1. Harris's literature review

The first half of Harris's essay is a literature review, in which he makes a number of concessions regarding the potential reasonableness of conspiracy theories. He concedes that, “many of the supposed faults associated with conspiracy theorising are not faults at all” (Harris Reference Harris, Barker, Crerar and Goetze2019: 235). And so, “[T]he faults standardly associated with conspiracy theorising do not warrant the sort of criticism to which the practice is often subjected” (Harris Reference Harris, Barker, Crerar and Goetze2019: 235). More specifically, Harris considers the idea that conspiracy theorizing is a straightforward epistemic vice, concluding that:

In addition, regarding the argument against conspiracy theories based on their unfalsifiability, Harris concludes: “If there is an epistemic problem with conspiracy theories, and belief in them, it is not that such theories are unfalsifiable” (Harris Reference Harris, Barker, Crerar and Goetze2019: 246). Brian Keeley had previously suggested that “unfalsifiability is only a reasonable criterion in cases where we do not have reason to believe that there are powerful agents seeking to steer our investigation away from the truth of the matter” (Keeley Reference Keeley1999: 121). And this does not apply in the case of conspiracy theories. For, Keeley argues, “we have reason to believe that there exist forces with both motive and capacity to carry out effective disinformation campaigns” (Keeley Reference Keeley1999: 121). Harris actually takes the defense of conspiracy theories on this point even further. He explains, “[R]esilience to falsification is hardly unique to conspiracy theories. Scientific theories in general are resistant to falsification. This point is emphasized by Imre Lakatos, among others … The need to conjoin theories to auxiliary hypotheses to derive testable predictions ensures that scientific theories are not straightforwardly falsifiable” (Harris Reference Harris, Barker, Crerar and Goetze2019: 245). This is refreshing. The purported unfalsifiability of conspiracy theories continues to be used against them even though falsifiability has, in the philosophy of science, been repudiated as a useful criterion for literally decades. Indeed, the theory that unfalsifiability renders theories unwarranted is itself unfalsifiable, and thus is self-refuting.

Harris also addresses the suggestion, made by Steve Clarke, that believing a conspiracy theory is like holding on too long to a degenerating research program. He writes:

Given all this, Harris concludes, “Standard criticisms of conspiracy theorising are misguided” (Harris Reference Harris, Barker, Crerar and Goetze2019: 247).

* * *

At this point, Harris offers three new critiques. He argues that (1) conspiracy theorizing involves a kind of invalid reasoning: probabilistic modus tollens. (2) It involves a particular combination of intellectual virtues and vices which, although the vices are not unique to conspiracy theorists, in combination with the otherwise virtuous intellectual traits of conspiracy theorists, the result is epistemically problematic. And (3) the skepticism conspiracy theorists exhibit toward authoritative sources leaves them with no basis for affirming their own theories. He summarizes his critiques this way:

Below, I examine each of these critiques (and a couple of variants of them) and find them wanting, just as were all previous efforts of this kind. Then I ask whether Harris's project was, despite his explicit rejection of generalism, nevertheless really just another ill-fated attempt at a generalist critique of conspiracy theories.

2. Critique #1: Conspiracy theorizing relies on probabilistic modus tollens

Modus tollens (aka denying the consequent) refers to the following valid inference form: If p then q; not q; therefore not p. Given Harris's examples, it is best to add a couple of extra “nots,” such that the q and not q switch, which does not actually change anything. It will nevertheless help to think of it this way: If p then not q; q; therefore not p. So, for example, suppose it is true that: if Rex is dead then Rex's heart is not beating (if p then not q). Suppose it is also true that: Rex's heart is beating (q). If we can be sure of both of those statements, then we can infer, with certainty, that Rex is not dead (not p). Because if he had been dead, we had agreed in the first premise, his heart would not be beating, and yet it is. That's just Logic 101.

The problem that Harris raises is that this inference is not valid when dealing with probabilities, rather than certainties.Footnote ⁵ The conjunction of “if p then probably not q” and “q” does not entail “probably not p.” Harris gives the following example, “If any given lottery with many participants is fair, it is improbable that any particular entrant will win [if p then probably not q]. However, it would be absurd to conclude, once a winner is named [q], that the lottery was probably unfair [probably not p]” (Harris Reference Harris, Barker, Crerar and Goetze2019: 250). Harris is right: that would not be a valid inference.

However, there are problems with Harris's critique. For one, the example Harris provides is misleading, but I will address that later. First let's address two central problems: (1) Harris's analysis implicitly assumes, wrongly, that if evidence is not conclusive, or at least not very strong on its own, then it doesn't count as evidence at all, and (2) it is not clear that conspiracy theorizing involves the form of reasoning Harris implies. Although conspiracy theorizing may be superficially similar to probabilistic modus tollens, nuances make a difference. And, although an argument form that is not valid cannot establish its conclusion, a form of reasoning similar to it, with the proper nuance applied, may legitimately provide evidence for it.

This can be seen by considering another invalid argument form, affirming the consequent: If P then Q. Q. Therefore P. A standard example is this: If it rained recently, the streets are wet. The streets are wet. Therefore, it rained recently. This is invalid, of course, because it is possible that the streets became wet some other way. Although this is clearly an invalid form, much scientific reasoning seems, superficially, to follow it. A scientist generates a hypothesis, P; Then she reasons that if her hypothesis is true, she should see result Q. She then devises an experiment to produce condition P and to see if Q occurs. Of course, if Q occurs, she does not conclude that P must be true. She recognizes that would be an invalid inference; it is possible that Q just happened to occur for some other unknown reason. However, the experiment is nevertheless evidence that her hypothesis, P, is true. And piling up evidence of this kind does legitimately increase her confidence in her hypothesis. Conspiracy theorizing using something that looks like probabilistic modus tollens is analogous.

Further, the reason the conjunction of “if p then probably not q” and “q” does not entail “probably not p” is that there are at least some instances in which we cannot draw that conclusion. But that does not mean that there are no circumstances under which we can reasonably draw that conclusion. For example:

It may be possible to advance some technical quibbles, but the really important point is just that the conjunction of A and B is legitimate evidence for C. And under some circumstances this kind of evidence is sufficient to reasonably support belief.

To make his point about the invalidity of probabilistic modus tollens relevant to conspiracy theorizing, Harris ties it to what Brian Keeley has termed “errant data,” that is, data that seems unexplained by or inconsistent with the official account of an event (Keeley Reference Keeley1999: 118). Harris writes, “The observation of errant data is, I suggest, analogous to the observation that some lottery entrant has won. The observation of errant data is improbable – perhaps extremely improbable – given the official account, but this alone does not provide reason to abandon the official account” (Harris Reference Harris, Barker, Crerar and Goetze2019: 250).

Now it's time to consider why Harris's lottery example is misleading. In his version of it, it is not surprising that there was a winner – indeed, if the lottery is fair, someone should eventually win. And if there is a winner, it had to be somebody. So, there is really no cause for suspicion. But suppose we change the lottery story just a little, to make it resemble more closely conspiracy theorizing. Instead of some random person winning, let's assume that a close relative of someone involved in selecting the winner won.Footnote ⁶ That may legitimately raise some eyebrows. But suppose it happened again and again. At some point, surely, it is reasonable to conclude that the lottery is probably unfair. And yet Harris seems to suggest that that inference is never reasonable, no matter how many times this situation is repeated. He writes:

What has gone wrong? In the revised lottery example, which better represents conspiracy theorizing, it is not something simply improbable that happened. It is something suspicious that happened. It is something that suggests an alternative theory of events. In the article that first explicitly distinguished generalism from particularism, Buenting and Taylor label this “fortuitous data,” sometimes referring to it colloquially as being “lucky” (or even “fishy”; Buenting and Taylor Reference Buenting and Taylor2010: 576). They explain, “If something is lucky in the way we intend, then merely stating it will suggest an alternative explanation of that fact. … some evidence is just so lucky that it cannot be shrugged off as mere chance; the luckiness of events requires explanation” (Buenting and Taylor Reference Buenting and Taylor2010: 573).Footnote ⁷ One of the examples they offer is the following:

Whatever one thinks about the particulars of this case (for example, whether the passport was found before or after the collapses, and whether it was essentially undamaged), one can imagine circumstances under which such a situation would suggest that al-Sugami was being set up. So, the mere fact that this is “errant data” does not show that it does not represent significant evidence of conspiracy. Its strength as evidence depends on the particulars.

Now let's consider an example provided by Harris: the collapse of Building 7 was reported as having occurred about 26 minutes before it actually occurred, with the building actually visible still standing in the background as the reporter reports that it had collapsed. This is errant data in the sense that it seems highly improbable, if one assumes the official account is true. Nevertheless, it does not follow that the official story is probably false. Harris explains:

It is not at all evident that there are a significant number of 9/11 conspiracy theorists who believe the official account is false on this basis alone. It is more likely they have other reasons for suspicion, and this is regarded as a striking anomaly, suggestive in that context of suspicious foreknowledge. So, the suggestion that conspiracy theorizing involves concluding the official story is probably false “on this basis alone” is a bit of a red herring. Further, whether or not this situation is reasonably regarded as suspicious depends, once again, on the particulars of the situation. Were there good reasons to expect Building 7 would fail, and thus reasonable explanations for how a report of its potential collapse may have innocently been converted to a mistaken report of its actual collapse shortly before it happened? If so, that is the reason this is not strong evidence against the official account. In general, it is just such particulars that will determine whether and to what degree errant data provide evidence against an official account.

Also, we must remember that the important question is not whether a situation such as the premature report of the collapse of Building 7 shows that the official account is probably false; it is whether such an example provides any evidence at all. Is it completely irrelevant? Can it be that, no matter how numerous and outlandish the errant data, we would never be justified in believing, based on this kind of evidence, that the account relative to which such data was errant was wrong? It is not too difficult to recognize that that just can't be right.

Let's briefly consider a couple more real-life examples for clarity and emphasis. A prominent argument cited by members of the 9/11 Truth Movement is that various lines of evidence suggest that there were significant quantities of molten steel (or iron) present at the WTC sites around the time of the collapse and after. When NIST representative John Gross was asked about reports of pools of molten steel at the WTC sites, he seemed to implicitly admit that, if true, this would constitute a problem. He responded, with a clearly dismissive air, suggesting that the reports are not true, “Steel melts at around 2,600° F. I think it's probably pretty difficult to get that kind of temperatures in a fire” (Walter Reference Walter2015: 33).Footnote ⁹ This suggests that if the claims of molten steel or iron held up, they would constitute significant evidence (though still not proof). What is important is whether or not such claims do hold up, not whether they count as (merely) errant data.

Now consider some facts regarding the Robert Kennedy assassination. While the witnesses placed Sirhan's gun about three feet away, the coroner determined that Kennedy was shot from about one to three inches away, toward the back of the head, behind the right ear (Noguchi Reference Noguchi2014). Surely that is evidence of Sirhan Sirhan's innocence, with reference to firing the fatal shot. It is not by itself proof, of course. It is not necessarily even sufficient evidence to make it probable that the official account is wrong. But it is evidence nonetheless. That is, it gives us a reason to increase our assessment of the likelihood that the official account is wrong. Whether or not one is warranted in believing the official account is wrong, or that it is probably wrong, depends on the sum total of evidence – including this kind of evidence.

There are also less directly relevant instances of errant data, which are less telling, but not necessarily without significance. In the RFK case, for example, Sirhan's lawyer, Grant Cooper, was under pending indictment at the time he was (some say incompetently) “defending” Sirhan. Though Cooper was potentially facing jail time for possession of stolen grand jury transcripts, he eventually got off with merely a fine. Are circumstances that suggest Cooper may easily have been pressured to throw his client under the bus, combined with reasons to think that he did just that, epistemically irrelevant? No, even if this could never by itself justify rejecting the official story, its degree of significance depends on the details and further context. However inconvenient, that is what has to be carefully considered.

Let's look at one last example of errant data: statements by doctors and others at Parkland Hospital after the attempt to save President Kennedy's life. Many of these statements seem to suggest that the President was shot from the front (see Palamara Reference Palamara2015). These statements, which are “errant” relative to the official story, do constitute evidence that Oswald was not the lone gunman. They are not, by themselves, conclusive, of course. They do not even show, by themselves, that Oswald was probably not the lone gunman. After all, many reasons have been given to explain away these statements. I'm not trying to adjudicate the strength of this evidence or the plausibility of the responses to it. But I am asserting that such adjudication, taking into account all the relevant evidence, is precisely what is necessary to determine whether or not the lone gunman theory is the best one. Simply making a point about probabilistic modus tollens just isn't going to do it. In fact, it does not help at all. Indeed, it is positively misleading. For it wrongly suggests that data that seems to problematize an official account cannot be relevant.

3. Critique #2a: The combination of intellectual effort and confirmation bias is problematic

Next Harris considers whether conspiracy theorizing involves some kind of epistemic vice. Ultimately, he concludes that it does. But he suggests that it is not a straightforward vice, such as has been alleged by others. For example, Harris rejects the view that conspiracy theorists, as a group, are intellectually lazy. He remarks, “Conspiracy theorists exhibit a sort of intellectual diligence, a motivation to uncover the truth, that seems downright praiseworthy” (Harris Reference Harris, Barker, Crerar and Goetze2019: 252). That doesn't mean they have no straightforward intellectual problems; Harris sees that they do, and that the rest of us do too. He concludes, “[T]o the extent that proponents of conspiracy theories and official accounts alike are both prone to confirmation bias and closed-mindedness, appeals to such things cannot fully ground the relevant sort of criticism of conspiracy theorising” (Harris Reference Harris, Barker, Crerar and Goetze2019: 253). After all, he writes, “[T]he susceptibility of the conspiracy theorists to these intellectual shortcomings is not, I would suggest, a dimension in which conspiracy theorists are clearly distinguished from their counterparts” (Harris Reference Harris, Barker, Crerar and Goetze2019: 253).

However, Harris believes he has found a more complex way in which conspiracy theorizing is problematic. It is the particular combination of vices and virtues that characterize conspiracy theorists, he suggests, that results in problematic thinking. “In short, the fact that conspiracy theorists’ enthusiasm for the pursuit of truth is not matched by a correspondingly heightened sensitivity to their own cognitive biases and potential for error exposes conspiracy theorists to unique epistemic criticism” (Harris Reference Harris, Barker, Crerar and Goetze2019: 253–4).

Harris does not argue that conspiracy theorizing generally involves paying less attention to potential biases than does, say, official account activism or conspiracy theory debunking. Rather, Harris suggests that the problem of confirmation bias, though shared by conventionalists, is less problematic for conventionalists precisely because they tend, overall, to be more intellectually lazy, at least with respect to investigating issues related to conspiracy theories. The reason being lax with respect to one's attention to one's own biases is more problematic for the conspiracy theorist than the conventionalist is, Harris explains, “precisely the fact that conspiracy theorising typically involves a greater degree of intellectual activity than that involved in acceptance of an official account” (Harris Reference Harris, Barker, Crerar and Goetze2019: 253) and “greater intellectual effort in general has the potential to go awry if it is not matched by greater intellectual caution” (Harris Reference Harris, Barker, Crerar and Goetze2019: 254).Footnote ¹⁰ However, some conventionalists are also intellectually active, such as debunkers, and even journalists. And, Harris admits that conspiracy theorists may not be more biased than conventionalists.Footnote ¹¹ So, this argument should also undermine intellectually active conventionalists, unless Harris means to suggest that we should make our judgement based on the average level of intellectual activity of different groups. But that makes little sense. Intellectually active conventionalists would be exonerated because of the large number of passive conventionalists, even though that is irrelevant. Further, perhaps the biggest problem is that intellectual passivity is not generally preferable to intellectual activity, at least so long as one strives to be careful and attentive to one's biases. And, again, Harris has conceded that conspiracy theorists are not less attentive to their biases than others.

The proper conclusion one should draw from this is not that there is something wrong with conspiracy theorizing, but only that one should be careful when one exerts intellectual effort to determine the truth of contested claims, whichever side one initially tends to favor. Certainly, we should not conclude that the best thing to do, given the pervasive propensity for confirmation bias, is to remain as intellectually disengaged as possible. Rather, it seems, when people engage with evidence relevant to conspiracy theories, both those who are inclined toward conventionalist thinking and those inclined more toward conspiracy theorizing should do so with as much intellectual fairness and caution as they can muster.

4. Critique #2b: The view promoted in the mainstream is unbiased

There is one other element in the above argument: the suggestion that when conspiracy theorists are intellectually active, they tend to consult non-mainstream sources. However, this does not really distinguish them from intellectually active conventionalists. And it is not clear that those who come to believe conspiracy theories are less attentive to their own biases than journalists who are compelled to be active, or at least ought to be active, in their inquiry regarding the issues in question. Harris is nevertheless optimistic about the reliability of the mainstream media.Footnote ¹² And so, the assumption is, it is better to adhere passively to the mainstream perspective, than to actively consult non-mainstream sources.

Is such optimism warranted when it comes to contested interpretations of events? In general, the mainstream media's reliability regarding contested accounts may be difficult to adjudicate. Are we to rely on the mainstream media itself for determining its own reliability? That seems methodologically questionable. And I'm not sure what the alternative is if we are not to include the consideration of non-mainstream perspectives, which Harris seems to imply we should generally steer clear of. In any case, there are mainstream journalists, and former mainstream journalists, who argue from experience that American mainstream journalists who challenge sensitive official narratives run into a “buzzsaw” (see Borjesson Reference Borjesson2004).Footnote ¹³

That shouldn't be too surprising. After all, the vast majority of the mainstream media is controlled by a few giant corporations that have interests in selling certain products – including pharmaceuticals and instruments of war. Nevertheless, as Harris points out, they have at least some interest in getting the facts right. How “powerful” that interest is may be questionable. It would seem that they would indeed want to avoid blatantly misreporting “simple questions of fact.” But the media's ability to ignore facts, underreport them, or report them with a subtle spin is significant. In addition to viewers and readers, they have advertisers to please. And “access journalists” don't want to step on the toes of those who provide them with access. Sometimes news outlets are pressured by government agencies to hold back on stories, such as the warrantless wiretapping program, which was not revealed until after President Bush's reelection. In short, the mainstream media has conflicting interests and is susceptible to influence by external pressures. The situation is not nearly so simple and unproblematic as Harris implies.

But it gets worse. When it comes to the allegations often involved in conspiracy theories, the alleged truth can be downright toxic, in the sense described by Lee Basham: “‘Toxicity’ is the likelihood that some conspiratorial scenarios, even if well evidenced, are too ‘toxic’ for our usual institutions of public information to disseminate to the public, or even investigate. Cover-up by intentional neglect, not descending control, is the easily predictable consequence” (Basham Reference Basham2018a: 271). According to Basham, “The problem of toxic truths can threaten to invert our apparently warranted expectations of public institutions of information, our public trust” (Basham Reference Basham2018a: 275–6, 286–7). The idea is this: Up to a certain point, the more momentous a story, the more attention will be paid to getting it right. However, for certain kinds of issues, there is a threshold point after which the situation is suddenly perceived to be too momentous.Footnote ¹⁴ At this point it has become toxic, or radioactive.Footnote ¹⁵ It becomes like the so-called “third rail”; everyone knows not to touch it. While Harris worries that when people venture beyond mainstream sources they risk reading biased sources, when it comes to accusations that would be significantly disruptive if acknowledged to be true, one also has reason to expect the mainstream to act in a biased way so as to support the status quo. And though one can expect it to be heavily scrutinized indeed, the harshest scrutiny may come from those concerned that the story not deviate too far from the official line or be too offensive to patriotic sensibilities.Footnote ¹⁶ Putting the point more mildly, political scientist Joseph Uscinski writes:

Although, in the case of Watergate, the investigation continued and the conspiracy was exposed, if things had been just a little different it may well have ended much more ambiguously,Footnote ¹⁷ as in some ways the Iran-Contra Affair did.Footnote ¹⁸ There is a survival bias issue here, as well. We can only definitively cite cases that did end up being pursued. But the tenuousness of the process of these and other casesFootnote ¹⁹ should give us pause.

But let's consider Harris's argument in support for his optimistic view. He reasons as follows:

Based on this, Harris seems to suggest we should just defer to the account promoted by mainstream sources, suggesting that “The collection of additional evidence may leave one worse off, epistemically, when that evidence is drawn from a biased pool.”

There are many problems with this.

First, nearly everything is “subjected to at least some degree of truth-oriented filtering.” So, without some quantification, applying this to the mainstream media is pretty meaningless. The degree to which the mainstream does a better job than other sources, after all, presumably varies. Some of these “other sources” may be published in respectable venues. Peter Dale Scott's The Road to 9/11, for example, is published by the University of California Press.Footnote ²⁰

Second, the question is not whether the media generally gets things right, it is whether they are reliable when it comes to certain kinds of controversial issues, such as war provocations, or political assassinations.Footnote ²¹ The US mainstream media's promotion of the Iraq war, for example, was seriously flawed. And it is not the case that the mainstream media quickly corrected all their errors. Even today there is a tendency to whitewash the relevant history by framing the Iraq war as an honest mistake, despite strong reasons to think it was not.Footnote ²² Although it is commonly reported that everybody thought Iraq had WMD, there was never strong evidence supporting this, and plenty of people knew it.Footnote ²³

Third, there is at least some tension between the idea that the mainstream media can be trusted because it is heavily scrutinized, and the idea that, because it can be trusted, it need not be scrutinized. Or that, at least, one should not expose oneself to information outside the mainstream media, where one is likely to encounter ideas that challenge the mainstream view.

Fourth, the fact that the mainstream media has incentives against misreporting simple matters of fact does not make it an unbiased pool of information, as Harris implies.Footnote ²⁴ FOX and MSNBC, for example, frequently support contrasting, and seemingly biased narratives. One study even suggests that exposure to Fox News and MSNBC, which are described as “partisan sources,” “has a negative impact on people's current events knowledge.” Most people would do better to watch the comedic Daily Show.Footnote ²⁵ Also, the perspectives promoted in the mainstream media of other countries differ from those promoted in the US media,Footnote ²⁶ despite the fact that they could appeal to the same “powerful incentives” to get things right. Rather than developing one's views based on a single mainstream source, it is probably wiser to base it on multiple mainstream sources with different perspectives. But if that is true, might not expanding one's net a bit further, if done carefully, be wiser still? Harris has not given sufficiently strong reasons to confidently answer “no.”

Fifth, the idea that the mainstream media gets “simple questions of fact” right, and that errors are often corrected (though without fanfare) is not so much wrong as it is a non sequitur. An “official account,” after all, is generally a complex theory, not “a simple question of fact.” Although the media does sometimes significantly misreport facts relevant to conspiracy theories,Footnote ²⁷ perhaps the greater worry, leading people to seek additional information, involves what the mainstream leaves out or fails to emphasize.Footnote ²⁸ The media may faithfully report selective matters of fact that support a particular theory, while largely ignoring so-called errant data. For example, many in the 9/11 Truth Movement believe the collapse of Building 7 is particularly significant, and they often point out that, though shown on TV as it happened, it was seldom shown again for many years.

Finally, there is also the reality of spin. When purported “errant data” is addressed by the media, it is not always addressed in a fair and competent way. For example, the National Geographic documentary, “9/11 Science and Conspiracy” employed “experts” to perform an experiment to answer the question: “Can thermite of any kind burn through steel beams?” They succeeded in showing that even large quantities of the weakest form of thermite cannot cut through a steel beam when uncreatively deployed. The clear intent is to leave the impression that thermite “of any kind” cannot have been used in the destruction of the WTC towers. But the underlying reasoning is patently invalid. And the implied conclusion that thermite cannot cut through steel beams appears to be falsified by more creative attempts using relatively small amounts of stronger forms of thermite (Cole Reference Cole2013).Footnote ²⁹

As an example of the mainstream media apparently getting even basic matters of fact wrong, consider NOVA's “Cold Case: JFK.” The program makes much of the supposed three-centimeter entrance wound on Governor Connally's back. This is used to support the single bullet theory by suggesting that the bullet yawed after passing through Kennedy's neck. One has to go outside the mainstream media to find a critique, by Gary Aguilar and Cyril Wecht,Footnote ³⁰ that points out that the House Select Committee on Assassinations report indicated authoritatively:

Aguilar and Wecht also point out that this is corroborated by the size of the holes through Connally's jacket and shirt, 1.5 cm and 1.3 cm respectively. So, it seems that much of NOVA's analysis is based on a false premise. I'm not aware of any correction or retraction on the part of NOVA or PBS regarding this matter. If there was one, few would know. Aguilar suggests that this case exemplifies the fact that “junk science continues to snooker the mainstream media” (2016). In fairness, it may be that a convincing response to Aguilar and Cyril is possible, and that NOVA and its experts are right after all.Footnote ³² But the question at hand is this: Are we really justified in just assuming this to be the case simply because mainstream media sources have an incentive not to present simple matters of fact incorrectly?

Let's briefly consider one last example. Glenn Greenwald, a Pulitzer Prize-winning journalist who played a significant role in bringing to light Snowden's NSA spying revelations, argues that the mainstream media (1) got the RussiagateFootnote ³³ story wrong, (2) has not owned up to it, (3) will face no consequences for irresponsible and misleading reporting, and (4) all this has “destroyed the credibility of the U.S. media.”Footnote ³⁴ It is beyond the scope of this article to establish that Greenwald's perspective is correct. But it is not obvious that he is wrong. How should we decide? According to Harris's reasoning, we don't need to even try to weigh the evidence. Greenwald is most probably wrong, and we should leave it at that, because what he is asserting is not compatible with a simple heuristic: the mainstream media has an incentive to get things right, and to quickly correct themselves when they screw up. But that is just not adequate.

The bottom line is this: Harris has not shown that consulting non-mainstream sources is problematic, or that we are best advised to uncritically trust views promoted in our mainstream media. A conspiracy theorist may have specific reasons, as well as general reasons, to think that mainstream sources are biased on a particular issue. They may have spot-checked facts and found at least some of the mainstream media reporting to be skewed. In some cases, their reasoning might be circular, and criticizable on that count. But this need not always, or even typically, be the case. In any particular case, we would have to scrutinize the details. Likewise, the quality of reporting can only be determined by looking at the specifics of particular cases. We cannot conclude, based on the combination of virtues and vices suggested by Harris, and the assumed reliability of the mainstream media perspective, that conspiracy theorizing is problematic. And we cannot conclude, without consideration of the evidence in question, that there is something criticizable about any particular instance of conspiracy theorizing. And so, as far as providing guidance for thinking about proposed conspiracy theories, Harris's analysis of epistemic virtues and vices and its relation to actively seeking information from non-mainstream sources is unhelpful, and potentially misleading.

5. Critique #3a: Conspiracy theorists have no basis for believing their information sources

Harris's third critique shifts the focus from how conspiracy theorists can disbelieve conventional accounts to how they can affirm alternative ones. Conspiracy theorists get their information from a variety of sources, including both mainstream media sources and alternative media sources. Harris is particularly concerned with how they can rely on the latter, but I will address both, starting with the former.

Few conspiracy theorists are radical sceptics or believe that the mainstream media is entirely filled with liars covering up for conspirators. Rather, they tend to believe (or, in any case, only need to believe) that there are powerful, or powerfully connected, interested parties that can influence information sources when deemed sufficiently important. For example, conspiracy theorists might suspect that, for fear of losing sponsors, news agencies would tend not to dig into issues involving pharmaceutical companies that might reveal something scandalous. But if mainstream news agencies do engage in reporting that generates such a scandal, there is little reason to think it is made up or exaggerated.Footnote ³⁵ After all, statements against one's interest are generally (and legally) regarded as more reliable than self-serving statements.

Suppose a defendant testifies that he was indeed at the scene of a murder but claims he didn't commit it. A juror may not believe the defendant is trustworthy, and reasonably doubt his claim of innocence, and yet nevertheless believe that his testimony is truthful regarding his presence at the scene. A conspiracy theorist may interpret media reports in an analogous way. They may believe that the mainstream media is biased toward certain official accounts (especially if the alternative is toxic), consistently trying to spin things to support the official view. And yet, when the media reports something that supports a counter-narrative, conspiracy theorists may (depending on the circumstances) legitimately regard that as relatively reliable.Footnote ³⁶

The model that governs the thinking of many, perhaps most, conspiracy theorists seems to be this: while some journalists may actually be intelligence assets, most are just trying to do their job responsibly. The latter type's tendency to support, or at least not challenge, certain official stories can be explained as a combination of bias and some degree of career-motivated self-censorship. After all, while journalists get rewarded for breaking most kinds of stories, they generally know better than to support toxic ideas, and so it is best for them to not even believe those. Thus, the conspiracy theorist may have good reasons to be skeptical of the mainstream reporting when they offer a certain take on certain issues, and yet regard them as more reliable when they provide evidence that cuts the other way.

Harris, however, seems more concerned about the level of trust he believes conspiracy theorists place in alternative media sources. In one formulation of this critique, Harris suggests, what is objectionable about conspiracy theorizing is that the skepticism involved is “often attended by, and even motivated in part by, a dogmatic acceptance of certain sources of information as reliable” (Harris Reference Harris, Barker, Crerar and Goetze2019: 256).Footnote ³⁷ This criticism, however, is only applicable to conspiracy theorists who actually do have a dogmatic acceptance of dubious sources of information, such as conspiracy theorists who uncritically accept everything said by Alex Jones. That leaves a lot of conspiracy theorists unscathed. In any case, one cannot legitimately criticize the activity of conspiracy theorizing itself by pointing to qualities that only apply to some conspiracy theorists.

However, in another formulation of this criticism, Harris casts a wider net. He writes, “[G]iven the level of scepticism that conspiracy theorists must adopt toward certain sources of information, it is unclear how a conspiracy theorist can maintain trust in the alternative sources needed to derive warrant for their preferred conspiratorial explanation” (Harris Reference Harris, Barker, Crerar and Goetze2019: 255). Here, it seems, the trust that conspiracy theorists have in alternative sources need not be dogmatic to be criticizable. But why would it be that some degree of trust in some alternative sources could not be warranted? As mentioned above, the quality of alternative sources varies. Further, as alluded to above, it is unclear what the “level of scepticism that conspiracy theorists must adopt toward certain sources of information” really is. Presumably most conspiracy theorists do not take a radically skeptical view of the mainstream media, as Harris's reference to Descartes implies they do. So, it is not at all clear that the conspiracy theorist's level of skepticism about mainstream media creates a serious “problem for deriving positive warrant for conspiracy explanations.” For one thing, it is often possible to check these secondary sources with primary sources. For example, much of the work produced by the JFK research community is well-sourced to official documents, including the 26 volumes of documents produced by the Warren Commission.

6. Critique #3b: There is no basis for choosing between alternative conspiratorial accounts

Harris considers “a second problem for deriving positive warrant for conspiracy explanations” (Harris Reference Harris, Barker, Crerar and Goetze2019: 256). Namely, “Because any number of conspiratorial explanations can be constructed, all of which account for the data equally well, the conspiracy theorist often lacks sufficient warrant for belief in any particular conspiratorial explanation” (Harris Reference Harris, Barker, Crerar and Goetze2019: 256). Harris rephrases it, with a bit more elaboration, this way:

While, strictly speaking, it is true that “any number of conspiratorial explanations will fit the data” it does not follow, and indeed is unlikely, that they will all be “equally supported,” or that “the conspiracy theorist has no basis for determining that some particular group of conspirators, as opposed to some other group, is responsible for the event in question.” Fitting the data, which is always going to be a matter of degree, is not the only relevant consideration. The underdetermination of theory by data is also an issue in science, in which one is often presented with competing theories that (at least largely) fit the data. Broadly speaking, prior probabilityFootnote ³⁸ considerations and explanatory power both have to be considered, and “fitting the data” is only one aspect of explanatory power. It is one thing to “fit” (not contradict), it is another to explain, and still another to predict.

In order to fit the data, one theory may posit that a key individual (perhaps one facing indictment) was pressured to act in a subtle but significantly corrupt way, while another theory may posit a large number of people with no apparent motives, or even disincentives, acting in a series of coordinated and outrageously corrupt ways. Clearly, the fact that they both “fit the data” doesn't imply that there is no basis for finding one theory to be better than the other.

In general, it seems true that the potential for variety in conspiratorial explanations is nearly limitless, but that does not mean that all account for the data equally well. It is true that one can always expand the scope of the conspiracy to explain a phenomenon, but that is done at an epistemic cost. An explanation that relies on a large number of unlikely conspirators to account for some data does not account for it as well as an alternative conspiratorial explanation which is more modest in scope and implicates only individuals with plausible motives. For example, one conspiratorial explanation of the RFK assassination implicates Thane Eugene Cesar, who was standing immediately behind Kennedy's right side, hated the Kennedys, and falsely claimed to have sold his .22 before the assassination.Footnote ³⁹ One could also construct an explanation involving interdimensional aliens. Both explanations can be made to account for the evidence, but that does not mean they do so equally well. Aliens with supernatural powers can nearly always be made to provide as much explanatory power as any more mundane explanation. But they do so at a severe cost in prior probability. Whether or not the view that Cesar shot Kennedy is true or warranted, it is certainly better than any view that implicates interdimensional aliens, even though the latter can also be made to “fit the data.” It is also better than an account that “fits the data” only by suggesting that all the eyewitnesses are lying because they are in on it.

It is the combination of prior probability and explanatory power that determines the best of competing explanations, taking into consideration all of the relevant particulars. Harris has not provided a sound reason to think that, when these considerations are weighed, alternative theories will not vary in their degree of warrant.

7. Is Harris's critique a generalist one?

Harris suggests that he is not putting forward a generalist argument. For he explicitly recognizes that “A generalist view, according to which conspiracy theorising is invariably irrational, is plainly unfounded. … [T]here are plainly instances of rational belief in conspiracy theories and rational conspiracy theorising” (Harris Reference Harris, Barker, Crerar and Goetze2019: 240). But if his points are not generalist ones – if they don't undermine conspiracy theorizing as a legitimate activity based on generalizations about it – then what force do they have? Although, some of his points have no force at all, or so I have argued, some of them suggest this: conspiracy theorizing should be carried out with a certain kind of care, and when this care is lacking, that is criticizable. Harris himself writes, “I do not mean to suggest that all conspiracy theorists commit the sort of errors described in the preceding sections. However, there are epistemic grounds on which to criticize those that do” (Harris Reference Harris, Barker, Crerar and Goetze2019: 257). Harris, then, has not, even by his own lights, shown that conspiracy theorizing is inherently problematic, but only that it can be done badly – a fact that is only too obvious. The proper implication, even if his analysis of the problems of conspiracy theorizing held up, would be that one ought to be cautious about it. But that doesn't tell us much about conspiracy theorizing in particular. For one ought to be cautious about one's belief formation in general, including accepting official accounts, especially if the alternative would likely be toxic. So, contrary to the rhetorical import of his opening paragraph, it is not merely “standard criticisms of conspiracy theorising” that “fail to demonstrate that the practice is invariably irrational” (Harris Reference Harris, Barker, Crerar and Goetze2019: 235). His own criticisms, he ultimately admits, wouldn't demonstrate that either, even if they had held up.

So, why then is a “relatively charitable attitude” toward conspiracy theories not thought to be appropriate? After all, if conspiracy theorizing is not itself inherently problematic, and if conspiracy theories might be right, and important, why should such theories be given a less than charitable hearing? Harris seems to suggest that the answer has to do with how pervasive the problems he has proposed are. The “epistemic errors” that he describes, he maintains, are “heavily implicated in conspiracy theorising” (Harris Reference Harris, Barker, Crerar and Goetze2019: 257). And so, Harris argues that “[The particularist] position, according to which whether conspiracy theorising is irrational must strictly be evaluated on a case-by-case basis, fails to recognize the extent to which conspiracy theorizing may involve problematic reasoning patterns” (Harris Reference Harris, Barker, Crerar and Goetze2019: 240). However, “the extent to which conspiracy theorizing may involve problematic reasoning patterns” remains uncertain and speculative. Harris's critique of probabilistic modus tollens turns out to be inapplicable. In addition, it is not clear that conspiracy theorizing involves less attention to one's biases than conspiracy debunking or other intellectually active efforts of conventionalists. And it is not clear that intellectual sloth, or blind acceptance of official stories, is preferable. Further, it is not clear that conspiracy theorizing cannot be grounded in judicious assessments of the credibility of relevant information sources.

But perhaps there are other problems with conspiracy theorizing. Indeed, Harris suggests that there are. He explains, “There is good reason to think that there are traits or reasoning strategies characteristic of conspiracy theorising. Belief in a given conspiracy theory strongly predicts belief in other conspiracy theories, even in cases where the conspiracy theories are incompatible” (Harris Reference Harris, Barker, Crerar and Goetze2019: 240; citing Goertzel Reference Goertzel1994 and Wood et al. Reference Wood, Douglas and Sutton2012). Unfortunately, there seem to be problems with this too. Social scientists who study conspiracy theorists have been slow to recognize that believing a particular conspiracy theory to be true gives one a reason to increase one's assessment of the prior probability of analogous conspiracy theories, and thus legitimately increases one's likelihood of believing other conspiracy theories. The type of reasoning seemingly employed here – the argument to best explanation – is not only legitimate, it is pervasive, and thus does not suggest there is a noteworthy reasoning strategy characteristic of conspiracy theorizing. Further, one of the two studies Harris cites (Wood et al. Reference Wood, Douglas and Sutton2012) misleadingly makes it appear as if conspiracy theorists are irrational, even though their evidence does not warrant that conclusion.Footnote ⁴⁰ Wood et al. drew a conclusion about beliefs based on data about how “plausible, convincing, worth considering, and coherent” (Wood et al. Reference Wood, Douglas and Sutton2012: 771) participants thought various claims were. Essentially, Wood et al. conflated believing with giving some degree of credence. It is not irrational, or in any way epistemically problematic, to find two contradictory theories relatively more plausible, convincing, worth considering, or coherent than some other people did.Footnote ⁴¹

Still, while some of the relevant social science research may be flawed, other research suggesting that conspiracy theorists have a tendency to certain types of poor reasoning that is measurably greater than average may survive scrutiny. So, let's assume, for the sake of argument, that there is some truth in the notion that conspiracy theorizing often involves greater than average problematic reasoning patterns. What are the implications? Harris reasons as follows: “If there are problematic traits or reasoning strategies characteristic of conspiracy theorising, there may be prima facie grounds for scepticism about the epistemic merits of conspiracy theorising even if certain instances of conspiracy theorising are epistemically unimpeachable” (Harris Reference Harris, Barker, Crerar and Goetze2019: 240).

Even if there are measurable differences in the average quality of reasoning between conspiracy theorizing and coming to accept official accounts, that would not justify treating any particular case of conspiracy theorizing uncharitably or giving any particular conspiracy theory anything less than a fair hearing. (A fair hearing could still be a short one if the theory had obvious serious problems – such as implicating lizard-people.) For one thing, quality of thinking and degree of bias varies considerably from individual to individual, and from context to context. The existence and extent of poor conspiracy theorizing does not reduce the quality of better conspiracy theorizing; the existence of Alex Jones does not give us a reason to dismiss Peter Dale Scott, Nefeez Ahmed, or Jim DiEugenio.Footnote ⁴² The strength of the arguments provided by more careful and reflective conspiracy theorists does not depend on whether the number of sloppy and unreflective conspiracy theorists goes up or down. Just as we should not judge the arguments of St. Thomas Aquinas by reference to the thinking patterns of average Christians, we should not judge particular conspiracy theories by reference to the characteristics of average conspiracy theorizing.