Normiefication of extreme speech and the widening of the Overton window

This research was done with and by Selin Ashaghimina, Agustin Ferrari Braun, Briar Dickey, Jasper van der Heide, Eleni Maragkou, Stijn Peeters, Bryan Steffen and Marc Tuters. Cross-platform visualisations by Andrea Benedetti.

On 23 August 2018, the president of the United States of America, Donald Trump, announced on Twitter that he had instructed his Secretary of State to investigate the ‘South African government seizing land’ and ‘large scale killing of’ white farmers. As is more often the case with Trump’s tweets, this was a response to a broadcast by Fox News anchor Tucker Carlson earlier that day, on the supposed expropriation of white farmers’ land in the country (no such thing had happened). A few months earlier, Carlson had dedicated an item to South Africa’s government’s project to make the country “less diverse” by condoning “brutal murders” (this is also not based on fact). 

Tucker Carlson’s broadcasts about this topic can be seen as part of a wider, increased attention for narratives about white people being systematically murdered – of which the more extreme versions are known as the ‘white genocide’ conspiracy theory. This theory has a long history, arguably originating with the theories of French fascist thinker René Binet in the 1960s. More recently such ideas have been gaining currency in the American conservative media ecosystem, through people such as author Ann Coulter and activist Lauren Southern. That Fox News (and by proxy, president Trump) are now discussing a version of this theory could be characterised as what is popularly called a ‘shift’ or ‘widening’ of the ‘Overton window’; the Overton window constituting the range of topics acceptable for mainstream discourse, which has then shifted or widened to now include this extreme theory. This can then, through increased exposure, lead to more support for the theory. Though the precise concept of the ‘Overton window’ does not have much currency in e.g. communication science, scholars like Whitney Phillips have argued for such an understanding of the media logic of extreme speech; in Phillips’ view, any discussion of extreme concepts gives them ‘oxygen’, amplifying them to a wider audience, and given the extremely problematic nature of these concepts, them having any audience at all is in itself undesirable. Phillips therefore warns that rather than scrutinising these concepts one should simply not give them attention, so not to risk some part of the audience adopting these ideas at their own.

The question then is – is this actually the case, and on what scale? Can we indeed see a process of ‘mainstreaming’ for the conspiracy theory of ‘white genocide’ and related concepts in the wider discourse, and thereby identify a ‘widening of the Overton window’? And does this indeed lead to more people supporting these concepts? The increased attention for this theory about white farmers being murdered in South Africa is but one example of a wider trend from the past few years, in which various types of ‘extreme speech’ have gained currency in political discourse. Extreme speech here refers to ‘incendiary and offensive language’ (Pohjonen & Udupa 2017), or in other words speech that is offensive, outrageous or particularly virulently hateful. Examples of this are various forms of bigotry, heavy usage of slurs, or content promoting white supremacy or other forms of racism.

In earlier work we have seen that in some cases this extreme speech is ‘cooked up’ on fringe internet sites, where – if it manages to gain currency – it can incubate until at some point it has become an established part of the discourse, at which point people that frequent such sites may also start using it elsewhere, e.g. on other sites they visit or in day-to-day conversation. If we characterise the web as a continuum with mainstream spaces such as news media on one end and obscure, subcultural, ‘deep vernacular’ (De Zeeuw & Tuters 2020) spaces on the other, these extreme concepts would originate in the latter, a ‘fringe’ web containing sites such as 4chan, 8chan and parts of Reddit. This can be seen with e.g. the Pizzagate conspiracy theory (Tuters et al. 2018), which was developed on 4chan before finding purchase elsewhere. Similarly, the QAnon conspiracy theory – which supposes that a White House insider has been posting cryptic messages on various web forums for the past few years – originated on 4chan but is now a regular sight at Donald Trump’s rallies (Hagen et al., 2018). This process can be understood as one of ‘normiefication’; from the perspective of 4chan, ‘normies’ are those outside the platform, or people that do not have the subcultural, vernacular capital that would be required to appreciate the layers of nuance, subcultural references, and ironic and sarcastic humour that they would argue underlie such concepts (De Zeeuw et al., forthcoming).

If they indeed undergo such normiefication, these instances of extreme speech can at some point reach mainstream discourse, or indeed the twitter feed of the U.S. president, and thereby attract a wider audience or at least make that wider audience aware that they exist. In the adoption of these concepts by people outside 4chan itself, they perhaps lose some of their definition – as the supposed irony or sarcasm is not appreciated by those who cannot observe it –  but conversely earn a wider audience, which is typically received with some ambivalence; while some despair at the misinterpretation of their discourse by what they see as the dumb masses, other welcome a wider audience for their ideas and opinions.

One reason one could applaud wider discussion of e.g. ‘white genocide’ – a racist conspiracy theory – is that as more people become aware of it, some of those will, through repeated exposure to it, appreciate it as a new and better way of understanding what is going on in the world, and adopt it as their truth. Even if mainstream exposure consists of mocking, ridiculing or harsh criticism, ‘any press is good press’ as before normiefication no one would even know about these ideas, and at least now it is being discussed, which is in itself progress. This is what is referred to as a ‘widening of the Overton window’ on e.g. 4chan. If the Overton window widens, this range increases, and topics or concepts that would have been considered too extreme, niche or taboo become part of the mainstream discourse. That in itself is – in the eyes of those promoting the concepts – an improvement over something being taboo, and one could expect that perhaps over time something that is initially only discussed to denounce it may, if it has some staying power, eventually be discussed on its own merits and find some people who are willing to defend it in e.g. a newspaper’s opinion section.

Our goal is then to investigate if concepts indeed normiefy; and if that ‘widening of the Overton window’ does eventually lead to a more positive reception of the concepts. Our case study for this comprises four concepts that have gained currency on fringe web platforms, even if they did not necessarily originate there. There is a broad far right conspiracy theory that says that – in very basic terms – white or ‘Western’ people are being replaced by non-white or ‘non-Western’ people, and that there is some sort of planning behind this process. This ‘gestalt’ theory has taken various more specific forms, of which we investigate four:

  • The ‘Kalergi Plan’; named after Count Von Kalergi-Caudenhoven, a pivotal figure in the formation of the European Union, this theory posits that at some high level the EU is a project to supplant white, ‘true’ Europeans with non-white immigrants, to achieve a higher goal of a ‘mixed-race’ continent.
  • ‘White Genocide’; a more straightforward conceptualisation of the theory where people see a concerted effort to eradicate the white population of the world or part of it. ‘White Genocide’ is sometimes used as a catch-all for many related conspiracy theory but has notably gained some prominence in the South African context, as a more specific idea that there is an effort to eliminate the white population of that country.
  • ‘The Great Replacement’, a similar idea but originating in France, positing that due to the influx of (primarily) Muslim immigrants the original peoples of various Western countries are slowly being replaced and marginalised. Notably this incarnation of the theory was promoted by a number of far right terrorists in the 2010s in their manifestos.
  • ‘Cultural Marxism’ is a concept that has a somewhat longer history than the other three concepts mentioned here, as an antisemitic conspiracy theory that states that through various institutions (e.g. universities) the left-wing elite is attempting to eradicate ‘Western culture’. As another theory that promotes the idea that there is a concerted effort to eradicate Western culture (rather than the population as such, in this case), it fits this wider ‘gestalt’ theory.

What we can then do is, first, track the activity around these four concepts on various ‘layers’ of the Web, where we take it as axiomatic that some parts of the Web are more in touch with mainstream, broad discourse while others are more fringe, vernacular and focused on small subcultural publics. The best example of the latter platform is 4chan, but based on prior work we did on extreme speech on 4chan it is already clear that all four of these concepts are an established part of the discourse there. What we want to know instead is to what extent these concepts have also penetrated discourse elsewhere. To this end, we capture and analyse data from three platforms:

  • Reddit: Reddit hosts a large number of communities, some fairly mainstream, some very fringe, and many somewhere between these two extremes. This complicates analyses in which it is taken as a unified platform, although it has been argued that generally Reddit communities skew towards the ‘geek sensibility’, nevertheless also containing many ‘mainstream areas’, but on the other hand parts of it can also be characterised as ‘toxic technocultures’ (Massanari 2015). As such it serves as a solid place to start an analysis of whether a concept has entered mainstream discourse, and to what extent. We take all of Reddit as our data set here (via the Pushshift API), and after initial data capture we can then take a closer look at the distribution of activity among separate subreddits.
  • YouTube: While ostensibly a platform with a very ‘normie’ audience, used by virtually anyone with access to it, YouTube can alternatively be positioned as another ‘bridge’ platform. It hosts a bewildering array of content creators, from global media conglomerates to bedroom-based vloggers, and thus likewise serves as a solid place to investigate what parts of the media ecosystem – if any – have adopted the concepts we study, which we do using their own API.
  • News media: ‘mainstream media’ is often used as a synonym for ‘old media’ or ‘legacy media’, originating as tv stations or paper newspapers as opposed to natively digital web platforms. But while many of these legacy outlets were originally slow to transition to online content, particularly newspapers nowadays combine a broad audience with a solid online presence, which makes them suitable for data analysis and comparison with other online platforms. We use the LexisNexis database here to capture relevant articles, and additionally employ CrowdTangle to study the response to these articles on social media.

For all of these platforms, we captured all posts, videos or news articles that mentioned any of the four concepts. We captured data within a four-year window, of which the start is marked by Donald Trump’s announcement of his candidacy for the 2016 U.S. presidential elections, in June 2015. Trump’s ascendancy has been correlated with a rise in prominence of far right (or alt-right) theories and publics, and as such serves as a suitable anchor for our investigation. Next to overall analyses of the data, we manually and inductively coded the attitude of a sample of the data for each platform to assess the attitude towards the concepts as hand, as neutral/unrelated, critical, hostile or positive. This study has an exploratory character, and partially serves as a proof-of-concept that one can study normiefication and the supposed Overton window in this way; at the end, we will make some suggestions on if and how this could be expanded into a more structured investigation of these issues.

Overall results

As a first look at the results, we can simply look at the frequency with which these concepts are discussed on the three platforms we investigate, and how that develops over time. Plotting these frequencies over time for all platforms combined provides an initial, global impression of whether indeed these concepts have been increasingly salient:

As the graph clearly shows, activity for all four concepts increases over time – the frequency of posts, articles and videos about them grows as time passes. There are clear differences between the four concepts; where ‘cultural marxism’ shows solid activity from the beginning, ‘great replacement’ only really picks up towards the end of the interval we study here. One explanation for cultural marxism’s comparatively high activity could be that it has a somewhat longer history than the other three concepts on display, as a conspiracy theory that originated in the late 1990s (Tuters 2018). As such we could be seeing the tail end of its own ‘normiefication’ here, whereas for the other three we are able to capture a fuller view of that process.

A number of particularly catalytic moments can be distinguished in the graph, as especially thick lines; some of these are annotated. One especially strong ‘catalyst’ occurs towards the very end of the interval, in March 2019 coincident with the terrorist attack on a mosque in Christchurch, New Zealand. The terrorist’s manifesto as well as inscriptions on his weapon mentioned the ‘great replacement’ theory by name, which explains the flood of attention this concepts and others adjacent to it receive at the time. Notably, one can see that this attention is sustained for a period of time after the event itself, suggesting that these catalytic events can drive attention for these concept in the long term too.

As an alternative to the visualisation above, one can ‘explode’ the graph and look at the frequencies per platform.

Two significant observations can be made here. First, it is clear that broadly speaking, activity is stronger in the ‘bottom layer’ of the internet – here represented by Reddit – than in the mainstream, which suggests that our understanding of these concepts as originating from the fringes is correct. As we ‘move up’ through the layers, activity becomes somewhat less frequent, and more strongly correlated with these ‘catalytic’ events such as the aforementioned Christchurch massacre. But crucially, we can additionally see that over time, activity increases across all three layers, and the level of activity in news media towards the end of the interval is comparable to the level of activity on Reddit at the beginning of the interval. In other words, the new media is discussing these concepts as actively as Reddit was doing four years earlier. This strongly suggests that these concepts have indeed become normiefied and – because they are all quite problematic and barely discussed at all, initially – this would constitute a widening of the Overton window. 

This is potentially alarming, as the ideas on discussion here are virulently racist and clearly at odds with e.g. democratic ideals. If the ‘all press is good press’ argument underlying ideas about the Overton window holds true, this could signify a shift of the public discourse in a very undesirable direction. As such, this question – is all press indeed good press? – deserves further scrutiny. After analysing the prevalence of these concepts on three different platforms, we therefore next take a look at how they are discussed. Can we indeed see that discussion ‘normalises’ over time, shifting from denouncement to serious consideration and perhaps even acceptance? As all three platforms we looked at have their own idiosyncrasies, next we discuss our more qualitative analysis of how these concepts are discussed per individual platform.


Starting at the bottom layer of our case study, Reddit is a web forum that comprises many individual ‘subreddits’, forums with their own theme (such as politics, funny pictures, a pop culture franchise or a genre of pornography). These subreddits can then be understood as representing a particular public; people interested in a particular topic will flock to the subreddit dedicated to it, and broadly we can then assume that activity in a subreddit means people interested in that subreddit’s theme are engaging with the concept at hand.

One way of understanding the reception of the four concepts we investigate is then to look at what subreddits engage with them, and if and how that changes over time. One thing that is immediately clear is that most of the activity around these concepts occur in subreddits that individually only contribute less than two percents of the overall activity. In other words, the ‘long tail’ is particularly long in this case; the concepts are discussed in many distinct subreddits. This indicates that discussion is diffused across Reddit, with no single nexus of discussion. No clear ‘top subreddit’ can be distinguished among the larger subreddits either; larger subreddits range from far left (ChapoTrapHouse) to centrist (politics) to far right (The_Donald), insofar as a political orientation can even be attributed to them; many are ‘meme subreddits’ that have no clear angle apart from ridicule (CringeAnarchy, TopMindsOfReddit). Activity seems to pick up in right-wing subreddits before it does so in left-wing spaces though, suggesting that left-wing activity may be a response to increased right-wing discussion.

Activity for ‘White Genocide’

On Reddit, these concepts then seem to have penetrated discourse across the board; they are not limited to e.g. far right groups. Activity increases over time, though left-wing engagement lags slightly compared to right-wing engagement. However, if we take a look at how Reddit as a whole engages with these concepts over time, a more nuanced picture emerges. Comparing a sample of posts from early in the studied period to a later sample, it is remarkable that as activity increases, most of that increase seems to be critical or hostile towards the concepts. From this perspective, ‘any press’ is not necessarily good press; increased attention does not seem to eventually give way to nuanced discussion of these concepts, but even after four years remains hostile and critical.

Discussion around the ‘Kalergi plan’ is more stable than for other concepts; as overall activity also develops less strongly, one hypothesis here is that for this particular concept we are still at the beginning of ‘normiefication’, and the concept may still be too ‘out there’ to get picked up by a wider Reddit audience. On the other end of the normiefication process, Cultural Marxism has developed two distinct types of use; it is used as a reference to the actual conspiracy theory, but also as a derogatory shorthand for conspiracy theorists. In both cases however uncritical usage of the term decreases and gives way to hostile and critical discussion over time. Overall, discussion around these concepts becomes more negative as attention increases.


As a platform that is home to a wide variety of content producers, for YouTube one can also investigate the kind of publics which engage with the concepts at hand, by looking at what channels produce popular videos about them. If we look at the top channels for all concepts combined, one finding is that most of them can be described as ‘alternative’, web-native channels, often with a partisan focus. Conventional ‘old’ media, i.e. the YouTube channels of traditional TV networks and newspapers, only occasionally appear within the top results, generally in the context of a particular event, such as the Christchurch massacre. As such, while YouTube is a relatively mainstream platform, the part of YouTube engaging with these concepts is not in itself mainstream. 

Number of videos and types of channels producing content related to ‘white genocide’ between June 2015 and June 2019.

It is important to note that this does not, in itself, mean they have smaller audiences or even that they completely eschew the conventions of traditional media; rather, it is an observation about their place in the wider media ecosystem. These alternative channels were further categorised as ‘alternative left’ and ‘alternative right’ (which here does not necessarily refer channels affiliated with the alt-right movement or an alt-left counterpart) in an effort to add nuance and gauge the potential emergence of progressive counterpublics, which counter these fundamentally extreme right ideas.

Top 20 alternative YouTube channels in the far-right conspiracism discursive space, ranked by subscriber count and comment engagement respectively and their affiliation. Visualization made with RAWGraphs.

Although right-wing channels dominate this space in terms of output, their quantitative hegemony is challenged by a number of progressive creators associated with BreadTube, a loose assortment of leftist YouTubers. As we are dealing with right-wing conspiracy theories, one can assume that the left-wing channels would be engaging with these concepts through criticism. Indeed the videos from these channels mostly serve to debunk, if not outright dismiss these far right ideas. While this might be interpreted as a shift of the Overton window to the right, legitimising in some sense these concepts in the field of ideas, it is important to note here that hostile or critical engagement remains the case over time, even increasing towards the end of our study interval.  

Of note in our study is the function some of these content creators serve as (political) influencers. Building on Rebecca Lewis’ work on the Alternative Influence Network, our research suggests that behind the channels stimulating engagement (both on the left and on the right) are personalities rather than uploaders of recycled content — figures like Natalie Wynn (ContraPoints) and Oliver Thorn (Philosophy Tube) on the left and Lauren Southern and Carl Benjamin (Sargon of Akkad) on the right, who cultivate and calibrate recognizable digital personae and aesthetics, and personalities like Jordan Peterson, who features prominently via uploads of his own lectures and interviews or commentary videos.

Type of engagement for the top 50 most viewed videos per query between June 2015 and June 2019.

While three of the four concepts investigated have at some point achieved a certain notoriety, especially in their associations with Donald Trump, the alt-right and white nationalism, once again ‘Kalergi’ is an outlier. Content about the Kalergi plan appears to be produced largely by smaller, European right-wing channels, and engagement with these videos is lower compared to the other queries. This could again be an indication that the Kalergi plan conspiracy is still rather obscure and has not yet penetrated wider discourse in the same way as the other concepts have.

News Media

If we finally take a look at how these concepts are covered in news media, one obvious observation is that more so than other platforms news media activity seems to be a response to ‘real world’ events that give cause to write about these concepts, e.g. the aforementioned Christchurch massacre and other political events (such as Donald Trump tweeting about White Genocide). This makes sense – the news media exist to write about news events. But it does underline the role news media (still) play in bringing these concepts to a larger audience. Not only do we see a boost in news media coverage, but this activity resonates in other layers as well, as the news articles are shared on Reddit, or YouTube content creators create videos about the events reported on in the news. This could be said to be the moment when normiefication is complete; mainstream news media reports on a hitherto fringe concept, and that in itself boosts discussion of the concepts on other platforms as well, providing ‘fuel’ to it that allows it to remain relevant as part of the discourse.

As we again take a look at the type of coverage the concepts enjoy, it is again mostly negative and hostile, though ‘Cultural Marxism’ again receives more ‘normal’ engagement as well, consistent with the findings on other platforms. As many of the articles in our data set were originally published online, we can gauge the reaction of their audience through CrowdTangle, a service that allows one to see social media activity (on e.g. Twitter, Facebook and Instagram) around a particular web link. Here too it is apparent that most of this activity is in fact critical and hostile; in other words, not only are the writers of these articles negative towards their topic, but this attitude seems to transfer to the audience of these articles too. As such, the findings here are more or less consistent with those for Reddit and YouTube, and reinforce the notion that while normiefication seems to be taking place, the idea that this would promote a gradual widening of the Overton window and increasingly positive reaction may need nuance.


We investigated the activity around four incarnations of a racist conspiracy theory over the course of four years on three distinct platforms. This served as a way of empirically exploring two related concepts. ‘Normiefication’ represents the normalisation and adoption by mainstream audiences of concepts that were cooked up or incubated on fringe, vernacular web platforms. ‘Widening the Overton window’ is how some on these platforms describe the entrance of ideas that used to be taboo because of their extremeness into mainstream discourse. By looking at the activity around our case studies and how they are adopted in various places we could provide a first impression of whether these indeed became normiefied and if so, whether that does indeed constitute a problematic widening of this Overton window where the ideas become more acceptable to these mainstream audiences over time as well.

What we found supports the notion that these concepts have been normiefied; particularly ‘white genocide’ and ‘the great replacement’ were marginal topics at the beginning of our four-year period but saw sustained uptake across all layers of the media we studied at the end of it. ‘Cultural Marxism’ saw a less pronounced increase in activity, perhaps because it is an older conspiracy that has already found a more stable media niche. ‘Kalergi plan’ also saw a smaller uptake, possibly because it is still a relatively obscure conspiracy theory that is still outside this Overton window.

On the other hand, the hope of some of those promoting these conspiracy theories, that over time they become more accepted and are discussed more seriously, is not confirmed across the board. Even as activity around particularly ‘white genocide’ and ‘the great replacement’ increases, based on the samples we took this discussion is predominantly critical and hostile towards these concepts. In this case not all press is good press.

This constitutes a preliminary exploration of these concepts and their changing niche in the media ecosystem. This first exploration proposes a protocol for empirically verifying these concepts of normiefication and the Overton Window. A next step here could be a more systematic investigation of audience responses to increase coverage of these concepts, for example; an investigation of other platforms besides the three on offer here; and an investigation over a longer time period, to capture the full process of normiefication. These could add more depth to the already suggestive findings on offer here.


De Zeeuw, D., & Tuters, M. (2020). Teh Internet Is Serious Business: On the Deep Vernacular Web Imaginary. Cultural Politics16(2).

Hagen, S., De Zeeuw, D., Peeters, S., Jokubauskaitė, E., & Briones, Á. (n.d.). Understanding Normiefication: A Cross-Platform Analysis of the QAnon Conspiracy Theory. Digital Methods Initiative.

Massanari, A. (2017). #Gamergate and The Fappening: How Reddit’s algorithm, governance, and culture support toxic technocultures. New Media & Society19(3), 329–346.

Phillips, W. 2018. The Oxygen of Amplification. Data & Society.

Pohjonen, M., & Udupa, S. (2017). Extreme Speech Online: An Anthropological Critique of Hate Speech Debates. International Journal of Communication11(0), 19.Tuters, M. (2018). Cultural Marxism. Krisis2

Tuters, M., Jokubauskaitė, E., & Bach, D. (2018). Post-Truth Protest: How 4chan Cooked Up the Pizzagate Bullshit. M/C Journal21(3).

One thought on “Normiefication of extreme speech and the widening of the Overton window

  1. South African farm attacks happen at a similar frequency as the amount of unarmed black people being killed by police in America each year, but according to mainstream media one is a conspiracy well the other isn’t.

Leave a Reply

Your email address will not be published. Required fields are marked *