What's Wrong with Bullshit

1. Introduction: Some Garden Varieties of Truth-Tracking Bullshit

Sebastian Marx is an American stand-up comedian who worked and lived in France, and loves to make fun of his experience there. In one of his shows, he recalls having worked in a French restaurant as a waiter, and mocks the “super complicated” names French restaurants sometimes give to their dishes. As an example, he mentions the “tasty fisherman’s hodgepodge on a bed of finely crushed puree”, before revealing: “it’s just fish and potatoes”. He then jokes about the fact that French people would say that such names are not “complicated” but “sophisticated”, and that Americans have another word for “sophisticated”: it’s… bullshit.

Sebastian Marx is far from being the first to make fun of such pompous dish names. French comic artist Gotlib has a cartoon in which he depicts someone at a restaurant ordering “the Prince of the Seas in his Sauce of the Fruits of Provence” and looking disappointed when the waiter brings him sardines in oil. This is another paradigmatic case of bullshit. Still, it’s an interesting one because it is true – and not even accidentally true: it actually tracks truth. Indeed, there is some connection between the name of the dish and its content, to the extent that trained consumers can guess what the dish contains. Rather than inventing names at random, the person who wrote the menu intended for the names to track the content of the dishes: they would never have given the name of “tasty fisherman’s hodgepodge on a bed of finely crushed puree” to a piece of deer with green vegetables (this would rather be “the hunter’s delight on its farandole of greeneries”).

Such mundane cases of “truth-tracking” bullshit are not limited to pompous restaurant dishes. We also use it to paint ourselves in a flattering light on our resume: we do not want to be caught telling something false, but we want our past professional experience to read as much more impressive than it actually was. You spent two months entering the results of pen-and-paper surveys into an Excel sheet in exchange for credits? Well, just put on your resume that you were “Responsible for data management, integration and integrity at Pr. Academic’s laboratory” (and do not forget to add “Advanced proficiency in Microsoft Excel” in the “Skills” section). You once gave a public speech in the street just in front of the UNO building? Just say that you presented your work in front of the UNO (Aberkane 2021). Again, these are not technically false claims – rather, they are fancy ways of presenting the truth.

In the same way that fancy dish names are clear cases of bullshit, fancy resumes are a paradigmatic case of bullshitting. And both cases seem to point at one crucial feature of bullshit: bullshit is something that sounds impressive at first sight but can be easily “deflated” on closer inspection. However, I will argue that existing philosophical accounts of bullshit have largely ignored these paradigmatic examples of bullshit.¹ In this paper, I will present a new account of bullshit that takes these examples into account and argue that this allows us to better understand what exactly the problem is with bullshit.

2. Limitations of Past Accounts of Bullshit

The examples I put forward have two purposes. The first is to highlight some limitations of former philosophical accounts of bullshit. As I will argue in this section, the most prominent philosophical accounts of bullshit fail to account for paradigmatic cases of “truth-tracking” bullshit.

3. An Alternate, Output-Based Proposal

Thus, it seems that most existing accounts cannot accommodate cases of “truth-tracking bullshit.” Now, all the accounts I discussed so far are “process-based” rather than “output-based”—meaning that they propose to define bullshit not based on the properties of the statements that are produced, but rather on the process that lead the agents to produce these statements. Maybe an “output-based” approach, focusing on the properties of the statements themselves, would be more successful.

This distinction between two approaches to bullshit comes from Cohen (2006). He bases this distinction on the Oxford English Dictionary, which gives the following two definitions of bullshit: (1) “nonsense rubbish,” and (2) “trivial insincere talk or writing” (Cohen 2006: 120). Whether something is bullshit in the first sense depends on the properties of what is produced, while whether something is bullshit in the second sense depends on the intentions, mental states and attitudes of the person who produces it. In this view, bullshit in the first sense is a noun that emphasizes the worthlessness of bullshit, while bullshit in the second sense is a verb that focuses on the process of bullshitting (Cohen 2006: 120–121). The first sense is output-based, while the second is process-based.

According to Cohen, Frankfurt’s process-based account is insufficient and should be completed with an output-based account of bullshit. He takes as an example his own experience with French Marxism, and more precisely with the Althusserian school:

The ideas that Althusserians generated … possessed a surface allure, but it seemed often impossible to determine whether or not the theses in which those ideas figured were true, and, at other times, those theses seemed capable of just two interpretations: on one of them, they were true but uninteresting, and, on the other, they were quite interesting, but obviously false. (Failure to distinguish those opposed interpretations produces an illusory impression of interesting truth). (Cohen 2006:118)

For Cohen, such obscure philosophical texts are paradigmatic cases of bullshit, but Frankfurt’s account of bullshit does not really capture them: it is not clear that people who produce such sentences are not concerned with truth. In fact, Cohen vouches for them, claiming that at least some of the philosophers who produce these kind of sentences are “honest thinkers.”

However, Cohen does not take such examples to constitute a counter-example to Frankfurt’s analysis of bullshit. Rather, Cohen concludes that we should distinguish between two different types of bullshit: bullshit in ordinary life, which is best captured by Frankfurt’s approach, and bullshit in academic work, which requires another kind of approach. To capture the second kind of bullshit, Cohen proposes that a statement or text is bullshit when it is nonsense, that is when it is

by nature unclarifiable, discourse, that is, that is not only obscure, but which cannot be rendered unobscure, where any apparent success in rendering it unobscure creates something that isn’t recognizable as a version of what was said. (130)

One objection often leveraged against Cohen’s definition of bullshit is that it is very narrow and that a lot of what we consider bullshit still has meaning. In fact, it is not even clear that Cohen’s proposal can account for his experience with Althusserian texts: according to him, such texts are clarifiable but, depending on how they are clarified, their claims are either obviously false or trivially true. As such, they correspond to what Dennett (2009) called deepities: misleading statements that can be interpreted in two ways: a true but trivial one, and a surprising but obviously false one.

However, Cohen himself admits that “defects other than unclarifiable unclarity can suffice to stigmatize a text as bullshit” (Cohen 2006: 131). He hints at other properties that might flag a statement as bullshit, such as “arguments that are grossly deficient, either in logic or sensitivity to empirical evidence”, or “irretrievably speculative comments.” As such, it might be possible to broaden Cohen’s proposal and improve on it with a more comprehensive output-based account of bullshit.

But where to begin? This is where I must emphasize the second purpose of the examples I presented in introduction: as I see it, they draw our attention to an important feature of bullshit—namely that bullshit is not necessarily meaningless, false, or lacking in evidence or respect for truth, but that it can always be deflated. Indeed, what makes the menu or the flattering resume instances of bullshit is both that they present the truth in an overly flattering light and that this impression will disappear as soon as we begin investigating what they really mean.

The same is true for other examples of bullshit. Lacan’s statement might sound deep and impressive but wondering “what they really mean” will lead us to conclude that they are meaningless. The Althusserian texts described by Cohen and Dennett’s deepities might also give us, at first sight, the impression of encapsulating some profound insight, but asking ourselves “what they really mean” will lead us to conclude that they are either false or trivial. Pseudo-sciences promise to give us some understanding of the world we live in, but this impression is misleading. Thus, the common thread between all these cases is the following: bullshit is something that seems to have much value at first sight, but scratch the surface and you will see that it is empty. To put it simply: bullshit is just hot air.

I propose to formalize this very simple insight in the following way:

What makes a given claim C bullshit is that (i) though C is presented as or appears at first sight as making an interesting contribution to a certain inquiry,⁵ (ii) C would turn out, on closer inspection by a minimally competent inquirer, to make a much less interesting contribution.

Here, I introduce the notion of “minimally competent inquirer”, because we don’t want all claims that ultimately fail to make an interesting contribution to count as bullshit. Imagine that you put forward a bold but plausible scientific hypothesis. A colleague then goes to great lengths to test your hypothesis. After years of effort, he finally succeeds in disproving your theory. Would it be fair for him to conclude that you were “bullshitting?” Probably not. Our intuition is rather that bullshit can be proved worthless via a very basic, elementary inquiry. You just have to scratch the surface to see that it’s empty.

We can thus define a minimally competent inquirer as someone who is not necessarily a specialist but has a basic comprehension of the terms involved in a bullshit claim and (when applicable) of the theories they play a role in. For example, psychologists Evans and colleagues (2020) designed a “receptivity to scientific bullshit scale,” which mixes scientific and pseudoscientific statements such as:

• In all thermal equilibria, if no surface tension is applied nor any refraction imposed upon the object, the capacity of that atomic structure disperses throughout the object.

• In a natural thermodynamic process, the sum of the entropies of the interacting thermodynamic systems increases.

In this case, the first sentence is bullshit, while the second is scientifically valid. If you guessed on your own—congratulations!—I was personally unable to tease them apart. However, the point is that most people who have a basic comprehension of the scientific terms involved in these sentences and of the theory they play a role in would certainly be able to. This is what makes the scientific bullshit statement bullshit and not some very speculative but really interesting hypothesis.⁶

In addition to this knowledge, the minimally competent inquirer should also master the basics of formal and informal logic, and be able to spot the most obvious forms of bad reasoning and arguments. They should also have a certain quantity of basic knowledge – meaning that they should know what most people in their epistemic community know.⁷ In some cases in which the target claim contains no technical term, then the minimally competent inquirer is just a person with this basic knowledge and a decent mastery of logic and reasoning.

The phrase “on closer inspection” should also be clarified. This includes at least two steps. The first (conceptual clarification) is the willingness to clarify what exactly is said, to ask oneself the question: “what does it really mean?” The second (fact-checking) is the willingness to seek easily accessible information to verify whether the claims that are made are (i) non-trivial, and (ii) true or at least based on reasonable evidence. As we saw, some claims, such as Lacan’s claim, do not pass the first step. Others, such as the main tenets of homeopathy, might pass the first, but not the second.

With that in mind, I can thus summarize this new proposal in the following way: bullshit is what seems or purports to make an interesting contribution to a certain inquiry but can be identified as failing to do so under closer inspection by a minimally competent inquirer. But there are many ways in which a claim C can fulfill these conditions. Here are some of them:

• C is meaningless,

• C has meaning but is obviously false,

• C has meaning and is true but is trivial or uninteresting,

• C has meaning, is true, and is not trivial, but there is no way that it can be grounded in evidence,

• C has meaning, is true, is not trivial, is plausibly grounded in sufficient evidence, but is irrelevant to the context of the discussion.

Statements that purport or appear to make an interesting contribution but ultimately turn out to be meaningless are thus only one sort of case of bullshit statements, but are probably the most paradigmatic ones, as closer inspection will ultimately prove them ‘empty’ and ‘void.’ Note that there are several ways in which a statement can be meaningless. Some might contain meaningless terms or pseudo-words, such as “that’s a nice Tnetennba” or “the geodesic biocompensator was tested using a Teslameter 05/40 and the Voltcraft Vc 960” (Milgram 2021). Others might contain real, meaningful words but combine them in a way that fails to make sense, such as “the sum of the angles of a triangle is equal to the color red” or “quadruplicity drinks procrastination.” However, in line with Erwin (1970: 161), I propose that meaningless statements all share one common property: “we cannot understand what it would be like for them to be true”. More precisely, I propose to cash out this intuition in the following terms: a statement is meaningless when it is impossible for competent speakers to conceive a possible world or state of affairs that would make this statement true.⁸

4. Arguments for the Proposal

Now that I have put forward my own account of bullshit, I would like to list a few arguments in its favor.

(1) First, I think this definition allows for the integration of most of the examples of bullshit put forward in the literature. Because it includes meaningless sentences, it accounts for cases of “meaningless” and “unclarifiable” bullshit, such as Lacan’s claim (Cohen 2006). Because it includes claims that are meaningful, but notoriously false, it also includes most cases of pseudoscience, such as homeopathy, astrology, flat earth theories, and other kinds of conspiracy theories (Johnson 2010; Dieguez 2018; Moberger 2020). Because it includes claims that are true but that turn out to be trivial or uninteresting after clarification, it includes “deepities” (Dennett 2009). Because it includes claims that are true and non-trivial, but that are made without proper evidence, it also allows us to include the cases of “irretrievably speculative comment” put forward by Stokke and Fallis (2007). And because it includes claims that are not relevant to the context of discussion, it includes cases of evasive bullshit (Carson 2016).

One question, however is whether my account can accommodate two further types of bullshit identified by Moberger (2020): obscurantist pseudophilosophy (the kind of discourse produced by people in humanities and social sciences when they engage with philosophical issues without awareness of the relevant distinctions and arguments), and scientistic pseudophilosophy (the kind of discourse produced by people with background in natural sciences when they engage with philosophical issues without awareness of the relevant distinctions and arguments). What makes such discourses bullshit according to Moberger is that their author’s ignorance of the relevant philosophical literature shows that they do not seriously engage in the relevant epistemic inquiry (i.e., they are not “conscientious” enough). However, I disagree: Let’s imagine that a scientist ponders over a serious philosophical topic in isolation, being completely ignorant of the relevant philosophical literature, but produces insightful theories and arguments regarding this topic. Should we still label his proposals as “bullshit”? I don’t think so. Rather, what makes obscurantist and scientific pseudophilosophy bullshit is that anyone with a basic knowledge of the relevant philosophical literature can see that such claims are conceptually confused, trivially false, or insufficiently warranted by the available evidence (often because the existence of a large philosophical literature discussing these topics act as a defeater for most of the simplistic views and arguments presented by pseudophilosophers).

(2) Second, I think that my account allows for a satisfactory elucidation of the relationship between lies and bullshit. On the one hand, certain accounts (such as Frankfurt’s account or Stokke and Fallis’ account) draw a sharp distinction between the two: if a claim is a lie, it cannot also be bullshit. On the other hand, most prototypical cases of bullshit in ordinary language are also cases of lies. For example, one famous best-seller about bullshit opens with the following claim:

We live in an era of unprecedented bullshit production (…) Never in history have so many people uttered statements that they know to be untrue. (Penny 2005:6)

Thus, it seems that bullshit in the ordinary sense somewhat intersects with lies. My account explains how some lies can be bullshit, without counting all lies as bullshit: a lie is bullshit when it purports to convey very important and interesting information. But most lies are not of this kind, especially when people who lie do so to keep a low profile. For example, a cheating husband who is asked where he was at noon should opt for a mundane answer such as “I was having lunch with Bob from the Accounting Department”, rather than “I was having lunch with Barack Obama.”

(3) Additionally, my account of bullshit is more in line with the way bullshit is conceptualized and operationalized in other academic disciplines. For example, anthropologist David Graeber (2018) defines “bullshit jobs” in the following way:

a bullshit job is a form of paid employment that is so completely pointless, unnecessary, or pernicious that even the employee cannot justify its existence even though, as part of the conditions of employment, the employee feels obliged to pretend that this is not the case. (9–10)

In other words, a bullshit job is a job that gives the impression of or is touted as having an important impact while, in fact, making no contribution whatsoever. This is in line with my definition of bullshit and allows for a unification of the category of “bullshit” across objects and disciplines.

Moreover, in the past ten years, there has been a growing body of research on the psychology of “pseudo-profound bullshit”. In their seminal paper, which gave birth to this research program, Pennycook and colleagues (2015) characterize bullshit in the following way: “bullshit, in contrast to mere nonsense, is something that implies but does not contain adequate meaning or truth”. As can be seen, this is an output-based rather than a process-based definition of bullshit. But more important is the method Pennycook and colleagues use to produce their list of bullshit statements: they use statements that have been generated at random by non-human programs such as “The New Age Bullshit Generator” or “Wisdom of Chopra”. However, it is hard to explain why such statements (or, more generally, statements generated by large-language models) can be bullshit on the basis of most process-based accounts, for which the indifference of the locutor towards truth is what makes bullshit bullshit.

(4) The case of artificially-generated statements lead us directly to a fourth argument already advanced by Cohen (2006), which is that, to identify bullshit, we rely on the properties of what is produced, rather on inferences about the motives and dispositions of the person who produced it:

Think of attempts to vindicate Heidegger, or Hegel. The way to show that they weren’t bullshitters is not by showing that they cared about the truth, but by showing that what they said, resourcefully construed, makes sense. (130)

Conversely, when we try to demonstrate that something is bullshit, we do so by showing that it is meaningless, obviously false, or much more trivial than it appears at first sight. We do not appeal to its author’s intention or mental states. This is more in line with output-based accounts than process-based accounts.⁹

(5) My account can explain the bullshit metaphor: why is it always bullshit or horseshit, but not flyshit or mouseshit? It’s because the excrement of bulls and horses are impressively huge (when compared to the ones of flies and mice) – but big shit is still shit. They might seem impressive at first sight but are worthless on closer chemical analyses – which is basically what my analysis of bullshit is getting at.¹⁰

(6) Finally, one last argument for my account is simply that it provides a simple explanation for the attractiveness of process-based accounts. Indeed, once we realize that a claim is obviously meaningless, false or trivial, it is hard to accept that the person who made that claim could have done so while being genuinely concerned about the truth. Indeed, that would mean that this person is somewhat stupid, or at least much less clever than we are. For people who are charitable, it is thus difficult not to jump from the mere perception that a claim is bullshit to the conclusion that the utterer did not really care about truth. In other words, I think that defenders of process-based accounts are too charitable.

I’m not.

5. A Process-Based Account of Bullshitting

Still, one might argue that my account of bullshit is incomplete to the extent that it lacks an account of bullshitting. A simple solution would be to define “bullshitting” as “intentionally producing bullshit”. However, I don’t think that would be satisfactory, for I think it is possible to be a bullshitter without producing bullshit statements. This is the case, I think of many pseudo-experts who pass as experts: they often do so by memorizing a few facts or anecdotes that most people with a few years of training in a given field would know, and then repeat these facts and anecdotes in front of a non-expert audience to convey and give the impression that they are indeed experts (Fuhrer et al. 2021). To many experts, this would appear as “bullshitting” the audience— even though nothing that the pseudoexpert says is “bullshit” in a strict sense (the information might be trivial to the expert, but it is not to the general audience the pseudoexpert is trying to seduce).

But maybe a clearer example of “bullshitting without bullshit” might be found in cases of “moral grandstanding” (Tosi & Warmke 2016: 199). Imagine a public figure that takes very harsh (but correct or plausible) moral positions against people who violate a certain moral norm. At some point, it is discovered that this public figure never really cared about this moral norm (or moral norms in general) and that they engaged in moral condemnation only to improve their public image. It would be natural to think that this public figure was “bullshitting” the audience the whole time. However, none of their positions or moral statements need to be bullshit in the sense I defined it.¹¹

A last, but even more striking example of “bullshitting without bullshit”, can be found in advertising. Throughout the literature on bullshit, advertising is considered a prototypical example of bullshit. However, a lot of advertisements do not even make claims (Johnson 2010). Good examples can be found in advertisements for perfumes (such as advertisements by Chanel or L’Oréal), which are often sequences of beautiful but puzzling imagery without any words spoken. These advertisements make no claim—as such they cannot be bullshit according to my account (and most accounts I reviewed). Still, they make us feel like we are being bullshited. What gives?

The answer might lie in the fact that, though they do not make any claim, such advertisements seek to produce a certain effect: they want us to associate positive feelings and ideas to their products. This is in line with several suggestions according to which bullshit is more interested in producing effects than communicating ideas. For example, Neumann (2006: 203) writes that: “Bullshit is a certain kind of speech, intended to distract or obfuscate in a general way, in order to achieve a desired effect – often one that is nonrational and emotional, where emotions become reasons for a course of action”. Similarly, Preti (2006: 21) writes: “what matters to bullshit is that it should make something matter to us”. And, speaking of “clickbait” headlines, Bergstrom and West (2020:23) conclude that: “most successful headlines don’t convey facts, they promise you an emotional experience”.

Thus one could be tempted to define bullshitting in the following way:

X is bullshitting when X engages in a communicative act but is more concerned about the general (affective) impression their act will have on a given target T than about the particular propositions they will get T to endorse.

Such an “impression-based” definition might account for Frankfurt’s intuition that bullshit “is never finely crafted, that in the making of it there is never the meticulously attentive concern with detail” (Frankfurt 2005: 21), while “the liar is inescapably concerned with truth-values” (51). Indeed, while the prototypical liar is concerned with implanting a particular belief in a particular proposition (e.g. that he was out of town when a certain murder was committed), the prototypical bullshitter is more concerned with leaving a general impression (e.g. that he’s competent in his field).

However, despite its advantages, this definition of bullshitting might still be too broad to be adequate. Imagine someone who is composing a CV and a motivation letter to apply to a job but has enough accomplishments and genuine motivation to avoid writing bullshit in the sense I defined it. Still, his main goal when engaging in those actions is to leave a positive impression on potential employers. But it would be strange to conclude that this person is “bullshitting”.

I think we can avoid such strange conclusions by introducing into my definition of “bullshitting” something I left out of my definition of “bullshit”: a certain modicum of dishonesty. Thus, we could say that the bullshitter knows that the impression he wants to convey is not clearly warranted. Taking into account that one can bullshit not only to give positive impressions, but also to minimize bad impressions (for example when Putin announces “a special military operation aimed at demilitarizing and denazifying Ukraine” rather than crudely speaking of “war” and “invasion”), or to give terribly negative impressions (for example of political opponents), I can offer a more appropriate definition of “bullshitting” along these lines:

X is bullshitting when X engages in a communicative act C but is more concerned with the general (affective) impression their act will have on a given target T, than about the particular propositions they will get T to endorse AND X is aware that X has no good reason to think that the impression X is trying to convey is warranted by reasons presented in or hinted at by his communicative act C.

This definition allows both “bullshit without bullshitting” (by reintroducing the idea that bullshitting involves some kind of dishonesty, while bullshit doesn’t) and “bullshitting without bullshit”. Still, it keeps a unity between bullshit and bullshitting by having both revolve around the idea of “something that is more impressive at first sight than it really is or purported to be”.

6. The Problem with Bullshit

So far, I have mainly motivated my account of bullshit on the basis that it allows me to accommodate paradigmatic instances of bullshit that are left aside by other existing accounts. However, one might object that the aim of a philosophical account of bullshit should not be to capture our ordinary linguistic intuitions about what counts as “bullshit”, but to engineer a concept of “bullshit” that is able to make some philosophical work by (i) identifying a certain phenomenon that is worth investigating, (ii) furthering our understanding of this phenomena, and (iii) furthering our understanding of related phenomena. For example, Stokke and Fallis (2017) are interested in bullshit because it allows them to explore the varieties of non-alethic speech beyond lying, while Moberger (2020) tries to put forward a concept of bullshit that would allow him to better understand pseudoscience.

I actually agree with this idea: my inquiry is not an inquiry into ordinary language. I think the goal of a proper philosophical account of bullshit is to provide theoretical insight into phenomena themselves, not on how people use the word “bullshit”. However, I will argue that the account I provided in this paper is better suited to fulfill some of the expectations that we can have about a philosophical account of bullshit.

Indeed, most of the philosophical literature on bullshit touts bullshit as a formidable threat to truth. For example, Frankfurt writes that “bullshit is a greater enemy of the truth than lies are” (2005: 61). Thus, we would like an adequate account of bullshit to explain why bullshit is such a threat, and how it spreads, so we can properly understand and react to it. However, I don’t think that most philosophical accounts of bullshit succeed in this regard. (Though they might succeed with respect to other criteria.)