Twitter, the July Days and the Arab Spring

Crowds in revolutions are fascinating for the historian as we try to unpick and explain to our students how the “Paris Mob” operated so the tweets from the revolutions of the “Arab Spring” are not only a wonderful document of a struggle for democracy, but also a case study in revolution.

I am frustrated by standard works on the French Revolution which often reduce the action in Paris in July 1789 to the Fall of the Bastille, as if this dramatic turning point emerged from nothing and changed everything. My students are not up to reading primary documents in French, but I can send them back to works like Mignet’s ‘History of the Revolution’, which is freely available on the web and for which he had the advantage of access to many of the participants. Reading Mignet, it is clear that Paris in the July Days before the fall of the Bastille was a city in turmoil, with crowds, fueled by fear, rumor, and street orators like Desmoulins, ebbing and flowing while the loyalty of the French Guards was in question and the cavalry of the Prince de Lambsec struggled to maintain order in a city that was increasingly hungry. From a few pages of Mignet, it is clear that while the Fall of the Bastille was a turning point, it was part of a process which modern textbooks all to often reduce to the simple, often repeated and wrong idea that the mob stormed the Bastille to gain arms.

How then will the history of events in Tunisia, Cairo, Libya and Syria be written. How will the writers of textbooks reduce it all to one sentence?

From following the events as they unfolded on twitter and in the ‘real media’, certain things emerge.

The character of the tweets differs between city and desert. In the Tunisia and Egypt, much of the events of the revolutions unfolded in urban settings, in some cases with ‘real media’ present alongside the popular voice coming through twitter. In a revolutionary crowd, there will often be a limited core of activists and a larger group of followers who have more time to tweet. In a city, there are more witnesses to the movements of activists, police and army, and action focuses around key locations – Presidential residences, TV stations, police and government buildings and choke-points. Events, real or assumed, around key places or people generate twitter traffic.

In some respects, all cities are one. It is an exaggeration to assert that the world is urban, but there is a global city. Thus, as the crisis unfolded in Egypt, the movements of diplomats and leaders in other regional capitals became part of the twitter stream. Leaders moving to meet were presumed to be consulting on the crisis. A story that the Egyptian ambassador in Riyadh was en route to the Airport there was retweeted as solid evidence that Mubarak was in the process of fleeing to Saudi Arabia, a story which proved to be false but which helped to keep the fires lit for a few hours.

On the other side of the coin, no one tweets from a desert firefight. In Libya, as the struggle changed from an urban revolution in Benghazi to a civil war in along the coastal highway and in the mountains, the tenor of the tweets changed. Whereas the tweets in Cairo were from the frontline, in Libya they were increasingly from the rear, from people under siege or under heavy weapons fire from Gaddafi forces in towns and cities, but little or nothing from the front line fighters.  It was therefore harder to get any sense of what was happening from hour to hour, although NATO was able to use some twitter information as a feed into their targeting of Gaddafi’s heavy weapons. As the war assumed a slower pace, it was possible to track the ebb and flow west along the coast from the placenames that came up in tweets listing casualties.

Another major impact of social media was the channel it provided to the world. Up to the end of the Cold War, news from dictatorships only came out when people managed to get out to west and find people who were willing to listen to their stories – news traveled by personal, physical contact. Even as recently as the Rwandan genocide, early news of what was happening only got out to the world through often harrowing phone calls from people about to become victims. here, again, news traveled from one person to another, and without that first link, the story could not get out. Advances in broadcast media meant that revolutions from ’89 onwards could catch world attention as long as the TV cameras were willing to cover the story. This meant that there was potential for wider coverage, but it flowed through large media organisations and was affected by their scheduling priorities.  But by the Spring of 2011, people around the world could follow the tweets from the revolution as they appeared. At least in theory, this meant that voters in the West could be influenced to support the democratic revolutions and might bring pressure to bear on elected politicians to support intervention. Large scale networks, at once personal and also public, could be built and maintained. Exploring how this worked in practice will provide interesting questions for historians and political scientists in years to come.

An interesting question is how this affected the old regimes –  while social media may well have contributed to the downfall of the Ben Ali government in Tunisia, who caved in relatively quickly, it seems no one in the Presidential Palace in Cairo was paying much attention to the “twittersphere”, and as long as the army supported the regime, nothing in the virtual world would make much difference. In Libya, the Gaddafi government has made some efforts to put it’s message out through social media, although it has been limited, obvious and not successful. How governments are influenced by the electronic ‘salon of the barricades’ in the future will be an important question not only for students of revolutions, but for the ordinary political process.

Getting from a huge, unsorted archive of tweets from the Arab Spring to an analytically useful end product is an interesting research process, and useful for teaching. The sheer size of the material makes it almost unmanageable. Much of the interesting work on the tweets focuses on mapping them with Google Maps mashups. There are also a range of tools out there to visualise the twitterstream, all of which are interesting and may draw attention to trends with exploring. However, for serious historical or political science work you need to get a collection of tweets down in plain text to slice and dice, and this is harder.

There are several ways to search the twitter archive, and some to search and export tweets. Most are limited in how far back or how many tweets they can dig out – 3-4 weeks or 1,500-3,000 tweets are common enough limits. This means that if you didn’t get the material while it was happening, you may have a problem. Google realtime search provides some handy search filters, but no easy export for the results. Searchtastic and Twapperkeeper have good search options – Twapperkeeper is better for finding older tweets on a topic, but it’s export feature is disabled. Searchtastic has more limited searching, but will export. If you are doing serious research, I imagine Twitter would facilitate you by providing decent access to their archive. However, before you download a million tweets, you might want to work out methodologies for attacking this archive.

Once you export a set of tweets, the magnitude of imposing some analytic framework becomes clear. Even a modest search and export will produce more rows than most spreadsheets can handle – I broke the row limit on OpenOffice Calc and Excel easily. All of the search, visualisation and export tools will grab the timestamp, and many the location of the tweeter from their profile, but none of the available export tools will pull out the full metadata attached to a tweet which means the geolocation data for a particular tweet, if it is enabled, is not easily accessible. While the available search and export tools are good, the full content of a “tweet” is much more complex, and serious research will need the whole tweet.

Some analysis is simple – it is easy to map who is following whom, and which tweets are retweeted. Items which are retweeted without editing can be tracked to show how a tweet moves from place to place. the use of hashtags is alomst useless – some relevant tweets use none, while some are almost indiscriminately tagged multiple hashtags  –  #Libya #Egypt #Yemen #Syria is not uncommon for example – and this means reading and manually coding the tweet. Developing ontologies to organise the archive and track memes as they emerge and flow will be interesting, but the Twitter Annotations feature will allow an XML compliant framework into which researchers might be able to hook coding schemes to allow for analysis.

Old style research manuals in the humanities say you should note all your data on 3″x5″ index cards, and sort them into piles to see what emerges. In the end, historians of the digital element of the revolutions of 2011 will be sorting a lot of virtual index cards into piles for many years to come.

One thought on “Twitter, the July Days and the Arab Spring”

Leave a Reply

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