Image from New York Times article

The Narrative of Terrorism in #Charleston

A domestic terrorist murdered 9 African Americans who were worshipping at the Emanuel AME Church in Charleston, South Carolina on June 17th, 2015. The racially motivated murders have been another cold reminder of racism’s constant presence in the United States and the destructive and traumatic results for black bodies.

I began following the news of the attack on Twitter, where as usual, most of the early information and reactions were being shared. As someone who has been closely following the uproar over the recent police killings of African Americans, this terrorist attack felt like a dramatic escalation of the dehumanization of black life that so many have tried to bring attention to over the past year.

From the moment I began to see the hashtag #CharlestonShooting being used, along with people’s stunned reactions, I felt it needed to be labeled an act terrorism.

There were several reasons I closely followed the events on Twitter. Of course I wanted to interact with others and share my own feelings about it. Twitter was also a comforting space in many ways, as others shared how this news was affecting them. I could tell early on that this latest attack felt very personal to many, especially African Americans, because we live with the constant awareness that racist violence against us can happen at any moment. These two tweets by @jmddrake were especially powerful for me.

But as an archivist and someone interested in documenting how African Americans build community and share experiences about these types of events in digital spaces, I was also interested in seeing how quickly reactions would grow, what would be said, and if that would affect when and how major media outlets picked up the story. I was particularly interested to see if the narrative of terrorism would take hold in relation to this racist act of violence.

As far as I could tell that evening there was no coverage on the 3 major networks (CNN, FOX, MSNBC) for over two hours after some of the first reports showed up on Twitter of an attack and possibly several casualties. I wasn’t the only one who noticed the blatant disinterest by the networks. It’s pretty much standard now for people of color to have to legitimize their own trauma on social media before major news networks give their stories any significant airtime.

In an attempt to capture some of the reactions and answer some of these questions, Ed Summers and I started collecting tweets tagged with #Charleston and #CharlestonShooting. We’re both interested in archiving social media and the potential for it to add a rich new layer to the work we do as archivists. After some very early analysis we wanted to share just some of what we are finding in the tweets. There are some important issues to consider for archivists as we continue to think about building collections and documenting historical events in this space.

As of 12PM ET on June 19th we had collected 2,075,608 tweets mentioning #CharlestonShooting since 9PM ET on June 17th. Here is a graph of tweets per hour so you get a sense of just how rapidly folks were sharing information about this event. Just nineteen minutes after the police responded to calls reporting the shooting we see the first #CharlestonShooting tweet:

The first tweet mentioning the word “terror” in conjunction with #CharlestonShooting came 9 minutes later from Wisconsin:

But note that the context for the tweet doesn’t really reflect the dialogue that emerged later. The first tweet mentioning terror that was retweeted came 11 minutes later from Florida:

To give you an idea of how the media eventually picked up on this theme of terrorism here is a list of some of the most retweeted tweets mentioning terror that were sent in the first three hours after the shooting.

Ed wrote a program to search through the tweets in chronological order to find the first one that pointed to a Web page that mentioned the word “terror” in it. It found this tweet from 10:43PM ET (just over an hour after the police were called in) which points at a Reuters story on Yahoo News byHarriet McLeod, a journalist based in Charleston, South Carolina:

But it’s important to remember that this analysis depends on the state of the Web page as it appeared at 3PM ET on June 19th, which is when our program was run, and was 28 hours after the tweet was sent.

Internet Archive has a snapshot of the same URL at 3:17AM ET on June 18th that has no mention of “terror” in it. In fact the content of the article is entirely different, with different headlines:

Gunman at large after killing nine at black South Carolina church.


White suspect arrested in killing of nine at black U.S. church.

Compare the first three paragraphs of the Internet Archive snapshot:

Text of the article at 3AM ET June 18, 2015

with the first three paragraphs just over a day later:

Text of the article at 3PM ET on June 19, 2015

This highlights just how fluid the Web is as a medium, and how important Web archives are for understanding the context of what transpires in social media spaces like Twitter. Usually archivists make the case for Web archives by drawing attention to rampant (and very real) problems with link rot. But as researchers such as Zittrain and Van de Sompel are noting, reference rot is also a huge and important issue. In this case what appears to be syndicated content from Reuters at Yahoo News, at a single URL, has radically changed in a very short period of time.

So exploring how this narrative of domestic terrorism emerged in social media and on the Web turns out to be kind of tricky. This use case represents a challenge for the archival community to build robust, high fidelity and (most importantly) timely Web archives for study. We’re hopeful small experiments like this Medium post will help inform the work we are planning to do on a tool that allows archivists and scholars to tune into social media events (like I did that night) and collect selected Web content. If we don’t do it who will?

One clap, two clap, three clap, forty?

By clapping more or less, you can signal to us which stories really stand out.