Word “fag” has undergone a series of semantic shifts from its first recorded usage in 1921, when it appeared as a shortened version of the word “faggot”. Through history, the longer word has referred to bundles of sticks, bonfires for burning witches, people that made living by gathering wood, old women and homosexuals.
Etymology of the word was popularized by The F Word episode of South Park, where the word’s meaning was provocatively relativized in spite of all emotional significance attached to it and the word was forced into a neutral semantic space. In fun way, Trey Parker, the writer of the episode, managed to cleanse the word of its homophobic discourse leftovers by redefining the word once again. In the episode, redefinition of word “fag” was done in formal fashion, by changing the word’s meaning in Webster’s Dictionary, because it was acknowledged that popular usage of the word was different from it’s official meaning.
South Park has criticized political correctness (PC) in similar fashion from its beginning, but perhaps it did so most directly in episode With Apologies to Jesse Jackson, in which a new racial slur for white people – “the nigger guy” – was introduced and then swiftly banned. In episode Chef Goes Nanners, instead of language, the problem was racist visual symbolism in the town flag, which depicted a mob lynching a black man. Solution to that problem was to make a lynch mob multi-ethnic. The entire 19th season of the show deals with PC issue by exposing it yet again as rigid ideology which values form over function. This time PC served as an instrument of terror placed in the hands of authority and frat house bullies. Making fun of strict rules of systems of language, law and belief by enforcing them in an unexpected context is one of the foundations of satirical humor driving this show.
After South Park was done with it, label “fag” could be applied to all kinds of obnoxious and inconsiderate people, at least in the show. Still, the the obligatory irony left bikers stigmatized as the new official fags. In the show, the new definition of “fag” is:
1. An extremely annoying, inconsiderate person most commonly associated with Harley riders.
2. A loud and obnoxious person who owns or frequently rides a Harley.
Without a group to be offended, a key element of political correctness, the satire wouldn’t work that well. Even though a specific group was left to bear the burden of faggotry, implied message is that regardless of their sexual preference, people can still be fags by choice. Being simultaneously progressive and offensive in this way is typical of South Park. Bikers were chosen probably because the first season of successful show Sons of Anarchy was nearing its finale and this served as parody of the show too. That and the fact that bikers are pretty gay.
Now let’s take a look at present day usage of the word in a specific online community where it can signify diverse groups of people. Even though the word keeps changing its meaning, old fags and new fags are all still fags on 4chan.
Fags of 4chan
4chan imageboard was launched in 2003, and pretty soon it gained ill reputation of consisting of inconsiderate people, being extremely chaotic and filled with all kinds of depravity. Mischievous spirit of its users and their love of the art of trolling were bound to be misinterpreted by mainstream media, which resulted in lots of goofy news reports. In media, 4chan was called Internet Hate Machine and it has even been mistaken for a single person (this 4chan person was recently v& and his real name is revealed to be Ryan Collins).
The site’s notoriety naturally attracted people who felt like social outcasts, edgy teens, and people who were intrigued by the content that this userbase posted on the imageboard. Starting point for this type of newcomers ended up being /b/ – Random (formerly Anime/Random, but its name was changed in November of 2003) board. Pretty soon it was clear that the site had potential to be more than an imageboard dedicated to discussion about anime.
Of course, people collectively labeled “Internet Hate Machine” must have been calling each other fags in everyday communication and in other online forums and boards prior to coming to 4chan. Level of their concern with politically correct language and mainstream interests can be easily inferred from the fact that initial population of 4chan consisted of users from “Anime Death Tentacle Rape Whorehouse” sub-forum of The Something Awful Forums, “the closest thing” it had “to a concentration camp”. If that isn’t convincing enough, here’s how moot, the former owner (replaced by Hiroyuki Nishimura, a.k.a. Chinese moot, on September 21, 2015) of 4chan announced registration of 4chan.net (top-level domain was changed to “org” in 2004) in a IRC channel:
<moot> regging 4CHAN.net
<moot> FOUR CHAN
<moot> brace for faggotry
These prophetic words are now directed at Google.
Anonymous mode of communication made it easy for people to insult fellow users due to so-called online disinhibition effect, and in turn, hard to take insults very seriously. Once rudeness has become the norm, even newcomers who were reserved about using politically incorrect terms had to accept them as part of the culture of the community. For example, replying to the original post in a thread with message “OP is a fag” is a site-wide meme and even if users wouldn’t call someone they disagree with a homophobic slur in any other context, they might do it on 4chan in order to express disagreement with original poster (OP), who is, as saying goes, always a fag. Anonymous mode of communication seems to encourage creation and usage of memes, since for people to stay anonymous, they need to act in similar fashion. Thus similar or often repeated situations tend to yield expected results – copypasta threads are answered with corresponding copypasta, OP that doesn’t deserve a proper response is simply called a fag, newfags are advised to “lurk moar”, sexual-deviant-flavor-of-the-day threads are bombed with pictures of Spiderman, and so on. Some memes are board-specific, local, and can serve as blockades for undesired or low-quality content by rendering the whole thread useless with what appear to be inane comments and discouraging OP from further engagement in the discussion. Calling OP a fag might have started out because copypasta was funny, but now it is also efficient. Ends justify the memes.
It’s worth mentioning that being a fag on 4chan isn’t necessarily a bad thing and users often call themselves a certain type of fag. In this context, the meaning of word “fag” is closer to colloquially “being gay for something”. Even though Mac users probably weren’t the ones who came up with “macfag” coinage, they might identify themselves as such in appropriate context. It seems that in a friendly discussion users are more likely to call themselves fags, while in hostile situations they will probably use the word to describe others. This process of reappropriation might even be the original usage of a word. Diversity of fag-words on 4chan is so vast, many of them might have been created by the very people they are meant to signify.
OK, this is nothing new. Everybody knows everyone is some kind of fag on 4chan. But let’s see what we can find out if we take these words seriously and map how often they are mentioned in different boards. I think they can be very useful in analysis of cultures of individual boards.
For gathering data, I wrote a Python script that makes a series of (decently delayed) queries to 4chan API and then defined a cron job that ran the script every 2 hours. The script is able to scrape new threads as well as old ones that have been modified since the last run of the program. Then it adds fresh comments in a database, divided into boards. Another script uses regular expressions to extract words ending with “fag(s)” from all comments that were stored in the database from a specified board. Comments without text, consisting only of images, were ignored by the latter script, since meaning of images couldn’t be automatically interpreted properly in this analysis.
I could’ve used data from an existing archive (shout out to Internet Archive!), but I wanted to have current data. I limited my analysis to a specific time period between February 24 and March 9 of 2016. I also limited my research to only five boards. Some of the boards update slower than others and some argue and use fag-words when arguing more often, but I decided to collect data from each of them one by one at the same pace. My choice of boards was the following:
- /b/ – Random – the most popular board, very active and filled with memes and bad language. This is the point of entry into 4chan community for most users, so it was guaranteed to be filled with newfags forming into more specialized kinds of fags. It is like a high school for fags in that sense, although you must be over 18 to post in it.
- /a/ – Anime & Manga – manga, anime and Japanese culture in general are important parts of 4chan culture that can be traced to the image board’s roots. I picked this board because it was fast and because fandom guarantees arguments, so I expected to harvest lots of fags from it in every run of the script.
- /g/ – Technology – “Flame War” – The Board. I knew users on /g/ took sides in discussions about operating systems, mobile phones, graphics cards, etc. For some people technology is a passion leading to previously unexplored levels of faggotry.
- /pol/ – Politically Incorrect – mandatory board for scraping inconsiderate comments and foul language. This one seemed like the perfect fag mine.
- /lgbt/ – Lesbian, Gay, Bisexual, & Transgender – I acted on a queer hunch that I would be able to find some fags on this board. I was also interested in how LGBT population used fag-words on 4chan.
Brace for faggotry
At first glance at the results, I was struck by the sheer variety of fag-words, ranging from well known “oldfag” and “newfag” – used for expressing one’s status in terms of how long the user has been a part of 4chan culture, to “tripfag” – reserved for posters with unique user identifiers called tripcodes, “poorfag” – often used insult based on one’s financial situation, to “normalfag” – a common term for “normal” people, which often serves as 4chan’s version of gaijin, and “samefag” – someone who pretends to be multiple different users in the same thread, to very specific and often downright esoteric neologisms such as “tarotfag”, “humanfag”, “betavirginjealousfag”, “newnamefagwhorecallsothersnewfag”, etc.
Even though every board is full of those strange one-time fags, each board has its own set of rules and values that can be seen in dominant fag-words on the board. 4chan users make it very easy for us to see whom they are calling a fag and why – they simply state it in the root word followed by the “fag” suffix! This is very convenient for researching trends and kinds of discussions that are going on in a given period and board. This might be treated as instant sentiment analysis because it is apparent from the data whom people are arguing with and about what, without needing to parse individual comments any further or trace relationships between comments. Sure, calling someone a fag might not be necessarily considered an insult on 4chan, or at least not a serious one, but a fag-word can still be used as a vector in conversation that’s going on in a given thread. So I used “fag” suffix to locate words from which I could extract root words I needed for detecting opinions. Even in cases when users apply the label to themselves, it is usually there because it is related to the topic of discussion.
But what about plain old “fags”? Since it has been established that on 4chan word “fag” generally doesn’t refer to a homosexual, the word standing on its own, although still looking offensive to outsiders, can be treated simply as a sign of heated discussion or disagreement. On the other hand, the word is definitely not completely cleansed from its homophobic roots and usage in gender policing, so detecting a genuinely homophobic context is impossible in my crude approach. Even discussion about homosexuality is difficult to detect, unless it happens as explicit conflict between gayfags and straightfags (which happens).
I was primarily interested in fag as a suffix, and you’ll notice that I didn’t include the longer version of the word, “faggot“, in my analysis either. I left it out because it is rarely a part of the compound words I was looking for. In the end I treated frequency of occurrences of standalone word “fag” (let’s call these “vanilla fags“) in a board as an indicator of general hostility of the board. Besides, constant presence of homophobic threads reflects the general hostility of the board too. (See how easy I got out of this one?)
Basically, I treated word “fag” standing on its own as an insult without specified direction. Perhaps those attacks were not necessarily based on personal preferences and specific type of fag that was insulted. This reasoning, however, works only in this case and you are hereby warned that avoiding ad hominem arguments by calling someone a fag doesn’t work that great in real life.
Here’s the summary of data per board:
- /a/ – scraped 517,417 posts in 7,674 threads.
- 11,709 fags in 10,329 posts.
- 1 fag in 44.19 posts.
- 1,410 variations of fags.
- 8.30 fags per variation.
- 2,095 (17.89%) vanilla fags.
- /b/ – scraped 938,774 posts in 17,663 threads.
- 25,818 fags in 13,679 posts.
- 1 fag in 36.36 posts.
- 758 variations of fags.
- 34.06 fags per variation.
- 17,998 (69.71%) vanilla fags.
- /g/ – scraped 232,981 posts in 6,455 threads.
- 3,030 fags in 2,841 posts.
- 1 fag in 76.89 posts.
- 296 variations of fags.
- 10.24 fags per variation.
- 876 (28.91%) vanilla fags.
- /pol/ – scraped 978,869 posts 19,835 threads.
- 11,038 fags in 10,267 posts.
- 1 fag in 88.68 posts.
- 797 variations of fags.
- 13.85 fags per variation.
- 5,072 (45.95%) vanilla fags.
- /lgbt/ – scraped 77,272 posts in 1,167 threads.
- 764 fags in 682 posts.
- 1 fag in 101.14 posts.
- 84 variations of fags.
- 9.09 fags per variation.
- 414 (54.19%) vanilla fags.
Boards sorted by variety of fag-words: /a/, /lgbt/, /g/, pol/, /b/.
Boards sorted by frequency of fag-words: /b/, /a/, /g/, /pol/, /lgbt/.
Boards sorted by percentage of vanilla fags: /b/, /lgbt/, /pol/, /g/, /a/.
Earlier I suggested that amount of vanilla fags might be used for measuring general hostility of a board. Topic of /lgbt/ board certainly makes things complicated here because of reappropriation of standalone word “fag” in gay culture. However, if we leave out /lgbt/ from the list of boards sorted by percentage of vanilla fags, as it needs closer inspection, the order of boards seems to reflect actual hostility of boards. General hostility might be inherent in the board’s topic or it could be a sign of very broad definition of the topic combined with loose rules and moderation. The latter allows groups with very different interests to inhabit the same board which could lead to fights between groups of users for limited board space. Comparing general hostility of a board before and after a new containment board (I’ll explain those in next section) is introduced might be a way to test my hypothesis that vanilla fags mean chaotic bickering. Anyway, I won’t deal with unspecified hostility and vanilla fags in the rest of this text, as it is not my main focus here.
Fags on individual boards
I inspected faggotry, made short analysis and visualization for each board. I omitted vanilla fags from charts because they are very dominant in every board. Their share will be mentioned again as a reminder, but we will focus on neologisms with “fag” postfix.
Let’s start with /b/ – Random, the infamous entry point to 4chan culture.
The number of fag variations on /b/ is not as high as you’d expect from a very dynamic board with all kinds of users posting random content. 69.71% of fags found on /b/ are vanilla fags, so they contribute most to the highest frequency of fag-word usage that this board holds. This can be explained by users migrating from this fag limbo to boards dedicated to topics that interests them. Whenever it becomes apparent that a significant number of users starts producing specific content all over the imageboard and many threads start getting railroaded into discussions about the same topic, a so-called containment board is created with goal of attracting all users interested in that topic. It is a common way of expanding online communities, but as the name suggests, containment boards are made out of necessity and for the sake of users who don’t display symptoms for which they would need to be placed in a quarantine. The lack of a well-defined topic seems to be the main reason for low fag variety on /b/.
Besides being general purpose and thus not very interesting for users interested in specific topics, /b/ is also full of new users getting acquainted with the imageboard and trying to prove themselves as active members of the community. This is why the biggest issue on /b/ by far is being new. “Newfag”, a reference to a newbie, takes 10.56% of all fag-words, with “samefag” following it with only 2.28% of all fag-words. After “samefag” (a concept especially intriguing to those new to anonymous mode of communication) comes “oldfag”, the complete opposite of “newfag”, with 2.09%. Other top fag-words signify populations often found on /b/, like people starting drawing threads, Americans and Europeans engaging in flame wars, furries, Boxxy fans, poor people, etc. “9fag” actually refers to 9gag, a meme aggregator website and online community regarded by 4chan population as terminally cancerous.
Now let’s take a look at /pol/ – Politically Incorrect, “a quarantine board disguised as a politics board“
After 45.95% of vanilla fags and 5.96% of newfags, a trendy topic of current presidential election is apparent – 5.70% of fags are trumpfags and 3.67% of fags are berniefags. Trump’s fags seem to be mentioned more often, even if we include other variations of berniefags, such as “bernfag” (0.88%) and “sandersfag” (0.27%) into Bernie’s fags. Almost nobody on 4chan cares about Hillary enough to be her fag (0.08% are hillaryfags). The elections were obviously the hottest topic in this time period here.
After these two, stormfags are next in line with 3.52%, then poorfags with 2.47% and christfags with 1.86%. “Chirstfag” obviously refers to a Christian and “poorfag” is a textbook example of argumentum ad crumenam – argument to the purse, commonly used in all kinds of discussions.
As for “stormfag”, its meaning is, perhaps ironically, closest to the original meaning of word “faggot” and “fag” as its shortening. Stormfag is a person that is a member of Stormfont community or that propagates ideological views similar to those held by its members – fascism, neo-nazism and white supremacy. Expanding on briefly mentioned etymology of the word at the beginning of the text, the origins of English term “faggot” can be found in Old French “fagot”, which in turn most likely originates from Italian “faggotto”, diminutive of Vulgar Latin “facus” and the original Latin word “fascis”, meaning “bundle of wood”. The bundle of sticks symbolizes strength through unity and it was a well known symbol in ancient Rome as well as in 19th century Italy where political groups were called “fasci” – or “fags”, if we allow ourselves an atemporal point of view – a number of which evolved into the fascist movement in 20th century. With this meaning in mind, fascism was the original faggotry and it looks like /pol/ is right by using the term in two historical contexts simultaneously.
The next board is /g/ – Technology – containment board for guro (NSFW) content that somehow ended up being worksafe board about technology.
Here, 28.91% of vanilla fags is followed by 16.47% of poorfags. The drop in the usage of vanilla fags is clear, but it is a slow board which means I collected less data from it. Poorfags are the main problem of /g/, with only a few mentions of richfags (0.63%), usually used in self-identification, in this period.
This doesn’t mean there’s a real struggle between poor and rich on /g/. However, here is definitely a struggle going on between poorfags and richfags on /g/. We can only speculate about economic status of some users based on boasting and photographs of technology they own, but what’s more important is the fact that there’s a constant poorfag vs richfag discourse going on in most threads on the board. This probably shows that /g/ is more about consumer electronics than general technology, since poorfag argument is most useful in discussions about gadgets, software and lifestyle choices these products are supposedly reflecting. One of the perfect examples of this connection of a product and calling someone poor is on the chart – “amdpoorfag” (0.56%).
After newfags with 8.65% share, macfags, users of Apple products, consist of 5.64% of all fags. “Applefag” (0.96%) and “ifag” (0.92%) could be added to macfags, boosting it to 10.52% of all fags, which makes them more frequent and thus more important than newfags to population of the board. Although it would require additional research, I think it is safe to assume that a large number of “poorfags” comes from Mac users defending their consumer choice with their wallets. A battle for encryption between FBI and Apple might have added to numbers of both “macfag” and “poorfag” words during this time.
/g/ is also filled with tripfags (5.05%) and samefags (3.63%), two opposites standing next to each other, one representing vanity and other a dirty tactic for creating an illusion of popular opinion. After them, 2.84% of fags are normalfags intruding into a realm of geeks, 1.95% are Windows users – winfags, followed by oldfags (1.52%), Linux users – linuxfags (1.02%), etc. Every major operating system is present in the list of top 20 fags.
The most creative neologisms came from /a/ – Anime & Manga board.
Scraping /a/ returned an enormous amount of posts containing word “fag”. Even from my relatively small sample, it looks as if this board has a fag for every character from every anime ever made. Since the nature of the board’s topic is to discuss and comment on users’ individual tastes (normalfag talk for waifu/husbando claiming), this might mean that a lot of fag-words is used in self-labeling and expressing one’s own taste. If that’s the case, it would support the board’s low overall hostility as I defined it. The opposite of this board would be something like /b/ – a high percentage of vanilla fags and a low number of fag variations. In this case, a high number of variations compared to a low percentage of vanilla fags is a testament of the board’s diversity of opinions that fit relatively nicely together. This certainly seems true when compared to /pol/, where low diversity of fag variations allows us to immediately spot one big debate (or a flame war) between supporters of Bernie Sanders and those of Donald Trump.
Here, vanilla fags are 17.89% of all fags, and after them the biggest issue seems to be that of newfags with 8.16% and after that the problem of normalfags with 5.38%. It seems appropriate to see this combination of dominant fags as outsider influence. Newfags are obvious intruders, and, as I said before, normalfags are treated as gaijin – another borrowing from closed Japanese culture by anime fans. Normalfags are regarded as “plebs” or commoners, with bad taste in TV shows and 3D women. Subculture of anime fans seems to stick together.
Next are drawfags (3.97%) and tripfags (2.34%), with possibility of some overlapping, because communicating with a drawfag is easier when the artist is able to stand out from the rest of the users. After them is similar percentage of waifufags (1.82%), people involved in romantic relationships with anime characters, and samefags (1.76%), followed by buyfags (1.14%), those who actually buy anime and toys, and almost the same shares of shipperfags (1.09%), fans with strong desire for fictional characters to be in a relationship, and yurifags (1.08%), fans of yuri anime and manga genre that portrays homosexual relations between women. Besides all-present poorfags, oldfags, moralfags, etc, the rest are all based on the fag’s favorite character or interest.
Finally, let’s examine the case of /lgbt/ – Lesbian, Gay, Bisexual & Transgender board added in March of 2013.
This is by far the slowest board I archived. In small sample I scraped, 1 fag was mentioned every 101.14 posts, with more than half of fags (54.19%) being vanilla fags. The board yielded a low number of fags per variation (9.09), meaning a high diversity of fags in the sample. Larger data is required for more reliable analysis, but let’s see what we can learn from my data. After the majority of plain old fags, there are tripfags with 13.87%. Next are the common newfags with 4.84%, followed by namefags (3.93%), oldfags (2.09%), samefags (1.31%) and then avatarfags (1.18%).
“Namefag” refers to an user posting under a name other than Anonymous and “avatarfag” to an user that often uploads pictures of the same character when posting, de-anonymizing him/herself as a result. Both are similar in function to tripfags, users identifying themselves with a trip code. Tripfags, namefags and avatarfags together make 18.98% of all fags in the board and reflect a big problem of user identity in the community. This, combined with very slow pace of the board, suggests that a number of users from already small userbase, often talking about intimate topics related to their sexuality, has a need to transcend anonymous mode of communication, probably for practical reasons. A few regular visitors can easily dominate a slow board and are likely to recognize each other anyway.
Amongst top fag-words, there are also those connected to LGBT issues, like “transfag” and “gayfag” with 0.92% share each, “bifag” with 0.78%, femfag with 0.65% and yurifag with 0.26%. Together, these words take up 2.61% of all fag-words on the board, but, as I stated earlier, we can’t really separate “neutral” vanilla fags from “homophobic” ones without further analysis. If we could, we would be able to add the percentage of homophobic ones to LGBT topics. Detecting reappropriation of homophobic words in order to see how many of them is used maliciously would be even harder and would require a deeper look at context.
You might be wondering whether fag-word data can be used in a practical way or not. I suggest that it can.
Meme tracking in general can be useful in shitposting (low quality posts) management, flaming detection and identification of long-term trends and hot topics. All this is useful to board administration, advertisers, propagandists, trolls, social scientists, etc.
Using this approach, we can also find overlapping topics (or at least sentiments about certain topics), quite possibly discussed by the same populations of users. After all, boards are designed to contain topics and not necessarily users who discuss them. Sudden change of the use of language in a board can mean a raid by users from another board and gradual change can mean migration caused by changes of rules or dominant topics. Tracking fags through history of 4chan would surely reflect such movements of populations of users. For example, when a new board is opened, permanent migration of users interested in the board-related topics could be reflected in the data, and we could measure effects of the change in the imageboard’s structure by comparing the frequencies of specific fags in old and new boards. Luckily, there are 4chan archives available so I might track a specific migration or a raid some time in the future.
Of course, fag-words are only an example of this type of research. We can compare it to similar tracking of word “cuck”, a popular insult on 4chan that is almost never used for self-labeling, especially in combination with other words. Similar research could be done with images, by looking for their hash values.
But this is enough from me for now. I hope to see more studies of faggotry and other phenomena in the future. I know I’ll be combing through archives.