Jordan Peterson is a respected professor of psychology at the University of Toronto. According to Google Scholar, his scholarship has been cited over 12,000 times — and more than 5,000 of these citations occurred before his emergence as a major public figure.

Peterson also plays an increasingly public role as an advocate for free speech, traditional values and personal responsibility. David Brooks, for example, terms Peterson “the most influential public intellectual in the Western world right now.” Between his best-selling book, his millions of followers on YouTubeTwitter and Reddit, as well as endless press ranging from fawning, to substantive, to critical, to muckracking, Peterson has attracted an unusual amount of attention relative to his academic colleagues — and controversy.

With his focus on corruption and ideological orthodoxy in the academy, especially his resistance to imposed speech on gender norms, many critics argue that he has attracted the interests of a male-dominated and white-supremacist following on social media sites such as Gab and 4chan. Others contend that Peterson serves as something of a ‘gateway drug‘ to more hostile alt-right content. These questions, whether Peterson attracts a disproportionately radical following (a portion of whom may possess misogynist and/or white supremacist sympathies) and whether consumption of his content serves as a gateway drug to more alt-right content, lend themselves to empirical investigation. This report represents an empirical investigation of these questions.

First, we will use machine learning to evaluate whether alt-right-associated Web communities promote Peterson, and we’ll describe how they feature him there. Next, we will perform trend analyses to examine whether new participants on YouTube encounter Peterson’s content as a ‘gateway drug’ or ‘red pill’ to alt-right material. We seek to be civil, curious and constructive about what we find, and we hope this piece can serve as a model for how to engage potentially charged topics with a scholarly and evidence-based approach.

How do alt-right-associated Web communities depict Jordan Peterson’s work?

We first sought to determine how Web communities friendlier to alt-right perspectives depict Jordan Peterson. Peterson has, on occasion made his outreach to these communities very explicit, explaining that his interest lies in “rescuing young men” lost to white supremacist and misogynist ideologies and insisting that his “angry young white men followers are a hell of a lot less angry and a hell of a lot less white than they would have been if they wouldn’t have been following me.”

Comments like these have provoked great outrage on white nationalist websites like The Daily Stormer. But might they be continuing to utilize his content despite frustration with these statements? Are Peterson’s words and arguments being coopted by these noxious actors nonetheless?

To find out, we turned to Gab — an alt-right associated Twitter-clone boasting nearly a million users and resident community for the Pittsburgh Tree -of-Life shooter in 2018. In a previous analysis looking at word associations with ‘White Genocide’ on Gab, Jordan Peterson emerged as a significant focal point in the discussions on this topic. He was the only academic or intellectual whose name emerged as a significant node in this network of associations.

We wanted more insight into how people were evoking Jordan Peterson on this forum. We ran a word/context association analysis using word2vec on the “#JordanPeterson” hashtag on 35 million posts in the community, made from August 2016 to January 2018. Word2vec is a machine learning tool which generates statistical embeddings from relationships of words in a given corpus. It models the likelihood that words share a given context through weighted measures of “cosine similarity”, their alignment and proximity in semantic space. For example, if the tool learned “the cat in the hat” as corpus, it would assign a high cosine similarity score to the words “cat” and “hat” since they tend to co-occur.

Our analysis generated a “two-hop-ego network” of the “#JordanPeterson” hashtag, allowing us to rapidly identify the terms and concepts most closely associated with the tag. Our analysis shows (see Figure 1) that the hashtag commonly appears in the contexts of misogynist tropes (purple), including “feminazi” and the “feminismiscancer” hashtag. It also associates closely with white nationalist tropes (yellow) including anti-immigrant rejoinders and the ‘white genocide‘ conspiracy.

While Peterson himself does not espouse these views, this analysis suggests that his material is digestible in that light and is distributed as such by a portion of people sympathetic to those ideas on the Web. Rather than “rescuing” people from misogyny and white supremacy in these communities, Jordan Peterson seems to be held up repeatedly by such communities in precisely their most misogynistic and white supremacist contexts.

Word associations with #JordanPeterson on Gab

Looking at Gab can provide some insight into how people who are already sympathetic to the alt-right seem to be evoking Peterson. This cannot, however, tell us much about those who do not demonstrate a history of engaging with alt-right material: do they become more likely to subsequently engage with that material after consuming Jordan Peterson content? We selected YouTube as our experimental community to explore this question because Peterson’s channel boasts over 2.3 million followers, making him the most influential academic on the entire site.

Does engaging with Jordan Peterson’s content serve as a vector to radicalization?

Recent evidence provides reason to suspect that engagement with Jordan Peterson, and other members of the “Intellectual Dark Web” (IDW) on YouTube actually does predict engagement with more radical content. Far from driving their members into more reasoned conversations, which is what Peterson and others in the IDW explicitly aspire to do, the data from the data from these studies suggest that people who engage with other viewers of IDW content tend to be drawn, over time, to engage increasingly with alt-right content on YouTube.

In order to gauge these claims, we ran a trend analysis (with Manoel Ribeiro) analyzing Peterson’s YouTube Channel. This analysis examined whether new participants (i.e., users who had no previous comment history and who had commented, for the first time on Peterson’s channel) would engage with more or less alt-right content over time compared to controls who did not engage with Peterson’s YouTube Channel. The results indicate that participants in Peterson’s comments sections on YouTube migrate twice as quickly to alt-right YouTube channels over time relative to controls (see Figure 2). In other words, it appears that participating in Peterson’s YouTube channel predicts increasing flirtation with alt-right content.

Discussion

“Do I believe that there is any connection between leftist identity politics of past decades and the emergence of the white identity politics of the alt-right? Yes, I think there’s a direct causal connection. It only stands to reason that if identity politics is going to be the acceptable mode of communication, that there are people who are on the white side of the equation who say, ‘we’re going to play exactly the same game.’ Clearly, obviously, and horribly — because I’m no fan of identity politics. But what’s sauce for the goose is sauce for the gander. But I would recommend that people don’t do it, because you know the radical leftists and their damn identity politics are unbelievably pathological — and if you decide to fight that by playing the same game,  you think ‘well I’ll play the same game and then I’ll win’ – no you won’t, because by playing the game you lose. That’s the thing about your political opponents: you don’t play their damn game. You play a different game.”

Jordan Peterson

While not conclusive, our findings conjoin two lines of evidence that suggest exposure to Jordan Peterson’s content may (inadvertently) serve as a vector for the spread of alt-right ideas. Alt-right-associated communities tend to associate Peterson’s material in the context of their most heinous misogynist and white supremacist ideas, and engagement with Peterson’s YouTube material predicts higher (rather than lessened) engagement with alt-right content on YouTube. In order to be more concrete, future work should seek to examine the nature of the comments in these analyses — for instance, to quantify the extent to which Peterson’s audience may be engaging critically with the alt-right content they grow more likely to consume, or whether misogynistic or racialized language increases for Peterson’s own YouTube audience over time.

Critically, these analyses do not support the notion that Peterson himself is an extremist or Nazi-sympathizer, and none of us believe that such accusations are credible. We also think that it is highly unlikely that Peterson himself knows about these trends. Indeed, we were ourselves surprised by the findings. Thus, these analyses should not be taken as an attack on Peterson’s character or motives.

Nor do these findings imply that Peterson needs to abandon his message, or steer away from controversial topics, in order to avoid providing ammunition for the alt-right. That is, the solution here is not for Jordan Peterson to self-censor, or for others to attempt to censor him. Instead, we would encourage Peterson et al. to consider ways they may be able to make the same points, just as forcefully, while avoiding a particular set of tropes.

For instance, Peterson wants us to remember the horrors of the communist regimes of Stalin and Mao in order to prevent us from repeating said horrors. He worries that many popular strains of leftist ideology predispose adherents, whether they recognize it or not, towards forcibly imposing their will on others via the state, suppressing dissent, etc. These are defensible arguments to make. Yet there is probably a way to do that without directly analogizing those one disagrees with to Stalin or Mao (which is also a popular tactic on the alt-right).

Peterson et al. might similarly consider avoiding dismissive and derogatory labels like ‘SJW’ or ‘regressive left.’ This kind of language is extremely common on the alt-right. Indeed, opposition to ‘social justice warriors’ seems to be one of the main associations people in that arena draw between themselves and Jordan Peterson (as reflected in Figure 1).

Granted, Peterson’s opponents readily brand him — and his colleagues — as ‘racist,’ ‘sexist’ ‘transphobic,’ etc. It can be difficult not to villainize or caricature them in turn. Yet Peterson et al. explicitly aspire towards a higher level of discourse and rationality than they perceive among many of their interlocutors. Embodying and modeling these alternative forms of discourse, even in the face of such attacks, may help Peterson be more successful in his aim of pulling people away from the fringes instead of towards them.

As a pertinent example, Peterson and his fellow travelers frequently position themselves in opposition to ‘cultural Marxists.’ However, the concept of ‘cultural Marxism’ essentially ties a host of ideas deemed ‘threatening’ to Western civilization to a subversive plot formulated by a cabal of Jewish intellectuals (Derrida, Marcuse, Adorno, et al.). It shares a lineage with ‘cultural Bolshevism’ theories propagated in Nazi Germany… the same Nazi Germany that many of those currently branded as founders of ‘cultural Marxism’ had to flee, upon threat of extermination. As shown in Figure 1, concern about ‘cultural Marxism’ seems to be one of the significant threads white nationalists draw between themselves and Jordan Peterson.

Now, Peterson does not rely heavily on the exact term, ‘cultural Marxist’ (although he does use it occasionally) – preferring instead to warn about ‘post-modern neo Marxists.’ However, he may do well to come up with some other language that is less derivative of the ‘cultural Marxism’ frame to indicate the worldview he seems to be referring to (namely, the reduction of virtually all of society and human interaction into struggles of power, and questions about the distribution or exercise of power — centered around identity characteristics like gender, sexuality or race; the belief that human beings and human societies are infinitely malleable, to be shaped and reshaped according to the aesthetic tastes of those in power).

As a final thought, in addition to his regular evocations of Stalin and Mao, Peterson may benefit from drawing more attention to a U.S.-based movement that helped inspire Hitler – namely, the early 20th century progressive social reform movement. Of course, it would be important not to suggest in any way that those who call themselves ‘progressive’ today are the same as the racists and eugenicists who helped inspire Hitler (as that would be feeding into the aforementioned tropes). However, leaning on this historical example would allow Peterson to fold direct criticism of eugenicist and racist thought into his warnings about the dangers of self-righteous, self-assured, top-down interventionism; they’d be two parts of the same coin. That is, it could simultaneously help Peterson distance himself from (and delegitimize) white nationalist ideology without being diverted from his broader critiques.

Conclusion

The question of what, if any, responsibility authors have for how their work may be used is a complex question about which people can reasonably disagree. We are not asserting here that Jordan Peterson is personally responsible for other people gravitating towards the alt-right, nor that he has a special obligation to prevent people from being drawn to those views. We do not presume to dictate to Jordan Peterson or others in the ‘Intellectual Dark Web’ about what they should or should not say.

However, given that Peterson has himself expressed concern about young people being pulled towards white supremacist and misogynist views — and himself aspires to help pull people away from those views — we have presented a few strategies he might consider to help mitigate others’ attempts to coopt his work in the service of noxious agendas.

All said, our findings confirm the importance of using more objective and data driven approaches to dissect sensational claims more specifically. This will allow conversations about challenging issues to be articulated in a more constructive, specific and less inflammatory manner. In that light, our methodology may yield meaningful insights for advocates of civil engagement, open inquiry and viewpoint diversity — to protect their important message from malicious influence and harmful misperceptions.

Note: The data on Gab in this report were collected by Savvas Zennetou. All data in this piece have been certified by the Network Contagion Research Institute. Any raw, de-identified data will be made available to researchers upon request.