- 09 Feb 2018
The Deep Vernacular Web
Theory Delirium: How 4chan/pol/ Cooked-up Pizzagate
A Day in the Life of 4chan /pol/: The Themes and Memes of the "Meme Factory of the Internet"
Theory Delirium: How 4chan/pol/ Cooked-up Pizzagate
- Daniel Bach
- Emilija Jokubauskaite
- Marc Tuters
On December 4th 2016, a man entered a Washington DC pizza parlor armed with an AR-15 assault rifle in an attempt to save the victims of an alleged satanic pedophilia ring, run by prominent members of The Democratic Party. This act was based on a conspiracy that had grown out of the online forum 4chan/pol/, a known online hangout and hub for the US ‘alt-right’ movement. On this forum, a sort of decentralized research effort had taken place (what we, as digital methods researchers, would call a ‘data sprint’), with its most intensive period having taken place between 2nd and 4th of November 2016. Within this timeframe elements of leaked Democratic Party emails were analysed to contain secret messages that could link high ranking members of the democratic party to an existing satanic pedophilia ring. While there has been some initial quantitative research into 4chan/pol/ from the perspective of computational linguistics (Hine et al. ‘17, Bernstein et al. ‘11) as well as significant qualitative research of its “trolling subculture” (Phillips ‘15) from the perspective of digital folklore studies, the current research represents the first effort to combine close and distant reading techniques by studying 4chan/pol/ using digital methods. Heavily promoted by the conspiracy theorist Alex Jones and the alt-right celebrity Mike Cernovich, the story of Pizzagate’s secret child sex ring based out of a DC pizza parlor has been covered extensively by all major news organizations, yet (to the best of our knowledge) the origins of the conspiracy has never been revealed… until now.
Initial Data Sets
- a third party archive of 4chan/pol content (claims to archive full content)
How was #Pizzagate created out of the arcane preoccupations and technical affordances of 4chan & what were the role of other platforms and tools?
Previous research conducted in the DMI Winter School 2017 has indicated an extended use of a variety of different tools by the users of subreddit r/The_Donald. Among the discussed infrastructure of this community, which is often defined as representing the alt-right, interesting cases of the use of pastebin.com and docs.google.com documents have been uncovered (Gestsdóttir, Indriunaite, Jokubauskaite & Nothnagel). These examples often showed explicit or implicit call for action, self-investigative content or reporting on issues (Gestsdóttir et al.) Inspired by these findings, the current research intends to look into the accumulation of facts on a different vernacular web platform - 4chan and its Politically Incorrect board (4chan/pol). It was the discovery of these peculiar seeming infrastructure of alt-right activists and self-described meme warriors (MYNUNUDONALDACCOUNT ‘16) that drew us into the present research.
The Welcome to /pol/ - Politically Incorrect post, stickied at the top of the board, states:
‘This board is for the discussion of news, world events, political issues, and other related topics. <...>The variety of threads allowed here are very flexible and we believe in freedom of speech, but we expect a high level of discourse befitting of the board’.
4chan/pol/ is known to have been a central figure in the outlandish 2016 US election season, as it has often been linked to the alt-right movement and its rhetoric of hate and racism (Hine et al.). Some of the previously studied characteristics of the platform consist of anonymity
(with regards to other users - even though users can enter a user name, they are, however, not required to), ephemerality
(‘<...> threads with the most recent post appear first, and creating a new thread results in the one with the least recent post getting removed <...>’), moderation
(‘<...> so-called janitors, volunteers periodically recruited from the user base, can prune posts and threads, as well as recommend users to be banned by more “senior” 4chan employees <...>’) and other (Bernstein et al. ‘11). It is important for our research to state that in relation to the ephemerality
characteristic, new posts contribute to bumping the threads up and ‘keeping them alive’, however after a certain limit of bumps or images is reached, the new activity no longer keeps the thread at the top of the board and it gets purged.
In the process of our research we came upon what we believe to be the origins of the famous Pizzagate conspiracy theory which we found to have been created through a collaborative process of knowledge accumulation over the course of 26 hours on 4chan/pol/. In order to trace the creation of this knowledge accumulation, a third party archive had to be queried for the needed data, since the platform itself does not store archived threads. Archive4plebs was chosen due to their claim to store all of the content from 4chan/pol.
Since previous research (Fisher et al. ‘16) has indicated that the creation of Pizzagate conspiracy is related to the Wikileaks release (2016) of Podesta emails (Oct 7), the time period of a month around that range was chosen for the initial analysis (2016 Oct 5 - Nov 10). By querying the term ‘pizza’ in that date range, it was possible to pinpoint the first instance of a thread (/pol/ - Politically Incorrect » Thread #95752720) that discusses a connection between the keyword and the leaked emails.
Starting with the timestamp of the first thread (Wed 02 Nov 2016 22:17:20), the next 40 threads were documented and analysed for tracing the creation of the conspiracy. Among the 40 threads, 18 ‘general threads were characterised as the most relevant and were chosen for further analysis (19 threads in total).
_Image 1 (click on image for full size)_
The ‘general threads’ (image 1) can be seen as an attempt to accumulate knowledge on a specific topic, in this case - the pizzagate conspiracy, by appending new information at the bottom of a new post and this way preserving previously found information. Creating new threads aids in keeping the same information up on the board after the previous thread gets purged. The last of the ‘general threads’ in this discussion seem to have quickly disappeared from the board, since the number of posts on them is very limited and their content discusses the fact that the new posts cannot bump the threads up anymore (a phenomena 4chan users call 'autosaging'). The reasons for this remain unknown, however this is where the current study draws the line of the end of the primary conspiracy creation.
By limiting the corpus to the ‘general threads’ and the first thread that is thought to have sparked the discussion, the final timeframe of the analysed ‘birth’ of the conspiracy has been set to be between
Wed 02 Nov 2016 22:17:20 and
Thu 03 Nov 2016 23:24:01 - a period of slightly more than 25 hours and a total of 19 threads. The content of the delimited corpus was then analysed.
Two different ways of analysis have been conducted on the delimited corpus.
First of all, a preliminary grounded-theory-based content analysis has been conducted by 3 independent researchers. After initially familiarising with the content, a bottom-up approach has been chosen to categorise the content in relation to various research practices, implemented in the academia. The following non-exclusive categories have emerged, and a number of examples have been chosen to illustrate each of them.
=Linking to primary sources= - providing links for the original leaked Podesta emails as information relevant for the build-up of the theory;
=Discussing primary sources= - mentioning already linked primary sources or linking to them and discussing their relation to the theory;
=Linking to secondary sources= - mentioning other than primary sources - for example posts by others that discuss relevant information;
=Using tools= - linking to external analytical tools, such as Gematria Calculator or Stegsecret;
=Compiling information= - gathering the already discussed information into collections, such as one post or a Pastebin document, also includes posts that urge to include some information into new ‘general threads’;
=Critique= - providing critique for the arguments with regards to the theory, findings uncovered etc.
=Calling for action= - asking for particular actions to be taken in relation to the theory build-up, for example the creation of thematic hashtags, infographics etc.
=Calling for expertise= - asking users with specific expertise to complete tasks (for example users with skills in visualisations, cryptography etc.)
=Archiving information= - reminders to archive the information (archiving of primary, secondary sources as well as original findings)
=Publishing on other platforms= - posts that provide links to other platforms, indicating that the findings are posted there for a different audience to see.
=Visualising= - posts that provide visualisations of information or ask for them
The categories that emerged were visualised in image 2 (below).
Image 2 (click on image for full size)
Secondly, in order to follow the information spread between 4chan/pol and other platforms (8chan, reddit etc.), the researchers analysed all the posts from the category of ‘publishing on other platforms’ and those that mention or include url links to other platforms. A total of 13 timeline points were chosen to represent the most key moments in the development of the Pizzagate story (image 3, below) -- for reporting on how the pizzagate conspiracy has spread readers can look elsewhere (See: Robb ‘17).
Image 3 (click on image for full size)
Findings and discussion
The decentralized methods of research that we observed on 4chan/pol/ in the course of the creation of the Pizzagate story could be thought of as seen as a kind of local vernacular rendition of digital methods. While implementing many of the same research techniques, as itemized above, their objective is however markedly non-scientific. As has been thoroughly documented elsewhere, practically none of the claims generated by the Pizzagate conspiracy theorists have stood up to analysis (Robb ‘17) -- in light of this scrutiny, even the notorious conspiracists Alex Jones and Mike Cernovich have tried to distance themselves from the “facts” of the story that they had once worked so assiduously to promote (Doubeck ‘17, Colbourn ‘17).
What our analysis seems to reveal is that the affordances and temporality of 4chan, alongside its unique cultural preoccupations with arcana and subcultural trolling humour, contribute to create an environment that is uniquely productive of bullshit -- here we are referring to a technical problem within a subfield of information science referred to as bubble studies (Hendrix & Vestergaard ‘18), in which balkanized online speech markets (Sunstein ‘01) appeal to an innate tendency towards emotion based decision making that confirms preconceived biases at the expense of shared epistemic principles (Lynch ‘16).
In terms, first, of the technical affordances of the platform, the key elements that makes 4chan/pol/ so incredibly productive of bullshit are the practice of creating so called “general posts” as well as the creation of data visualizations. Since posts are archived or ‘purged’ after a period of declining interaction, what we found was that the creation of the pizzagate conspiracy came about through the use of these compilation techniques.
_Image 4 (click on image for full size)_
The general posts serve as a way for the community to stay updated about what is going on in the current research effort. These posts compile the previous discussion threads posted as comments underneath them, and serve as the main thread for posting new research and aggregations of knowledge regarding existing lines of inquiry and theories. (Image 4 (above) shows an example of a general post.) This practice affords significant power for a single actor who is in the position to define the controversy for the community, deciding on which parts of a thread to include and what to exclude, in other words to develop a narrative out of the overabundance of data in the process of summing-up “the story so far”.
Image 5 (click on image for full size)
It would furthermore seem that this particular function can only be held by one person at at time so that redundant general posts are simply ignored (image 5 (above) shows how a compilation seemingly made by a different user does not make the cut, thereby supporting the influential status of one particular actor for the theory building). It is thus clear on close inspection that, far from reflecting the entirety of what had been posted in a given thread, these general posts are in fact highly selective. As might be expected, it seems that elements tend to make it into the general posts if they are repeated persistently throughout many different threads.
Image 6 (click on image for full size)
The visualisations made during the same process serve the same goal - to help combine and compile the research (image 6, above). To circumvent the process of purging, these images are often stored on other sites such as Imgur. Image 7 (below) shows how child-related businesses have logos of the same aesthetic, and are roughly shaped into an inverted pentagram, and used as proof of the conspiracies legitimacy.
Image 7 (click on image for full size)
#Pizzagate was born out of 4chan’s arcane preoccupations, which included a long running fascination with a sex scandal involving the Clintons known as "Orgy Island", which alleged the Clintons to have flown to a secret island for sex tourism aboard a private jet called "Lolita Express" owned by Jeffrey Epstein, a man who had served 13 months in prison for soliciting an underage prostitute. Elements of the pedophile story at the centre of the Pizzagate conspiracy thus predated #Pizzagate as such. Indeed, a conspiracy theorist by the name of Steve Pieczenik had been making peddling this narrative on the right wing liberation talk show circuit as far back as at least March 2015 (Anon ‘15).
During the 26 hour period that gave birth to #Pizzagate, then, these stories were combined with other clues found in the Podesta emails to create the famous narrative, so well covered by so many journalists. To pick only one such example, defiant this 26 hour period an email was discovered sent to Tony Podesta from the performance artist Marina Abramovich in which she wrote: "Dear Tony, I am so looking forward to the Spirit Cooking dinner at my place". A term used to refer to a sacrament in Aleister Crowley’s satanic religion of Thelema, the term "Spirit Cooking" fit perfectly with 4chan’s preoccupations with the occult.
Esoteric preoccupations are relatively ubiquitous across memes in 4chan/pol/ as represented by the concept of Kek, a term used to refer to chaos magic. In its preoccupations with esotericism, 4chan/pol/ could be thought of as a "cultic milieu", in the sense developed by sociologist Bryan Wilson (1990), for spaces that accumulate alternatives perspectives to those of the dominant culture. In its seemingly ironic tone, 4chan’s cultic milieu is reminiscent of Robert Anton Wilson’s Discordian religion that advocated a radically idealistic worldview and which embraced chaos. In light of the alt-right elements that took root in 4chan, it can more darkly be understood in relation to the reactionary strains of Traditionalism as represented by Julius Evola whose anti-modernist thought inspired 70's Italian Far Right terrorists.
Conclusion & Future Research
This findings of this research were significant and original from a number of perspectives. Journalistically we identified the origins of the much reported Pizzagate story as having been created the period of a single day on 4chan/pol/. In terms of new media studies we observed how the affordances of the platform, and its unique temporality might understood to have impacted upon the creation of this peculiar and highly successful conspiracy theory. As digital methods researchers we found ourselves studying something uncannily familiar, a data sprint process in which researchers create visualizations in order to narrativize complex data sets. While perhaps superficially similar to some of our own research methods, this delirious approach to knowledge accumulation is however characterized above all by its wilful rejection of objective standards of verification; in other words, to put it scientifically, it’s bullshit.
Given the impact of the Pizzagate story, not to mention the many other narratives and memes that have been born there, 4chan/pol/ is a highly significant site of study. That very little research has been conducted on 4chan may stem in part from the fact that it is not easy to study because the lack of a fully reliable and authoritative archive. In conducting this research, we consider ourselves to have been lucky to have found archives of the pages that we were looking for. Going forward however researchers interested in studying 4chan would benefit from having access to an archive that could be queried as a whole dataset. To this end, one of the current initiatives of the Digital Methods Initiative is to develop such a tool.
Anon_Resistance, (2016) "Epstein's Little Black Book," Pastern, July 2016, https://pastebin.com/m7FYj73Z
Bernstein, M. S., Monroy-Hernandez, A., Harry, D., Andre, P., Panovich, K., & Vargas, G. (2011). 4chan and /b/: An Analysis of Anonymity and Ephemerality in a Large Online Community. Presented at the Proceedings of the Fifth International AAAI Conference on Weblogs and Social Media.
Colburn, R. (2017, December 4). Celebrate the 1-year anniversary of the #pizzagate shooting by getting Mike Cernovich kicked off Twitter. Retrieved February 15, 2018, from https://www.avclub.com/celebrate-the-1-year-anniversary-of-the-pizzagate-shoo-1820983596
Doubeck, J. (2017, March 26). Conspiracy Theorist Alex Jones Apologizes For Promoting “Pizzagate.” Retrieved February 15, 2018, from https://www.npr.org/sections/thetwo-way/2017/03/26/521545788/conspiracy-theorist-alex-jones-apologizes-for-promoting-pizzagate
Fisher, M., Cox, J. W., & Hermann, P. (2016, December 6). Pizzagate: From rumor, to hashtag, to gunfire in D.C. The Washington Post. Retrieved from https://www.washingtonpost.com/local/pizzagate-from-rumor-to-hashtag-to-gunfire-in-dc/2016/12/06/4c7def50-bbd4-11e6-94ac-3d324840106c_story.html?utm_term=.ef9c2b1edc2f
Gestdottir, Anna Karolina, et al. “Alt_Activism: An Analysis of the Means of Mobilisation within the /R/The_Donald Community and Its Infrastructure of Digital Tools.” Digital Methods Initiative, Jan. 2017, https://wiki.digitalmethods.net/Dmi/WinterSchool2017TrumpAltRightActivism.
Hendricks, V. F., & Vestergaard, M. (2018). Postfaktisch: Die neue Wirklichkeit in Zeiten von Bullshit, Fake News und Verschwörungstheorien. Berlin: Karl Blessing Verlag.
Hine, G., Onaolapo, J., De Cristafora, E., Kourtellis, N., Leontiadis, I., Samaras, R., et al. (2017). Kek, Cucks, and God Emperor Trump: A Measurement Study of 4chan's Politically Incorrect Forum and Its Effects on the Web. Association for the Advancement of Artifical Intelligence.
Lynch, M. P. (2016). The Internet of Us: Knowing More and Understanding Less in the Age of Big Data. New York: W. W. Norton & Company.
MYNUNUDONALDACCOUNT, (2016) "Advanced Meme Warfare /cfg/," Pastebin, July 2016, https://pastebin.com/hack9Z6G
Phillips, W. (2015). This Is Why We Can't Have Nice Things: Mapping the Relationship between Online Trolling and Mainstream Culture. Cambridge MA: MIT Press.
/Pol/ - Politically Incorrect » Thread #95752720. http://archive.4plebs.org/pol/thread/95752720/#95752720. Accessed 18 Feb. 2018.
Sunstein, C. (2001). Fragmentation and Cybercascades. In Republic.com (pp. 51–88). Princeton.
Robb, A. (2017, November 16). Anatomy of a Fake News Scandal: Inside the web of conspiracy theorists, Russian operatives, Trump campaigners and Twitter bots who manufactured the “news” that Hillary Clinton ran a pizza-restaurant child-sex ring. Rolling Stone. Retrieved from https://www.rollingstone.com/politics/news/pizzagate-anatomy-of-a-fake-news-scandal-w511904
WikiLeaks (2016, October 7). WikiLeaks - The Podesta Emails - Part 1. Retrieved from https://wikileaks.org/podesta-emails/press-release
Wilson, B. (1990) The Social Dimensions of Sectarianism: Sects and New Religious Movements in Contemporary Society. Oxford: Clarendon Press.
A Day in the Life of 4chan /pol/: The Themes and Memes of the "Meme Factory of the Internet"
- J.P Guarnieri (11762497)
- Sal Hagen (sal.hagen@uvaREMOVE_ME.nl)
- Simoon Hermus
- Nicole Leeflang (11932376)
- Raquel Schilder
- Ayşe B. Tosun (11780142)
- Muus VIsser
Since the 2016 U.S. presidential election, the infamous imageboard 4chan, being known as an obscure and offensive Internet environment, managed to reach a mainstream public as the community that “memed Trump into office” (Ohlheiser, 2016). While 4chan has been in and out of the public eye for the last few years due to their various brigades and offline endeavors (notable examples: Anonymous, as well as 4chan’s takeover of the Scientology website, or “Project Chanology”
), the part it played in the election catapulted it into a forum of interest for broader audiences than bored teenagers or meme-makers.
4chan was founded in 2003 by the then 15-year-old Christopher Poole (KnowYourMeme
). While it began as a place to discuss anime, it soon devolved into different subjects, eventually becoming known for its free-for-all mentality, bordering on insensitiveness and politically incorrectness (Hine et. al, 92). Due to the ephemeral and anonymous affordances of the platform, memes thrive on 4chan, making it the “meme factory” of the Internet (Stryker 2011). A meme can be defined as a cultural unit that is perpetuated through society (Stryker 2011). On the Internet, memes usually take the form of images, GIF’s and catchphrases and are shared with small variations on the theme. An example of a 4chan-perpetuated meme is Pepe the Frog - a fictional cartoon character created by Matt Furie in 2005 for his comic “Boy’s Club”. Pepe gained widespread traction on Myspace, 4chan, and other online platforms, quickly becoming a popular Internet phenomenon. (Know Your Meme, 2016). It was soon assimilated by Trump supporters and popularized during Trump’s presidential campaign, particularly on 4chan’s board /pol/.
While 4chan has 70 boards that cover different topics, one of its most popular remains /pol/, a board for discussing politics, and adopting the surname “Politically Incorrect”. Put very mildly, this 'political incorrectness' is apparent throughout the board, which assaults the average user with shocking images of gore, Nazism, explicit sexual activity, and other controversial visuals that are often meant to repulse any “normies” or regular internet users (Urban Dictionary, 2018). This has led to 4chan gaining the reputation of being the “asshole of the Internet” (Philips, 2015). However, this should not discredit the fact that 4chan’s /pol/ has provided a hub for those with newfound right-wing sentiments, often being described as one of the most prominent 'alt-right' onlinse spaces. Some /pol/-anons claimed, in their imaginary world of ‘meme magick’, to have “memed the President into the Oval Office” through a grassroots attempt of spreading pro-Trump memes around the Internet (Burton 2016). While the /pol/'s influence is hard to measure, Trump did in fact retweet Pepe the Frog, the froglike mascot of the alt-right, and an image containing a David Star, both images appearing first on /pol/ (http://archive.is/E0Dto
). Therefore, the infamy and controversial opinions on /pol/ do not take away the fact that the community can cause off-platform influence.
While discussing these often untolerable imagery has the unfortunate effect of spreading them even further, like done here, this widespread influence renders /pol/ a space of direct political interest. Their role as a “meme factory” allows us to examine not just the use of older memes and how they have evolved over time, but also the creation of fresh memes. Although /pol/ is mainly known for its recent cultural involvement in Trump’s political campaign, now that time has passed since the election period, it is interesting to examine how political attitudes have changed and shifted focus since, if they have even done so at all. In order to do so, we decided that categorising the images used on /pol/ might provide a general sense of what topics the images relate to. Based on these reasons and the exploratory nature of this research, the following research question was formulated:
“How can we categorize and explain the memes used on 4chan /pol/ in both a qualitative and quantitative sense?”
To gain an understanding of the visual characteristics and narrative nuances of /pol/, we found that it was essential to capture the board's entire content, including detailed post data. All images of 4chan /pol/ were fetched within a 48-hour period between January 7th 2018 06:00:00 and January 9th 2018 06:00:00. These snapshots were taken at an interval of 3 hours, using a Python script. This 'chanscraper'-script uses 4chan API (https://github.com/4chan/4chan-API)
to store all active threads and images on 4chan/pol/ at a specific moment. All .png, .gif and .jpg images were stored in a folder (.webm videos were discarded due to size issues), and the post-data was stored in a csv file. The data stored in these snapshot csv-files includes the following columns:
- 'threadnumber': the unique number of the thread the post/comment appeared in
- 'no': a unique value assigned to the post/comment
- 'now': the time of posting of the post/comment in UTC
- 'time': the time of posting of the post/comment in UNIX timestamp
- 'comment': the text of the post/comment. All original HTML-values are escaped, generating plain text.
- 'subject': the thread 'subject' title (if an entry is the first in a thread))
- 'replies': the amount of comments in a specific thread (if an entry is the first in a thread)
- 'uniqueips': the amount of unique IPs that contributed to a thread (if an entry is the first in a thread)
- 'id': the ID of a poster, based on IP-number
- 'country': The country name of the IP's origin. Note: these are often spoofed by VPNs and proxies.
- 'imagefile': a unique value for the image file that was included with the post/comment (empty if no image is posted)
The code for the aforementioned 4chan-scraper tool is available at https://github.com/salhamander/4chanscraper/ .
The 'three-hourly' snapshot-csv files and the respective images were used as the basis for the subsequent research. As 4chan is ephemeral by nature and images and threads get bumped off every few minutes, the use of this script was essential. As a result of this process, roughly 56.000 images were captured.
While collectively thinking about how to tag the images, the group quickly noted that one could either tag the images on their formal attributes (e.g. 'is it Pepe or not?') and on the general theme the image is used for (e.g. 'is this image pro-Trump?'). Therefore, we decided to attribute the formal aspects to a 'meme'-tag, and the thematic use with a 'theme'-tag, especially because doing so allowed to relate the formal aspect to the themetic usage ('which memes are used for which memes, and vice versa?').
and jQuery code was written. The resulting 'Memetagger' allows uploading the snapshot csv-file and showing the data (using the library jquery-csv, https://github.com/evanplaice/jquery-csv)
. One row of the csv is 'read' to visualise the text and the image of the respective post. A sidebar shows the meme- and theme-tags, which can be selected to tag the image. Clicking 'next image' shows the next post/comment, and so on. The page allows exporting the results in JSON, generating a file with the the respective theme- and meme-tags. The Memetagger script is available at https://github.com/salhamander/imgtagger.
Each group member then analyzed their own unique 4chan-snapshot, being responsible for analyzing and categorizing the images according to specified theme- and meme-types. To narrow down and increase the likelihood of identifying similar images or memes, ImageSorter
was used to sort the images by color. By sorting these images by color, the memes in question became easier to discern, yielding a range of themes, namely antisemitism, (neo)Nazism, racialism, European nationalism, anti-Islam, antileft, anti-liberal, misogyny/sexual imagery, conspiracy, Trumpism, anti-(plebeian)USA, historical imagery, crusades/Christian imagery, Egyptian imagery, Nordic imagery, Vaporwave, LGBT-phobic, multi-meme and esoteric. In addition to themes, specific types of popular and emerging memes were also identified, such as Pepe the Frog, Happy Merchant, Spurdo, Picardia, Amerimutt, Feels guy, thinking emoji, big grin, Anime/hentai, Doge and Taylor Swift.
Before commencing the coding process individually, we also took a sample of approximately 200 images and sorted them as a group to establish a consensus on how to sort these images, as well as how certain memes and themes correspond to each other, and whether or not context and captions should be taken into consideration. Ultimately, it was decided that the images should be taken as self-explanatory and verge-cases should be labeled as 'other', avoiding cluttered datasets. Furthermore, once the individual sorting process began, any ambiguous images instantly were shared with the group and discussed before categorization.
A visual example from Image Sorter, with one subset of images, sorted by color
Upon establishing the themes- and memes-tags and tagging the images, another Python script was written to merge the original snapshot-csv file with the exported JSON-files from MemeTagger
, resulting in a new .csv with two extra columns: 'themetags' and 'memetags'. The posts/comments without an image were removed, as well as any duplicate entries (since there was some minor overlap between the snapshots). The resulting sheets with tagged image-posts were merged into one large Google spreadsheet. This file was then used for generating the visualisations.
An overview of the MemeTagger
After tagging +- 5000 images (obtaining a sample of roughly 10%), RawGraphs
data visualization application (available at http://app.rawgraphs.io/
) was used to visualize the relevant findings. Throughout this whole process, the website KnowYourMeme
was essential in helping recognize and identify the different memes used, as well as giving context and history to each meme.
Based on the output of the processes above, there were three main findings which proved salient.
Themes and Memes
Firstly, the links between themes and memes were clearly established, with variations per meme also becoming apparent. When looking at Image 3, it becomes clear that the link between themes and memes highly differs per meme. While some memes are used for more tactical or topic-specific meme-ing, others are more versatile, mainly being used as “vessels” to convey other meanings and perpetuate other themes. To clearly illustrate this phenomenon the use of the most-used memes will be addressed: Pepe the Frog, Amerimutt (aka the '56%-meme'), and the Happy Merchant (aka 'Shlomo Shekelberg').
Themes of Memes - Larger bubbles are memes, bubbles within are themes
Themes Related to Pepe the Frog Meme Image 5:
Themes Related to Amerimutt Meme
In images 4 and 6, we can see that although Pepe the Frog has often been associated with Trumpism, with even Trump tweeting Pepe, the meme relates to many different (though not less-troubling) themes, such as Nazism and racialism. When compared to an extremely theme-specific meme such as the Amerimutt, seen in Image 5 and 7, we can see that the main themes that persist throughout the Amerimutt usage are racism and anti-USA. This usually occurred in conjunction with each other.
Themes Present in Pepe the Frog Meme
Themes Present in Amerimutt Meme
Other prevalent memes contain such as Feels guy,
also fell under the umbrella of memes used for general purposes. Theme-specific memes contained the following: Taylor Swift, Historical Imagery and the Happy Merchant. Taylor Swift was constantly associated with Nazism, Trumpism, and racialism, despite there being no explicit link between this celebrity and these ideologies. Still, Swift embodies the ideal feminine picture for many /pol/ users. The use of Historical Imagery was mainly related to Nazism, with its second most popular theme being racialism.
Lastly, one of the most interesting relations between themes and memes is the Happy Merchant. The Happy Merchant
almost always entailed anti-semitism or conspiracy (usually correlated). But, although it was the biggest symbol of the rampant anti-semitism on 4chan, some examples of the meme indicate it might slowly turn from a topic specific-meme to a general-purpose “floating signifier”, a la Lacan (Mehlman, 1972).
As previously mentioned, 4chan /pol/’s claim to fame has largely involved its creation of pro-Trump memes during the 2016 U.S. presidential election. But, a year later, the sentiments towards Trump have shifted towards a more ironic and critical tone. As much as 4chan was linked to Trump’s success in the election, there were not much pro-Trump images on the board in our snapshots taken in January 2018:
Themes on 4chan /pol/. Trumpism (top) only comprises a small part
A Discussion by /pol/ Users About the Board’s Feelings Towards Trump
In Image 8, we can see that while Trumpism is still discussed, it is no longer the most common theme on the board. In Image 9, we can also anecdotally discern that this discussion of shifting political allegiances is also occurring on the board itself, with a 4chan-veteran predicting how feelings towards Trump will change pre-, during, and post-presidency.
Prevalence of Racism
The Amerimutt/Le 56% Face
Another noteworthy trend was /pol/'s prevalence of the rampant racism, which, although in many ways ingraned in 4chan's transgressive DNA, was now also embodied in a specific meme: the Amerimutt (Image 10). Originally stemming from the European migration crisis,
American users have been criticizing Europe for accepting refugees from countries that are non-white. In that light, the Amerimutt meme serves to point out the hypocrisy of these American channers by reminding them that their country is extremely diverse itself, with the United States’ population only being 56% 'white' (Know Your Meme, 2018). Unlike other 4chan memes such as Pepe or Feels Guy, this meme has luckily not widely reached mainstream platforms (yet).
Tagging the /pol/ images with both 'theme' and 'meme' labels proved a sufficient method to render visible some trends. Most noticeable is a difference not just in usage, but in internal variation among some memes. While some templates, such as Feels Guy and Pepe were very versatile, appearing on a wide range of themes with wild aesthetical variation, others, like Amerimutt, were linked to specific contexts. These tended to be formats inherently linked to a theme (Amerimutt, for example, is a racial caricature, and thus is inherently linked to posts that deal in racialism).
The Happy Merchant, depicted in a different context
However, worth noting is that older templates, even if very thematically charged, had a broader use (e.g. the Happy Merchant, which despite being an anti-semitic caricature, is also used in posts that do not display explicit anti-semitism. In one case, it had been edited to depict the final boss from the videogame Fallout: New Vegas, as seen in image 11 above, using the template for other purposes than anti-semitism).
Trump Tweets Pepe
Criticism on Trump’s Pro-Israël Policy
Despite the role /pol/ claims to have had in Trump’s election, Trumpism did not seem to be that popular on the board anymore - at least when analysing /pol/'s images. This might denote that 4chan’s 2016-Trumpism was indeed ironic and not a result of sincere political affiliations. Not many pro-Trump posts were found, and those that had Trumpist imagery seemed to use it ironically. An interesting comparison can be made if one regards Image 12, taken from Donald Trump’s official account, which shows how Trump was closely tied to the 4chan meme Pepe the Frog, whereas Image 13, taken from /pol/ as part of this research displays a shift in the attitude towards the U.S. president. As Angela Nagle argues in Kill All Normies
(2017), the transgressive style of 4chan's anons hardly generates stable political allegiances, but rather enables fluid countercultural subcultures. The shifting Trumpism might be seen as an example of this.
Pepe was so closely associated to Trump and the alt-right that its creator “killed” it in a comic (Vincent, 2017), reverted to its pre-election, more general use, with edits of the cartoonish frog depicting celebrities, pop-cultural characters, users of the board, and politicians. On /pol/, it seems Pepe is, once more, a “reaction face”, an image used to portray the emotion the poster is feeling, instead of a monolithic artifact with specific connotations. In that context, it is possible to regard a character such as Pepe as a “floating signifier”, which is defined by Levi Strauss as a "semantic function whose role is to allow symbolic thought to operate despite the contradiction inherent in it", namely "a symbol in the pure state, apt to be charged with any symbolic content, a symbolic value zero" (qtd. in Mehlman, 15).
There has not, however, been a decline in the sentiments that led 4channers to support Trump in the first place. The board is still populated mostly by racist, misogynistic, anti-semitic and anti-immigration ideas. European and American users constantly tease each other over their country’s racial mixing and their failures to respond to “islamic invasion”, respectively, leading to the birth of a new meme: the “Amerimutt”, a racial caricature which represents the United States being “only” “56% white” (which is why it is often accompanied by captions such as “le 56% face”).
Despite not being pro-Trump, the undercurrent of anti-immigration is still quite prevalent, which can be seen in the instance of the Amerimutt. Although it might look like a self-reflective criticism by the American public, it actually started out as a criticism aimed at them by the Europeans, who originated the meme. Amerimutt could also be interpreted as an example of the “gatekeeping” mentality that 4chan assumes. By mocking the plebeian, anti-intellectual populace, channers attempt to keep out undesirable users that do not 'understand' their rampant racism and European purity. It is also worth noting that despite the large amount of American users share the Amerimutt, there is a heavy European influence in terms of Amerimutt’s popularity as well, as can be seen in image 14.
Amerimutt-images by countryflags associated to the post
Tactical meme-ing with Amerimutt
The tactical use of posting Amerimutt is geared towards raining this mixed-race aspect of the USA. Since it is a funny image, some 4channers see it as a tactical device for posting on more mainstream platforms to spread their hidden racism, hence creating a “tactical meme-ing” strategy (Image 15).
This exercise has provided useful insights into techniques and considerations for capturing and classifying memes. On the one hand, the identification and subsequent categorization of popular themes allowed for better understanding of /pol/’s discourse, showcasing clear threads of thought that could be related to the content used to convey specific themes. However, classifying memes came with its own challenges. The variety of uses and forms among some templates creates a lot of ambiguity when attempting to fit them into solid categories. This does not even mention the troublesome aspects of irony, gatekeeping and in-group humour that might set the use of an image apart from how it appears on first sight.
The landscape of /pol/ is, simultaneously, ever-changing and stagnant. While new memes appear, support for politicians change and use of memetic templates fall in and out of fashion, the core style of the board seem to remain the same since its inception: transgression through ever-present transcending of moral, ethical and societal boundaries. 4chan users have always been concerned about outsiders entering their community, changing its values and ideas, but, looking at /pol/, they do not seem to have reason to worry: 4chan has either created such an online “brand” that users who fundamentally disagree with the discourse presented there do not even attempt to become part of the website, or they have become become good enough at assimilating users that it can remains “fluidly stable”.
Hine, Gabriel, Onaolapo, Jeremiah, De Cristofaro, Emiliano, Kourtellis, Nicolas,
Leontiadis, Ilias, Samaras, Riginos, Stringhini, Gianluca, and Blackburn, Jeremy. "Kek,
Cucks, and God Emperor Trump: A Measurement Study of 4chan’s Politically Incorrect Forum and Its Effects on the Web" International AAAI Conference on Web and Social Media (2017): n. pag. Web. 18 Jan. 2018
Know Your Meme, “Amerimutt / Le 56% Face.” http://knowyourmeme.com/memes/amerimutt-le-56-face.
Accessed 18 Jan. 2018.
Know Your Meme. 2016. “Pepe the Frog” http://knowyourmeme.com/memes/pepe-thefrog
Mehlman, Jeffrey. “The ‘Floating Signifier’: From Levi-Strauss to Lacan.” Yale French Studies, no. 48, 1972, p. 15. CrossRef
Ohlheiser, Abby, 2018. “We Actually Elected a Meme as President”: How 4chan Celebrated Trump’s Victory - Chicago Tribune. http://www.chicagotribune.com/bluesky/technology/ct-meme-president-4chan-trump-wpbsi-20161112-story.html.
Accessed 18 Jan. 2018.
Phillips, Whitney. This Is Why We Can’t Have Nice Things: Mapping the Relationship between Online Trolling and Mainstream Culture. Mit Press, 2015.
Stryker, Cole. Epic Win For Anonymous. New York: Overlook Press, 2011. Print.
Urban Dictionary, 2018. “Normie.” Urban Dictionary, https://www.urbandictionary.com/define.php?term=Normie.
Accessed 18 Jan. 2018.
Vincent, James. “Pepe the Frog Is Officially Dead.” The Verge, 8 May 2017, https://www.theverge.com/2017/5/8/15577340/pepe-the-frog-is-dead-matt-furie.