2. Understanding: a preliminary account

What is understanding, and how does it relate to disinformation? Let us begin with the first question. The philosophical literature recognizes different forms of understanding (Kvanvig Reference Kvanvig2003; Elgin Reference Elgin2007; Gordon Reference Gordon2012; Hills Reference Hills2016; Baumberger et al. Reference Baumberger, Beisbart, Brun, Grimm, Baumberger and Ammon2016). A common distinction is between propositional understanding, which takes a proposition as its object – as in “Lucas understands that it is time to go” – and objectual understanding, which in Kvanvig’s words, occurs “when understanding grammatically is followed by an object/subject matter, as in understanding the presidency, or the president, or politics” (Reference Kvanvig2003: 191). Recently, much attention has been given to explanatory understanding – or understanding-why – which is directed at a phenomenon or a question, as in “Alice understands why the space shuttle Challenger exploded” (Grimm Reference Grimm2008; Khalifa Reference Khalifa2012; Strevens Reference Strevens2013; Hills Reference Hills2016). For our purposes, however, it is not necessary to engage with the broader debate about different types of understanding or how they interrelate (for an overview of the debate, see Hannon Reference Hannon2021). In this paper, I would like to focus on understanding how things work and why they work the way they do. This may encompass elements of objectual understanding, explanatory understanding, or an overlap between the two.

How can we sketch an approach to understanding in broad strokes? While there is widespread disagreement in the literature on how to develop a detailed account of understanding, philosophers of different persuasions generally agree on one point: understanding is not merely a matter of knowing a list of facts. Rather, it involves a kind of grasping. More specifically, it involves grasping the connection between various pieces of information. Kvanvig (Reference Kvanvig and Starr2018: 699) puts the point in the following way: “Central to the notion of understanding are various coherence-like elements: to have understanding is to grasp explanatory and conceptual connections between various pieces of information involved in the subject matter in question”.

Elgin (Reference Elgin2007: 35) makes a similar point:

Grimm (Reference Grimm and Fairweather2016: 214) helpfully notes that what is grasped in understanding is a structure, such that in understanding, “one apprehends how the various elements of a system depend upon one another, so that one can potentially manipulate the system in various ways”.

Does every piece of information that constitutes the object of understanding need to be true? In other words, is understanding factive in the same way that propositional knowledge is? Elgin (Reference Elgin2007, Reference Elgin2017), for example, argues in favor of a non-factive account, according to which understanding can incorporate felicitous falsehoods. According to her, this is evident in scientific practice, where idealizations and models frequently play a central role. Many of our best scientific theories rely on deliberate falsehoods. For instance, the behavior of real gases is often explained by appeal to ideal gas, a theoretical construct that assumes properties that are hard to discern in actual gases (Elgin Reference Elgin2007: 40).

However, as Hannon (Reference Hannon2021) points out, we can acknowledge the role of idealizations in fostering understanding without committing to the stronger claim that deliberate falsehoods constitute the object of understanding. As he puts it, “a false theory may foster genuine understanding, but the phenomenon that is understood (or the predictions one is attempting to make) must still be true – or at least approximately true” (Reference Hannon2021: 278). If this is correct, we can reconcile Elgin’s insights about the function of idealizations in scientific practice with the factivity of understanding.

This is not the place to settle this debate. For present purposes, I will assume that understanding is factive – that is, that its objects are true propositions. However, not much hinges on this assumption. My central argument can be easily reframed in non-factive terms.Footnote ³

What about the relation between understanding and knowledge? A long-standing debate in the literature concerns whether understanding can be reduced to knowledge-why or knowledge of causes – a view tracing back to Aristotle and widely endorsed in the philosophy of science (see, e.g., Lipton Reference Lipton, de Regt, Leonelli and Eigner2009; Khalifa Reference Khalifa2017). However, a number of epistemologists argue that understanding is not in fact reducible to knowledge (see, e.g., Kvanvig, Reference Kvanvig2003; Elgin Reference Elgin2007, Reference Elgin2017; Pritchard Reference Pritchard2009; Hills Reference Hills2016). Hills (Reference Hills2016: 662), for example, argues that understanding – her focus is on understanding-why – “is different from this standard conception of propositional knowledge: it is consistent with luck, it can be based on defeated evidence and it cannot (easily) be transmitted by testimony”. In this paper, I remain neutral on this issue. My argument does not hinge on whether understanding reduces to a specific type of knowledge, such as knowledge-why. What it does require is the recognition that understanding is not merely the knowledge of a collection of discrete factual propositions, devoid of a grasp of the relationships between them. If some form of knowledge necessarily involves grasping such relationships, then it may well be compatible with the view put forward here that understanding reduces to that form of knowledge.

We are now in position to introduce a rough account of understanding – one that can be applied across different forms of understanding:

Let me clarify this thesis by applying it to a case of understanding how something works and why it works as it does. Consider Leo, an experienced Formula 1 mechanic. Leo grasps how various components of Formula 1 cars – engine, aerodynamics, suspension, electronics, tires – interact to produce their performance. This enables him to explain how performance is generated and predict how a car will behave under different conditions. He can also explain how and why certain mechanical systems work the way they do: why minor changes in tire pressure affect grip and handling; how aerodynamic tweaks improve cornering speed; why a particular engine misfire is likely caused by an issue in the fuel injection system rather than the ignition system.

Next imagine someone who understands how government agencies such as USAID work and why they work the way they do. More specifically, consider Roberta, a policy analyst specializing in international development. Roberta grasps the institutional structure of USAID – its funding mechanisms, partnerships, bureaucratic processes and relationships with other government agencies, NGOs, and private contractors. She understands how aid is allocated, why certain regions receive more funding than others and how shifting political priorities influence strategic decisions. This allows her to explain and predict a number of aspects of USAID’s operations, such as: why humanitarian aid is often prioritized over long-term development projects in times of crisis; how congressional budget cuts affect staffing and program implementation in different regions; how diplomatic and geopolitical considerations shape development strategies, such as why aid programs in certain countries are expanded while others are scaled back.

These examples illustrate what we can expect from someone who understands how something works and why it works the way it does – whether a complex piece of technology like an F1 car or an institution. Such an individual grasps the relationships between various components in a way that enables them to explain the system’s behavior, anticipate outcomes resulting from changes in its internal workings or external conditions, identify underlying causes of problems, and so on.

Now that we have outlined a preliminary account of understanding, we are in a position to examine its relationship to the phenomenon of disinformation. This is the focus of the next section.

3. Understanding as an epistemic shield

In this section, I will argue that, all else being equal, possessing understanding of how things work and why they work the way they do can act as a safeguard against falling prey to disinformation campaigns.

Let us begin by revisiting Leo. Now, imagine that Leo and other F1 mechanics are the targets of a disinformation campaign aimed at convincing them that a particular car crash was caused by human error rather than a mechanical failure, as they currently believe. More specifically, suppose that Leo and his colleagues are exposed to the following content aimed at reshaping their views on how and why the crash occurred:

Suppose that the crash was indeed the result of mechanical failure, so that the content above is one that has a disposition to generate or increase ignorance in normal conditions regarding the causes of the crash, fitting into Simion’s (Reference Simion2024) account of disinformation. As a result of his understanding of how F1 cars work and why they work the way they do, plus the available evidence, Leo is convinced that the crash was the result of mechanical failure alone, with no fault to be attributed to the driver. Is Leo likely to fall prey to the disinformation campaign that targets him and others?

I take it that answer here is clearly “no”. Leo’s understanding of F1 mechanics should allow him to resist the disinformation campaign in several ways. First, it should equip him with the means to assess competing explanations of the crash. Since he grasps how the various components of an F1 car interact – how tire degradation affects grip, how aerodynamic balance conditions handling, and so on – he can judge whether an account of the crash coheres with his body of information on the matter. Moreover, Leo’s understanding should enable him to recognize causal patterns. Mechanical failures follow characteristic trajectories; so do driver errors. If the evidence points to a sudden loss of braking power rather than a delayed reaction time, for example, Leo is in a position to tell that the failure is mechanical and not human. Since he knows how these systems work, he knows what kind of failure to expect when they malfunction. Also, Leo’s epistemic position is strengthened by his ability to assess the evidential weight of different sources. A core epistemic role of understanding is that it should allow one to differentiate between reliable and unreliable sources. The disinformation campaign relies on selectively framed testimonies and rhetorical insinuations – radio excerpts suggesting hesitation, leaks from engineers criticizing the driver’s style. However, Leo knows that such testimonial evidence is epistemically weaker than direct mechanical diagnostics, for instance. Understanding should enable him to give proper epistemic weight to different sources and to recognize that the strongest available evidence supports the mechanical failure hypothesis.

So, given his understanding of how F1 cars work and why they work the way they do, Leo is well-positioned to resist the disinformation campaign unfolding in Car Crash. Yet external pressures – such as institutional threats or potential professional fallout – might compel him to publicly endorse the human error explanation, if only to preserve his standing within the investigation. In more acute cases, these pressures might even lead Leo to want to believe the alternative explanation. More troublingly, Leo might display a disposition to shy away from confrontation or risk at the very moment when he ought to stand firm, thereby undermining his own conviction in the mechanical failure hypothesis. Alternatively, Leo might fall prey to conformity, seeking to adjust his beliefs to align with those of influential colleagues or authoritative voices. Over time, this pull may erode his confidence in the mechanical failure hypothesis, through the slow accumulation of doubt, seeded by those around him.

This illustrates a crucial point: possessing understanding does not render one immune to disinformation. Epistemic vulnerability can stem from more than lacking understanding; it can arise from reduced self-confidence, intellectual cowardice, a disposition toward conformity, and related factors.Footnote ⁴ Nevertheless, this should not be taken as a counterexample to my central claim that, all else being equal, possessing understanding of how things work and why they work the way they do provides an epistemic shield against disinformation campaigns. When the factors mentioned above enter the picture, conditions are no longer equal. My claim is not that understanding invariably grants immunity against disinformation, but rather that it constitutes a crucial epistemic resource – a resource that, provided it remains uncompromised by external pressures or internal vulnerabilities, for example, significantly enhances one’s epistemic resilience.

Similarly to Leo, Roberta, our policy analyst specializing in international development, who understands how US government agencies such as USAID work and why they work the way they do, should be able to see through the fog in the case of a disinformation campaign against such agencies. While disinformation may be widespread – alleging corruption, inefficiency, or covert operations – Roberta’s grasp of USAID’s funding mechanisms, policy priorities, and bureaucratic constraints equips her to critically evaluate misleading narratives. Of course, external pressures, such as political influence or media saturation, may complicate her epistemic position. But all else being equal, her understanding can function as an epistemic safeguard against such campaigns.

At this point, one might object that both Leo and Roberta are specialists, and it is hardly surprising that they should be somewhat insulated from disinformation campaigns within their respective domains. But what about non-experts? Are they entirely defenseless against such campaigns? Not necessarily. Understanding how things work and why they work the way they do comes in degrees.

Consider an F1 enthusiast, someone obsessed with car mechanics but lacking direct experience with F1 vehicles. They may not be able to diagnose a malfunction by listening to an engine’s sound alone, but they might still be better equipped than most to assess competing explanations regarding the causes of a car crash. Their understanding, though partial, provides some resistance to misleading content. The same holds for someone with a passing but structured grasp of how international aid organization’s function. Even if they lack Roberta’s expertise, they may still recognize certain patterns, inconsistencies or implausibilities in a disinformation campaign. Understanding, even when partial, is often enough to place constraints on what one is willing to accept uncritically.

I would like to suggest, however, that there are cases in which a subject falls below a certain threshold necessary to be considered as having some understanding of how something works and why it works the way it does. In such cases, it is appropriate to regard the subject as lacking understanding. These, I will argue next, are precisely the cases in which individuals are particularly vulnerable to disinformation campaigns.

4. Epistemic vulnerability: when understanding is lacking

I have characterized understanding as relation between a subject S and a structured body of (true) information, such that S grasps the relationships among its elements in a way that enables explanation, inference, and application across various contexts. This broad account accommodates the possibility of partial understanding – cases in which a subject has some grasp of the relationships between elements within a body of information, allowing for a limited degree of explanatory capacity, inference-making, and cross-contextual application. But take away that grasp, and what’s left? If a subject fails to discern these relationships, if they cannot generate explanations, draw meaningful inferences, and apply information beyond the immediate case at hand, then they do not merely have less understanding; they lack it altogether.

In order to illustrate what a lack of understanding may amount to, consider the following case (I will call it Mercury’s Oddities), introduced by Wilkenfeld et al. (Reference Wilkenfeld, Plunkett and Lombrozo2016) in an experimental investigation on when and why we attribute understanding to others.Footnote ⁵

Participants in Wilkenfeld et al. (Reference Wilkenfeld, Plunkett and Lombrozo2016) were hesitant to attribute understanding to Richard in this scenario – what the authors refer to as the ignorant condition – and for good reason. We are told that Richard knows plenty of facts regarding Mercury, its orbit, and general relativity. But something crucial is missing. He fails to see how these facts connect, how they come together to explain Mercury’s peculiar orbital behavior. The structure is there, the pieces are in place, but the grasp is absent. Because he cannot integrate this information into a coherent explanatory framework, he is unable to generate explanations, derive meaningful inferences, or apply what he knows to new contexts. Thus, it seems that Richard does not understand why Mercury has some oddities in its observed orbit, or more generally, why Mercury’s orbit behaves the way it does.

I would now like to suggest that, despite knowing numerous facts about Mercury’s orbit and general relativity, Richard’s lack of understanding—his failure to grasp the connection between these facts and why Mercury’s orbit behaves as it does—may leave him vulnerable to disinformation campaigns. Consider the following case, which I will call Dark Twin:

Dark Twin seems plausible. Richard, despite possessing extensive factual knowledge, is unable to integrate those facts in a way that enables explanation, inference, and application across various contexts when it comes to the question of why Mercury has some oddities in its observed orbit, or why Mercury’s orbit behaves the way it does. The crucial point is this: had Richard understood why general relativity explains Mercury’s motion, the dark twin hypothesis would very likely have struck him as implausible from the start. He would have recognized that general relativity predicts the precise deviations in Mercury’s orbit and that there is no leftover anomaly crying out for a new celestial body. But because his knowledge is compartmentalized and devoid of a grasp of the relationships among its elements, he is left epistemically vulnerable – ripe for manipulation by a well packaged but misleading alternative.

If this is so, we can conclude that factual knowledge alone, without understanding, may not offer enough protection against disinformation. A subject may possess an extensive repertoire of factual knowledge while lacking the structured grasp that enables explanation, inference, and cross-contextual application. Without this integrative grasp, even a well-informed individual may be unable to distinguish between a well-supported theory and a misleading but superficially plausible alternative. Thus, it is not mere factual knowledge but understanding that serves as an epistemic safeguard against disinformation campaigns.

5. Disinformation and the exploitation of ignorance

As noted earlier, Simion (Reference Simion2024) offers an account of disinformation in terms of its disposition to generate or increase ignorance. More specifically, she defines disinformation as follows: “X is disinformation in a context C iff X is a content unit communicated at C that has a disposition to generate or increase ignorance at C in normal conditions” (2023: 1213).

Simion (Reference Simion2024:1214) identifies different ways in which disinformation can generate or increase ignorance:

• By presenting content that induces false beliefs.

• By providing content that undermines knowledge by defeating justification.

• By inducing epistemic anxiety.

• By reducing doxastic confidence or defeating justification.

• By conveying false implicatures.

What I want to suggest is that these means of ignorance generation or amplification often exploit a fundamental epistemic deficit: a prior lack of understanding of how things work and why they work the way they do. In many cases, disinformation’s effectiveness depends on an epistemic landscape already shaped by gaps in understanding. When individuals lack a grasp of the relationships between relevant facts, they are more susceptible to misinformation that distorts those connections, undermines justifications, or exploits epistemic insecurities.

The case of Richard illustrates this point. Richard possesses an extensive collection of factual knowledge about Mercury’s orbit and general relativity, yet his failure to grasp how these facts interconnect leaves him vulnerable to the dark twin hypothesis. Because he does not understand why general relativity explains Mercury’s orbital anomalies, he does not immediately recognize the hypothesis as implausible. The disinformation campaign exploits this epistemic gap—Richard is prone to being misled by false claims precisely because of his lack of understanding.

Of course, other forms of ignorance may also facilitate the spread of disinformation. A straightforward case is a sheer lack of factual knowledge. If one has never encountered accurate information about a given topic, one may be at the mercy of whatever claims happen to reach them first. Someone who has never heard of USAID before, for example, may be more likely to accept without question a claim that it is a covert intelligence operation rather than a development agency. Similarly, an individual unfamiliar with the principles of epidemiology may be easily swayed by misleading claims about vaccines.Footnote ⁶ Moreover, ignorance about the mechanisms of disinformation itself can further entrench susceptibility. If one is unaware of common tactics – such as the deliberate manufacture of controversy, selective presentation of evidence, or emotionally charged framing – one is less likely to recognize when one is being manipulated.

My aim here is not to provide an exhaustive taxonomy of the various forms of ignorance that disinformation exploits. My claim is a more modest one: the means by which disinformation generates or amplifies ignorance often rely on a pre-existing lack of understanding – specifically, a failure to grasp how things work and why they work the way they do. This does not imply that other forms of ignorance – such as a lack of mere factual knowledge, or unfamiliarity with disinformation tactics – do not also play a crucial role.

6. The challenge of acquiring understanding and the limits of deference

Understanding is typically difficult to attainFootnote ⁷ – certainly more so than mere factual knowledge. I can come to know that there is a cat on the table simply by opening my eyes, but I cannot acquire an understanding of how F1 cars work and why they work as they do in anything like the same effortless manner. Understanding requires more than simple observation; it demands integration of information, and a grasp of the underlying relationships that structure a given domain. As put by Hazlett, according to many epistemologists,Footnote ⁸ “understanding requires a suite of intellectual abilities, including the ability to give and follow explanations or to provide theoretical accounts, the ability to draw conclusions about similar cases, and the ability to answer explanatory and modal questions” (Reference Hazlett, Lackey and McGlynn2025: 609). But developing these abilities is far from straightforward. It certainly takes patience, practice, and often expert guidance. It can be impeded by cognitive biases, limited attention, unreliable information sources, or simply by the many other demands competing for our time.

Thus, we may appear to be in a bind. Understanding can shield us from disinformation campaigns, and so fostering it is surely desirable. Yet if understanding is characteristically so challenging to acquire, relying on widespread improvements in public understanding as our main defense against disinformation may be overly optimistic.

At this point, however, one might object that the problem is not as deep as it initially appears. After all, we can defer to those who understand how something works and why it works the way it does—or at least to someone who understands it better than we do. If understanding is difficult to attain firsthand, perhaps our best strategy is to rely on those who already possess it.

Consider Richard once again. Despite possessing a wealth of factual knowledge about Mercury’s orbit and general relativity, he lacks the understanding necessary to integrate these facts into a coherent framework. This epistemic gap leaves him vulnerable to the dark twin hypothesis. But suppose Richard, recognizing his own limitations, decides to defer to an expert in celestial mechanics—someone who not only knows the relevant facts but also grasps how they fit together to explain Mercury’s orbital behavior. By deferring to this expert, Richard could avoid being misled by the disinformation campaign, even without personally attaining the relevant understanding.

This suggests a possible solution: when we lack understanding, we can mitigate our epistemic vulnerability by deferring to those who possess it. If understanding shields one from disinformation, then relying on those who understand – scientists, engineers, historians, and policy analysts—should provide at least some degree of protection. In principle, Richard need not understand why general relativity explains Mercury’s orbit to reject the dark twin hypothesis; it should suffice for him to trust those who do.

Of course, for Richard to successfully shield himself from disinformation campaigns through deference to experts, he must first reliably identify who the genuine experts are. Without this ability, his deference risks being misdirected – potentially reinforcing rather than dispelling his epistemic vulnerability. But how should he go about doing so? The problem of recognizing expertise has received some attention in the epistemological literature (see, e.g., Goldman Reference Goldman2001; Anderson Reference Anderson2011; Guerrero Reference Guerrero2007; McKenna Reference McKenna2023). Several criteria have been proposed as indicators of genuine expertise, including formal credentials and institutional backing, a proven track record of accuracy, argumentative rigor, and alignment with the broader expert consensus.

However, as Levy (Reference Levy2022) observes, we live in epistemically polluted environments where “the cues for expertise don’t correlate well with its actual possession” (Levy Reference Levy2022: 112). A central feature of such environments is the proliferation of pseudo-expertise – where individuals and institutions successfully mimic the outward signs of genuine expertise, whatever they happen to be. This epistemic mimicry makes it increasingly difficult for laypersons to distinguish between those who in fact understand how things work and why they work the way they do, and those who merely play the role of an expert without the requisite understanding.

A striking example of epistemic mimicry can be found in Merchants of Doubt (Oreskes and Conway Reference Oreskes and Conway2010), which meticulously documents how individuals without genuine expertise successfully played the role of scientific authorities to mislead the public on issues such as tobacco, climate change, and acid rain. They did not need to possess genuine understanding of the phenomena in question; they only needed to appear authoritative enough to sow confusion.

In epistemically polluted environments, as Levy (Reference Levy2022) warns, the usual indicators of expertise—academic credentials, institutional affiliations, or media presence – may be strategically co-opted by non-experts to project credibility. The result is an erosion of the reliability of traditional epistemic cues, making it increasingly difficult for laypersons to separate genuine expertise from its mere performance. When disinformation campaigns exploit this mimicry, they do not merely introduce falsehoods; they create an epistemic fog in which deference to expertise becomes unreliable.

This complicates Richard’s situation. If he were to attempt to shield himself from disinformation through deference to expertise, he would first need to navigate an epistemic landscape rife with counterfeit authorities. Lacking understanding himself, he is especially vulnerable to mistaking the merchants of doubt for genuine experts. The very strategy that could, in principle, protect him – trusting those who understand – may in practice lead him deeper into epistemic entanglement.

The broader lesson is clear: in an environment where the cues of expertise can be co-opted, mere deference is not enough. The capacity to identify genuine understanding – and to recognize when expertise is merely being mimicked – is itself an epistemic skill. Of course, possessing understanding of how something works and why it works the way it does may by itself allow one to identify experts who have an even deeper grasp of it. If I understand how government works, I am in a better position to recognize a true specialist on government policy, say. But here we confront the same problem we started with: understanding is typically difficult to attain, and so most of us must rely on deference. Yet if deference alone is unreliable in polluted epistemic environments, where does that leave us?

The task may seem to be clear: we must clean up the epistemic environment. But doing so is no simple matter. As Levy (Reference Levy2022) puts it, we are faced with the problem that “while everyone might be better off if no one pollutes, no individual can make a significant difference on their own, and any individual who pays the cost of clean-up locally is worse off than others who don’t cooperate’ (Reference Levy2022: 127). Cleaning up the epistemic environment, then, requires large-scale cooperation. And large-scale cooperation, in turn, ‘may require some degree of coercion, from government or other institutions with the clout to impose costs on those who don’t cooperate” (Levy Reference Levy2022: 127). What policies or regulations might help address this issue is, of course, beyond the scope of this paper.Footnote ⁹

Are we, then, in a hopeless predicament – fated to remain easy prey for disinformation – if understanding is difficult to attain and cleaning up our epistemic environment equally challenging? Perhaps not. From a more optimistic vantage point, we might recall that neither task is utterly beyond our grasp. Cleaning up the epistemic environment may yet prove achievable, provided we identify and implement effective policies and regulations. Likewise, even if robust understanding is demanding, understanding comes in degrees. Some grasp of how things work and why they work the way they do – even if partial or incomplete – is surely preferable to none at all.

Moreover, even if understanding cannot simply be passed on through testimony from those who possess it,Footnote ¹⁰ testimony can nevertheless help promote it. Gordon (Reference Gordon, Grimm, Baumberger and Ammon2016), for instance, argues that speakers can facilitate understanding in their audience by encouraging “(i) the acquisition of new true beliefs, (ii) the rejection of false beliefs, (iii) the grasping of new connections (and rejection of mistaken connections), (iv) overcoming blocks to grasping, and (v) the acquisition or enhancement of abilities linked to grasping” (Reference Gordon, Grimm, Baumberger and Ammon2016: 294). Thus, even if testimony alone does not transmit understanding, it can still pave the way for its acquisition.