whether someone else holds enough true beliefs about a particular subject matter or has an understanding of it. Laypeople, by definition, do not have the required grasp of the subject matter.
We therefore argue for looking at the definition of expertise from a different perspective. Instead of asking what the specific cognitive abilities of an expert are, we propose to look at this problem from the perspective of how one becomes an expert. As non-experts, we might not be able to judge whether a purported expert has the required knowledge or understanding, but we might be able to assess whether someone has undergone the necessary steps to become an expert. Thus, the first question is, is there something that everyone who wants to become an expert has to do to acquire expertise, no matter the discipline? In order to become a scientist, one usually has to obtain some sort of degree from a university. At first glance, asking for such credentials seems to be an easy approach to identifying experts. However, credentials are neither a sufficient nor necessary condition for expertise. First of all, it is not hard to think of a case where someone develops a high degree of expertise in a domain without ever gaining an official certificate as evidence for it. More importantly for our case here, it does not help us find a criterion to detect epistemic trespassing, because often the real expert and the epistemic trespasser have very similar credentials. To the non-expert, the credentials of an expert in virology and an expert in bacteriology can look very similar (see section 2.2).
So let’s consider more broadly the process of becoming an expert. Research done by psychologists shows that one thing expertise requires is “deliberate practice” (Ericsson and Pool Reference Ericsson and Pool2016). However, contrary to popular belief, studies do not show that there is a fixed number of hours one has to spend to become an expert (Watson Reference Watson2021).Footnote 3 So we might instead ask whether there is a common process that everyone wanting to become an expert has to undergo.
2.1. How to become an expert
One approach that expands on this question comes from sociologist Harry Collins. Based on Collins’ research on the significance of tacit knowledge (Reference Collins2010, Reference Collins2014) in science, Collins and his co-author Robert Evans (Reference Collins and Evans2007) point out that expertise of the type scientists haveFootnote 4 requires “specialist tacit knowledge”, which cannot be learned from books alone. Instead, much of the knowledge required from a scientific expert can only be acquired in practice. Further, it is only accessible through being part of a scientific community and by being in conversation with others. This is due to the fact that at least part of this knowledge is tacit knowledge.
Knowledge may be tacit for a variety of reasons. It might be that the knowledge is not made explicit because we do not wish to do so or because we are not aware that someone is missing some important bit of information (Collins Reference Collins2010: 91–98). For instance, if scientists fail to replicate an experiment, this is often because they have either intentionally or unintentionally been given incomplete instructions (Collins Reference Collins1985, Reference Collins2001). In practice, knowledge is also frequently transmitted tacitly because it is easier to show someone how to do something than give verbal instructions. Oftentimes, however, much of the knowledge we have remains tacit for more fundamental reasons. This might be because our brains are not structured in a way that allows us to process this knowledge (Collins Reference Collins2010: 99–117). For this reason, for instance, we cannot learn how to balance on a bike from a book. Other knowledge might be non-explicable because it “can only be acquired through social embedding in society” (Collins Reference Collins2010: 123). This is why we will still have difficulties navigating the traffic in China on our bike if we have learned to ride it in Germany.Footnote 5
Tacit knowledge is not just essential for our daily life; it also plays a crucial role in science. Doing science is a skill that can only be learned from and in conversation with others. Further and critically, tacit knowledge plays a role both in the discovery and in the justification process in science (Sojka Reference Sojka2023: 176f., Soler Reference Soler2011). The former concerns the practical skills of doing science: mastering the specific scientific methods of one’s domain, being familiar with handling specific instruments in and outside of a lab, and becoming all-round proficient at performing the practical tasks performed by a scientist in the specific field. These are skills that fundamentally cannot be learned from books. It is the reason why students of science spend so much time on lab work. They are skills that can only be acquired in practice by observing those who already master them and by trying them out oneself (Polanyi Reference Polanyi1958: 55).
Even fully trained scientists often have difficulties establishing particular protocols or novel experimental practices successfully in their lab at first, even if they have written or oral instructions. Frequently, scientists run into problems that they can only solve by visiting the labs of other research groups (Collins Reference Collins1974) or by taking a trial-and-error approach (Collins Reference Collins2001). Oftentimes, there are some factors essential for the experiment to succeed that remain hidden to the scientist, as they are done unconsciously and accidentally (Collins Reference Collins1990).
Through doing science, scientists also develop a sort of ‘intuition’ about which research approaches might be fruitful and which can be disregarded. In practice, this intuition is an essential feature of the day-to-day life of a scientist. Scientific progress would advance significantly slower, if not grind fully to a halt, if scientists always explored every possible research route. This would also practically significantly reduce science’s relevance to society, as it would make it difficult to find quick solutions in a crisis.
Further, as Léna Soler (Reference Soler2011) points out, scientists also rely on tacit knowledge when justifying scientific research results. Based on her conversations with scientists, Soler concludes that through practicing as scientists, they develop what she calls a “compass,” for example, for when to consider a hypothesis sufficiently tested to accept it. From our own work with scientists, we can also – at least anecdotally – report that the same applies to decisions regarding the processing and interpretation of data, for instance, when it comes to distinguishing noise from signal in a raw dataset. There is usually some epistemic leeway in these sorts of decisions, and scientists have to resort to their experience. Importantly, Soler notes, how this “compass” works is at its core opaque to the scientists (Reference Soler2011: 406). That is, if pressed to rationalize their decision-making process (which rarely happens in day-to-day science for obvious reasons), they will, in the end, only be able to resort to their experience or their ‘intuition’ that the decision is the right one. Tacit knowledge is also essential to less conventional experimental practices, such as the construction and evaluation of computer models. Lenhard (Reference Lenhard, Gramelsberger, Lenhard and Parker2020) argues that climate scientists develop a “feeling” for the models that is sharpened through working with the models. This feeling makes it practically possible for scientists to work with models of which they do not have a full “analytic understanding” (Lenhard and Winsberg Reference Lenhard and Winsberg2010).Footnote 6
Importantly, though the tacitness of these decisions might make them seem like an arbitrary choice or a source of bias, such concerns are unwarranted. Just because the decisions are not fully explicable does not mean that scientists do not have good reasons to make these decisions the way they do (Sojka Reference Sojka2023: 188f.). Quite the contrary, these decisions are still scrutinized by other scientists and are evaluated through the various forms of peer review, just like any other (fully explicable) claim made by someone of a specific scientific community. However, in critically assessing these decisions, scientists also rely (in part) on specialist tacit knowledge.
Collins and Evans (Reference Collins and Evans2007) emphasize the significance of “linguistic socialization” to acquiring specialized tacit knowledge (Watson, Reference Watson2021). Through talking to and discussing with experts, apprentices learn the unique language and conventions specific to the respective discipline. An essential skill one learns this way is how to assess the quality of an argument made by a peer. For that, one needs to know what specific terms entail and the reputation of specific publications and particular scientists, as well as a good understanding of the non-explicit features of the research. This is knowledge that cannot be acquired by studying the literature in solitude. Specialist tacit knowledge gives one the skill to ‘read between the lines’ of a scientific paper and understand what is omitted and what is implied. It also gives one the skill to situate a specific argument in the general debate.
We consider this a particularly apt approach to understanding expertise in the context of epistemic trespassing for two reasons. First of all, it helps us to better understand what essential element of expertise epistemic trespassers are actually missing. While some (if not many) epistemic trespassers also lack important explicit knowledge, what they are all missing by definition is specialist tacit knowledge, as they are not (or no longer) members of the specific scientific community. We will further expand on what follows from this for the definition of epistemic trespassing and the debate about it at the end of this article. For now, we are interested in what insight we can infer from the fact that specialist tacit knowledge is essential to expertise for the recognition problem. Secondly, we argue that it tells us something fundamental about what we should look out for when we are concerned about epistemic trespassing or fake experts more generally: experience. Based on their assessment of the significance of acquiring specialist tacit knowledge through immersion in a specialist community, Collins and Evans (Reference Collins and Evans2007: 68) argue that experience can function as “meta-criteria” to distinguish between experts and non-experts. Collins and Evans argue that experience is a better meta-criterion than, for example, track record or credentials, as it also includes experts who accumulated relevant experience through unconventional means and who have not (yet) emerged as recognized experts in the domain. In science, this might, for instance, apply to the inclusion of indigenous knowledge in climate science (e.g., Head et al. Reference Head, Adams, McGregor and Toole2014; Nunn and Reid, Reference Nunn and Reid2016). People who hold such knowledge often do not have credentials like a PhD or a track record. Nevertheless, their expertise is considered very valuable for climate change assessment (Chen et al. Reference Chen, Rojas, Samset, Cobb, Diongue Niang, Edwards, Emori, Faria, Hawkins, Hope, Huybrechts, Meinshausen, Mustafa, Plattner and Tréguier2021: 177 f.).Footnote 7
The argument we want to make here, however, is somewhat different. Instead of arguing that experience is one auxiliary criterion among others for identifying experts as non-expert, we propose that criteria such as track record, credentials, or peer recognition are more accurately understood as indicatorsFootnote 8 of experience. As non-experts, we usually cannot determine whether a particular statement by a purported expert in a specific situation is true. However, we can assess whether the person has the required experience to be an expert in the first place. Experience has to be understood in a broad sense here; it includes all the activities (from talking to practicing science) one undertakes to acquire and maintain the required specialist tacit knowledge in a recognized scientific domain. This includes specialist (formal or informal) training (Watson Reference Watson2021, Reference Watson2022a) but also crucially continual immersion in the expert community. Credentials, peer recognition, track record, and dialectical superiority are indicators of whether someone has the experience in a specific domain. Credentials can indicate that someone has undergone some formal training in the course of which they would have acquired the necessary experience needed to become an expert. Peer recognition can indicate that someone is accepted as a member of a scientific community in which one maintains the relevant experience. A good track record can indicate that someone has, in the past, had the required experience to make correct expert judgments in a domain. And dialectical superiority can indicate that someone is proficient in the language of a particular domain, which can only be acquired by being immersed in a particular expert community.
2.2. Indicators of experience
In what follows, we identify and discuss the most frequently cited indicators. For a more in-depth overview and discussion, see Anderson (Reference Anderson2011), Collins and Evans (Reference Collins and Evans2007), Croce and Baghramian (Reference Croce and Baghramian2024), Goldman (Reference Goldman2001), Grundmann (Reference Grundmann, Lackey and McGlynn2025), Guerrero (Reference Guerrero and Peels2016), and Walton (Reference Walton1997). These indicators have been taken up in related debates regarding assessments of purported experts – for example, in contexts involving public trust in science, the democratic legitimation of science-based policies, and the need to distinguish genuine experts from epistemic trespassers and fake experts.
One of the most widely cited indicators of experience is academic credentials (Anderson, Reference Anderson2011: 146; Collins and Evans, Reference Collins and Evans2007: 67 f.; Goldman, Reference Goldman2001: 97 f.). These refer to academic degrees – ranging from bachelor’s to doctoral levels – and serve as indicators of specialization and sustained experience in a particular field. Credentials often encompass achievements such as peer-reviewed publications, competitive grants, and academic prizes – all of which depend on (often anonymous) evaluation by other experts.
A second indicator is dialectical superiority. Goldman (Reference Goldman2001: 95) suggests that non-experts observe how the disagreeing experts respond to one another. Experts displaying dialectical superiority typically answer with precision and clarity, using appropriate technical language and domain-specific vocabulary (Collins and Evans Reference Collins and Evans2007).
A third and frequently cited indicator of experience is a scientist’s track record (Collins and Evans Reference Collins and Evans2007: 46; Goldman Reference Goldman2001: 106; Guerrero Reference Guerrero and Peels2016: 173). For example, when researchers accurately predict a solar eclipse or how a virus will spread through a population, the public can verify the outcome without understanding the underlying models or methods.
Another underexplored indicator of relevant experience is sustained engagement with a topic, such as indigenous communities with deep medical knowledge or patients with rare diseases who become experts on their own condition. We argue that sustained engagement is a particular significant indicator of experience, as it applies also in those instances where experts neither have (official) credentials nor (accessible) track records (Collins and Evans Reference Collins and Evans2007).
There are four central limitations to these frequently cited indicators that we address here. The first is their context dependency: some indicators are limited either by the type of scientific domain they address or by the presence (or absence) of formalized systems for recognizing expertise. For instance, non-experts can only meaningfully refer to a purported expert’s track record in predictive disciplines. In fields such as biology, mathematics, or philosophy, where the primary work does not involve making predictions, this indicator is less applicable (Croce and Baghramian Reference Croce and Baghramian2024: 6). Similarly, credentials – though highly accessible – are only useful in domains with formalized systems for credentialing expertise. As Collins and Evans (Reference Collins and Evans2007) note, even in domains where most experts hold academic credentials, there are cases where individuals acquire substantial expertise without ever being awarded an official title.
The second limitation concerns the limited accessibility of information needed to assess an expert’s dialectical superiority, track record, or sustained engagement. Scientific debates are rarely conducted in public or made available for scrutiny, restricting non-experts from evaluating the dialectical superiority of (purported) experts. Moreover, unlike academic credentials, there is no formalized system for recording a scientist’s track record or sustained engagement; such records either do not exist or are not publicly accessible (Guerrero Reference Guerrero and Peels2016: 175).
Which brings us to the third limitation. Even if such records were publicly accessible, they would still be difficult for laypeople to interpret, as doing so entails a requirement of esoteric knowledge of the academic system. Consider academic credentials: they are highly accessible, yet this very accessibility can be exploited by epistemic trespassers. For example, a professor of virology may have numerous publications and research projects to their name regarding a specific virus, yet – as Drosten notes – still lack the expertise to comment responsibly on other viruses or related domains. Such credentials reflect experience only within a specific subfield (Hennig and Drosten Reference Hennig and Drosten2020). Assessing whether this background qualifies someone to speak on coronaviruses, or another virus altogether, requires time and familiarity with technical distinctions. The same applies if past track records were publicly accessible: a strong track record in one domain does not guarantee success in another, nor even in future work within the same domain. Caution is particularly required if someone leaves the field. Remember, maintaining expertise usually requires engaging with the specific expert community. Likewise, information on sustained engagement is only meaningful if one can judge whether it involved acquiring the relevant specialist tacit knowledge.Footnote 9
The fourth limitation concerns the risk of instrumentalization of the indicators themselves to create the appearance of experience, building on the challenge of interpreting them without specialized knowledge. Actors may strategically invest in the creation of institutions, journals, or conferences designed to signal academic credentials.Footnote 10 Similarly, individuals seeking to appear dialectically superior can be trained to project confidence, demonstrate rhetorical skill, and obscure topic shifts – regardless of the actual strength of their arguments (Guerrero Reference Guerrero and Peels2016: 171).
The indicators discussed so far – academic credentials, peer recognition, dialectical superiority, track record, and sustained engagement – aim to capture whether someone has relevant experience. While expertise itself is often inaccessible to non-experts, these indicators provide indirect ways for laypeople to assess it. Yet, there is something else about experts that laypeople can also evaluate, which is not about experience as such: epistemic integrity, respectively the lack of epistemic integrity. Behaviors such as failing to disclose conflicts of interest or fabricating data are sometimes treated as indicators of a lack of expertise (Croce and Baghramian Reference Croce and Baghramian2024; Goldman Reference Goldman2001), but we argue they point to something different. We suggest that these additional indicators reveal something distinct from experience and competence. Specifically, they can sometimes gauge the extent to which a person adheres to principles of good scientific practice – that is, whether the person demonstrates epistemic integrityFootnote 11 .
3. Indicators of lack of epistemic integrity
Drawing on the work of authors such as Anderson (Reference Anderson2011), Goldman (Reference Goldman2001), and Guerrero (Reference Guerrero and Peels2016), we propose that epistemic integrity is best assessed in negative terms. For non-experts, it is often more practical to watch for red flags that signal a lack of epistemic integrity than to try to positively confirm it. These include documented instances of scientific misconduct (such as plagiarism, data falsification, or fabrication), refusal to share data or methodology without good reason, or evasion of peer review. Additional concerns include undisclosed conflicts of interest – both material, such as funding, invitations, or gifts, and non-material, such as close personal relationships within professional contexts. It is often argued that non-experts can still assess whether secondary interestsFootnote 12 introduce significant bias. For example, consistent patterns of promoting claims that align with a sponsor’s interests may serve as a warning sign, even without expertise in the scientific field (Anderson Reference Anderson2011: 147 f.). Another indicator for a lack of epistemic integrity is the systematic exclusion of certain perspectives – whether due to bias or testimonial injustice (Fricker Reference Fricker2007).
A lack of epistemic integrity can also manifest in public discourse through deliberate misrepresentation of other researchers’ views (e.g., cherry-picking data or quoting out of context), the repeated defense of refuted claims, presenting oneself as an expert outside one’s domain of expertise (i.e., epistemic trespassing), or voluntarily affiliating with individuals who engage in such practices or who openly support conspiracy theories and science denialism. A related reason for skepticism is when someone bases their claimed expertise on fields widely regarded as pseudoscientific – such as alchemy or homeopathy. We would like to add that a positive indicator of a scientist’s epistemic integrity is their refusal to answer questions that exceed the limits of their own expertise. Interestingly, this is also an indicator of their experience, since one needs experience to accurately judge the boundaries of one’s own expertise – thereby actively avoiding epistemic trespassing.
As with the indicators for experience, there are also limitations regarding the indicators used to assess a purported expert’s lack of epistemic integrity. Similarly, a first major obstacle is the limited accessibility of relevant information. As there is no formalized system for documenting whether a researcher has failed to disclose conflicts of interest, exaggerated findings due to sponsorship, or misrepresented data (Guerrero Reference Guerrero and Peels2016: 174).
Second, even if such thorough publicly accessible documentation were available, the requirement of esoteric knowledge – whether of the academic system in general or of the specific discipline in question – would remain. Although interpreting records of a lack of epistemic integrity does not require expertise in the relevant scientific field, it is still far from straightforward. Laypeople may find it difficult to distinguish genuinely concerning patterns from legitimate scientific disagreement or correction. For instance, retractions or revised claims can be misinterpreted as signs of misconduct rather than as evidence of responsible self-correction. This is illustrated by the evolving assessment of COVID-19 vaccines: early public communication often emphasized their potential to significantly reduce transmission, but subsequent studies showed that while the vaccines were highly effective in preventing severe illness, their impact on infection rates was more limited than initially hoped. Some observers wrongly took this shift as evidence of deception or incompetence, rather than as a normal part of scientific refinement under conditions of uncertainty (Goldin and Fichera Reference Goldin and Fichera2022).
Third, there is the vulnerability to politicization: accusations of a lack of epistemic integrity are themselves easily weaponized. In polarized contexts, where laypeople are confronted with conflicting testimony – such as during the debate on public health measures regarding COVID-19 – accusations of lacking epistemic integrity may be strategically employed to discredit researchers.Footnote 13 This can occur even when the allegations are unfounded.
A common criticism of these negative indicators of epistemic integrity is that they focus primarily on individuals. However, non-experts may struggle to identify group-level biases within a (sub-)discipline or to recognize the exclusion or underrepresentation of certain perspectives due to bias or epistemic injustice (Fricker Reference Fricker2007; Goldman Reference Goldman2001: 105). Moreover, structural problems within the scientific system itself can give rise to epistemic distortions that are largely opaque to outsiders (Contessa Reference Contessa2023: 2951). In particular, one structural problem within science is the way publications are incentivized. This can lead to publication bias and to changes in how research is communicated. For example, studies show an increase in the use of strongly positive or negative adjectives in abstracts, where previously more neutral language was used to describe findings (Vinkers et al. Reference Vinkers, Tijdink and Otte2015). These developments make it even more difficult for non-experts to assess the significance of research results.
The indicators for experience and lack of epistemic integrity discussed here may not be the only kinds of indicators. Our aim is not to present an exhaustive list but to summarize the strengths and weaknesses of the most commonly cited indicators in the existing literature and public discourse. It reveals that these indicators should not be treated as a checklist for expertise, for two reasons: first, they are not a fixed set of criteria that every expert has to fulfill. As we have shown, there may be experts who do not have, for instance, credentials or a track record. Further, there can be legitimate reasons for intentionally not fulfilling the standards set by certain indicators – for instance, withholding data due to intellectual property concerns or national security. Moreover, in many cases assessing whether a purported expert meets specific indicators can be difficult for outsiders because they lack the necessary esoteric knowledge and information. In fact, applying these indicators in practice requires a context-specific, case-by-case assessment of their applicability and usefulness.
Second, and more importantly, what our discussion of these indicators has shown is that they are not indicators of expertise at all. Rather, they are indicators of experience and (lack of) epistemic integrity. Distinguishing between these kinds of indicators is not just conceptually useful. We would also like to draw the reader’s attention to another aspect about the process of expert assessment that our analysis of these indicators exposes: evaluating the experience of a purported expert as a layperson or even their expertise as another expert requires first assessing their epistemic integrity. If a person is systematically biased or has manipulated their research, it becomes much harder for experts to assess their expertise and for non-experts to assess their experience. Indicators of a lack of epistemic integrity reveal whether any further assessment, be it of expertise or of experience, is even possible. In this sense, we argue, epistemic integrity serves as a precondition for the meaningful application of other indicators and any assessment of expertise.
In the following, we argue that we can even further specify the function of these indicators. We suggest, taken together, indicators of experience and of a lack of epistemic integrity serve as indicators of epistemic trustworthiness. To support this, we turn to the literature on public epistemic trust in science, where a similar distinction has been discussed.
4. Indicators of epistemic trustworthiness
Trust plays a crucial role not only in science but in society more broadly. It forms an invisible matrix that enables cooperation and coexistence. We cannot do everything ourselves, nor can we verify every piece of evidence or acquire the necessary expertise in every domain. We also cannot constantly live in fear or anxiety in areas where such verification is not possible. Trust is essential for mental well-being; it facilitates the division of labor (Fukuyama Reference Fukuyama1996; Gambetta Reference Gambetta1988), and it reduces social complexity (Luhmann Reference Luhmann1968). However, it is important to note that trust is only instrumentally good: it unfolds its constructive power only when placed in those who are trustworthy. We argue that the ongoing debate about how laypeople can identify experts is, at its core, a debate about how to identify experts who are worthy of trust.
For the purposes of this paper, we will not delve into the ontological and ethical debate about what trust is or how it can be distinguished from mere reliance.Footnote 14 What we adopt from the general literature on trust is the understanding of trust as a three-place relation: A (the trustor) trusts B (the trustee) with regard to x (the object of trust). In the context of public epistemic trust in science, this can be understood as (members of) the public trusting (members of) the scientific community with regard to specific knowledge claims. For example, Anna may trust Prof. Drosten with regard to his statements about the infection rate of coronaviruses.
There are two notable particularities about public epistemic trust in science. First, we are often dealing with testimonial or speaker’s trust, where the object of trust is not an action but a proposition.Footnote 15 Second, we encounter a wide range of actors on both sides of the trust relationship – on the trustor’s side, for instance, individual citizens, groups within the public, public institutions, or ‘the public’ more broadly; and on the trustee’s side, individual scientists, research teams, scientific institutions, or science more broadly.
According to Hardwig’s (Reference Hardwig1991) influential account of the role of trust in knowledge, epistemic trust occurs when A believes a proposition p on the basis of B’s testimony. For this to be epistemically successful: (1) B must know that p, (2) B must communicate this truthfully to A, and (3) A must have good reasons to believe that conditions (1) and (2) are satisfied.Footnote 16 The first two conditions specify the circumstances under which the trustee is trustworthy.
The first condition reflects that the trustee must possess the relevant expertise – in particular, the ability to judge what qualifies as good reasons to believe that p (Hardwig Reference Hardwig1991: 700). As we have argued, such expertise typically requires experience in the respective domain, enabling the acquisition of relevant tacit knowledge.Footnote 17 As in the literature on expertise, the debate on public trust in science widely acknowledges that laypeople are often not in a position to judge whether an expert believes p for good reasons. Irzik and Kurtulmus (Reference Irzik and Kurtulmus2019), for instance, emphasize that while expertise is a necessary condition for epistemic trustworthiness, it typically remains opaque to non-experts. For this reason, they propose that laypeople must rely on what they term second-order criteria – features that serve a similar function to what we call indicators of experience.
The second condition concerns the trustee’s epistemic integrity – specifically, their commitment to the public’s interest in being honestly and accurately informed about research findings and their limitations. As with expertise, we argue that epistemic integrity is difficult for laypeople to assess directly, which is why indicators play an important role here as well. A truthful transmission of knowledge is seriously compromised when someone systematically avoids peer review, conceals conflicts of interest, or deliberately misrepresents their own or others’ research findings.
This requirement – that scientists act with honesty and in the public’s interest – has prompted debate about whether trust in science is necessary in the first place. Some argue that it would be sufficient for the public to merely rely on scientists’ expertise, without needing to place trust in their willingness to serve the public’s interest by communicating research accurately and transparently. According to this view, it is enough to assume that scientists will behave reliably out of self-interest – because the institutional structure of science provides incentives for reliable behavior (Blais Reference Blais1987: 370; Koskinen Reference Koskinen2020: 1193). Scientific misconduct, so the argument goes, would be detected through self-correcting mechanisms such as peer review and replication. In order to avoid the reputational damage and sanctions that follow from exposure, researchers are motivated to act reliably, even without an explicit commitment to the public’s interest. We believe there are compelling reasons to hold that the relationship between scientists – and especially between science and the public – extends beyond mere reliance on expertise and is, in fact, one of trust.
First, Hardwig (Reference Hardwig1991: 707) argues that it is highly unlikely that institutional control mechanisms could be developed that reliably detect all forms of scientific misconduct. He notes, on the one hand, practical limitations – for instance, that well-executed data fabrication may be extremely difficult to uncover. On the other hand, he points out that even if such mechanisms existed, one would ultimately have to trust the integrity of those operating them. Thus, trust – and with it the condition of epistemic integrity on the part of those involved in science – remains essential, albeit possibly redirected to different actors within the system.
Second, Wilholt (Reference Wilholt2013: 16–18) emphasizes that in a research process based on division of labor and cooperation, it is insufficient to rely solely on researchers’ adherence to methodological standards. He highlights that hypothesis acceptance or rejection is often not strictly codified and inevitably involves uncertainty. Following Rudner (Reference Rudner1953), Wilholt argues that such decisions require value judgments – especially when the consequences of accepting a false hypothesis or rejecting a true one affect individual or public well-being. In such cases, trust is required because expertise alone is not sufficient; what is also needed is that scientists are willing to take public interests into account when weighing inductive risks (Douglas Reference Douglas2009: 112).
We conclude this excursus on trust in science by emphasizing the third condition of Hardwig’s definition of epistemic trust: it specifies when trust in science is justified. Trust is justified when there are good reasons to believe that a scientific actor possesses both the relevant expertise and epistemic integrity. In the debate on trust in science, several of the indicators discussed above are cited as providing such reasons. They help the public – whether individual citizens, groups, or institutions – assess whether a scientist’s testimony is trustworthy. As we argue in this article, these are not indicators of expertise but of experience and (lack of) epistemic integrity. However, together, we maintain that they should be understood as indicators of epistemic trustworthiness.
5. Returning to epistemic trespassing
We began this article by considering the concept of epistemic trespassing. We now return to it to see what we can infer from our analysis of the indicators of epistemic trustworthiness about the definition of epistemic trespassing. Much of the philosophical debate about epistemic trespassing has been concerned with the question of when it is epistemically permissible or even desirable for experts to venture outside their domain and when it is epistemically problematic (Gerken Reference Gerken2023; Pavličić et al. Reference Pavličić, Dimitrijević, Vučković, Đorđević, Nedeljković and Tešić2024; Watson Reference Watson2022b). Taking into account that having specialist tacit knowledge is a necessary condition for expertise helps us further understand where and when epistemic trespassing ‘goes wrong’.Footnote 18 In many cases, harmful epistemic trespassing occurs despite the fact that the trespassing person has read some scientific publications on the topic. They may claim to “have done their own research,” but they do not have the required specialist tacit knowledge to properly interpret what they have read. Note that Collins (Reference Collins2014) points out that having just what he and Evans (Reference Collins and Evans2007) call “primary source knowledge” is particularly dangerous because it can give one the false impression that one understands more about a specific subject matter than one actually does. This fits well with Levy and Varley‘s claim that those who engage in epistemically dangerous trespassing usually do not “possess the capacity to defer appropriately” (Reference Levy and Varley2026: 165). That is, they do not take “opinions and the shape of debates very seriously in deliberation, and as appropriate caution in dissent” (Reference Levy and Varley2026: 153). The “skills of apt deference” are, as Levy and Varley rightly note, acquired tacitly “as a byproduct of professionalization” (Reference Levy and Varley2026: 161).
We might also add that many of the standards and conventions of science are epistemically underdetermined, so they are established and adjusted by scientists through being in conversation with each other. As many of the standards and principles of science are fluid and change over time, they can only be acquired by being active members of a scientific community. This is why even scientists who were once active members of a scientific community – but no longer are – risk epistemic trespassing, because they are no longer up-to-date on the necessary specialist tacit knowledge required to assess new arguments made in the domain in which they were once experts.
This does not mean that detrimental epistemic trespassing automatically occurs when scientists engage in interdisciplinary research. Quite the contrary, what makes interdisciplinary research so valuable is that scientists with a variety of different perspectives and experiences look at the same problem. But in these instances, these experts all have important specialist tacit knowledge that pertains directly to a specific question or problem. However, as anyone who has ever engaged in interdisciplinary research can attest, it requires quite a bit of work to bring everyone involved up to speed on the necessary specialist tacit knowledge to overcome communication hurdles, because the specialist language, methods, and standards differ across disciplines. Further, interdisciplinary research is most successful if everyone involved is aware of the limits of their own knowledge and values the knowledge that everyone else brings to the table. If this does not happen, they risk engaging in counterproductive epistemic trespassing. This also means that experts have good epistemic reasons to step into debates in other domains if and only if their experience means that they can contribute to the discussion, even if their opinion contradicts prevailing beliefs in the specific domain.
To avoid the occurrence of harmful epistemic trespassing, there are certain standards of good scientific practice that we expect scientists to adhere to. We require them to show signs of “apt deference” (Levy and Varley Reference Levy and Varley2026) by considering other experts’ assessments and opinions and, very importantly, by showing epistemic prudence when it comes to knowing what they don’t know. By contrast, genuine, problematic epistemic trespassing occurs when an expert in one domain wanders into a different domain where they have no relevant specialist tacit knowledge related to the problem or question at hand. Yet, they pretend, or are convinced, that they have the necessary experience. That is, they neither have the required experience nor the required epistemic integrity. As we have shown in this paper, there are indicators of both experience and a lack of epistemic integrity that non-experts (at least in principle) can apply to assess whether someone is a fake expert.
Acknowledgments
We thank Gottfried Vosgerau for discussing various earlier versions of this paper with us. We also received helpful comments from Alexander Christian, Chiara Lisciandra, and Nastasia Müller. Maria Sojka’s research for this paper was supported by the research project ACCeSS, which is funded by the Ministry of Culture and Science of the State of North Rhine-Westphalia in Germany (Förderlinie: Profilbildung2022, Förderkennzeichen: PB22_028B).
Open access
