On 27th March 2012, The Guardian published an article called “Twitter: Twitter racism: how the law is taking on the ‘Twacists'”, which discussed the growing rise of racist abuse towards young black male sports people and celebrities on Twitter. In this article, they described this emerging form of Twitter racism: “Twacism”, and the people that posted these abusive tweets: “Twacists”:
A spate of racist comments directed at Premiership footballers and broadcast on Twitter has prompted concerns about use of the social media network.
…two cases have come to court involving students who sent out offensive messages insulting the former Liverpool striker Stan Collymore and the Bolton Wanderers midfielder Fabrice Muamba, who collapsed during a recent FA Cup tie.
A variety of laws are being used by the Crown Prosecution Service to deal with offenders as police forces move swiftly to deal with high-profile cases. A number of footballers and celebrities are reported to have closed their accounts on the micro-blogging site after becoming targets of abuse.
Before attending a Downing Street summit last month tackling homophobia and racism in football, the culture secretary Jeremy Hunt said: “The behaviour of crowds has been that something that was socially acceptable 20 years ago is now socially unacceptable. Unfortunately, it seems still to be socially acceptable on Twitter.” The term “Twacism” – for “twitter racism” – has gained a limited currency online.
While my study of Twitter abuse on this blog has seemed to weigh more heavily towards racism on twitter, as this seems to be the most widely media reported form of people abusing each other on the Twitter platform – I thought this would be a good opportunity to see if this new descriptive term had the potential to either influence great discussion, general derision, or be a term that would eventually be used as a short-hand word to describe this specific form of Twitter abuse. To this end, I conducted a series of network analysis experiments during the height of publicity about the term to see who were the most vocal actors in discussion about it, and if, a month later this was a term that stuck around as a frame of cultural reference, or turned into something even more embedded in incidents of abuse on Twitter.
Methods & Discussion
On 27th March, I used Jeff Clark’s Tweet Spectrum platform to perform a quick analysis of the paired terms “twacism” and “twacists”. The image below shows that the term “twacism” was slightly more prevalent across the Twitter-sphere, with other key words, like ‘student’, muamba’, and ‘jailed’ mentioned in relation to the Liam Stacey vs. Fabrice Muamba court case and jailing of Liam Stacey, as well as other key words from the Guardian article:
As news of the article spread, I also used the Digital Method Initiative’s Google Scraper to scrape the top 100 results for the term “Twacism”, and retrieved an interesting mix of news reporting and blogging sites, all making reference to the term. The top site linked to and mentioned, was not surprisingly the original source of the The Guardian’s article:
I then used Google Scraper’s Tag Cloud Generator tool to create a Wordle that would better show how prminent The Guardian was as a source for the term “twacism” in the days surrounding when the article was published:
However, I also wanted to analyse the words surrounding the term “Twacism” in a way that was easy to for one to view at a glance. To do this, I used Tagxedo’s “Twitter Search” feature in it’s online word cloud platform:
This tool is a lot easier to style and shape into meaningful word infographics than Wordle, and enabled me to produce an interesting word/issue cloud, that showed key discursive words such as ‘offensive’, ‘racist’, ‘stupid’ ‘RT (retweet)’, ‘twacists’, ‘joke’ and the names of blog sites such as ‘Wendi Writes’ and ‘LondonDiva’ – were all associated with the term “Twacism” in the days around the publication of The Guardian’s article:
An interesting turn of events was how much the data scraped from the height of publicity for The Guardian’s “Twacism” article changed over time from a pretty balanced mix of retweets of the article with short-url links back to it – into a torrent of abusive and racist comments to a specific Twitter account called “@Twacism“, with the simple description “@twacism
Racism on Twitter”. A browse of this Twitter accounts’ feed shows a growing collection of racist tweets from across a broad spectrum of people – with abusive tweets coming from many different races, and all tweets consisting of racist abuse towards other races. The owner of this Twitter account is anonymous, and one is unable to establish whether this is a cynical collection/observation of racism on Twitter, an covert research project, or someone seeing how far they can go in encouraging Twitter users to post racist tweets.
Because this source has now replaced The Guardian article as the dominant actor in the web sphere, data outputs have also changed. A Twitter scrape via Twendz using the term “Twacism” shows this change quite clearly. Below is the Twitter discussion of media coverage from 27th March 2012, where the most popular words associated with “Twacism” were words like “commie”, “hell”, and “pipedream” (in relation to the then controversy over President Barack Obama’s healthcare plans), and then “racist” and “word”. The scrape also shows that on 27th March, the sentiment around the terms “Twacism” was fairly neutral at 57%, while 17% negative and 27% positive:
Contrast this with a Twendz scrape for “Twacism” on 23rd April 2012. Key words associated with “Twacism” are now words found in abusive tweets, such as “Asians”, “Hate”, “N*ggers”, “People” and “Sh*t”. WHile the word cloud has particular emphasis on the largest visible words “hate” and “people”. Sentiment for the term is now 43% negative and 56% neutral, with zero percentage of positive sentiment:
A more visual representation of this new trend can be shown using the relatively new tool “Spot” by Jeff Clark. A quick search for the term “twacism” brings up a cluster of negative words, the majority coming from the racist tweets that are been tweeted to or retweeted from the Twitter account @twacism. Again, most prominent here is the word “Hate”, and then follows racist terms for different racial groups:
According to Jaewong and Leskovec, online content exhibits rich temporal dynamics, and diverse realtime user generated content further intensiﬁes this process. My small study above supports this view by showing that content on micro-blogging platforms like Twitter, is very volatile, and that pieces of content become popular and fade away in a matter of hours. Like the terms “Twacists” and “Twacism”, Yang and Leskovec argue that short quoted textual phrases (“memes”) rise and decay on a temporal scale of days, and represent the integral part of the “news cycle”. Yang also argues that because of this, uncovering patterns of temporal variation on the Web is difﬁcult because human behavior behind the temporal variation is highly unpredictable. This makes the overall picture of temporal activity on the Web even more complex due to the interactions between individuals, small groups, and corporations. My small study has shown that while mainstream media and bloggers are both producing and pushing new content into the system, one cannot always determine how that content then gets adopted through personal social networks and altered as it diffuses through the Web. In the future, I would like to use models like Yang and Leskovec’s K-Spectral Centroid (K-SC) clustering algorithm to analyse the life of a meme in more detail, but for now it has been interesting (and from a human level a little saddening) to follow the rise and change of the terms “Twacism” and “Twacists” from coinage to their use as a focus for general abuse.