Riots and Tigers and Memes, oh my!
For a few hours in August 2011, a significant proportion of the UK population believed that tigers were roaming the streets of London. This led me to ask how rumours get started, what makes a rumour believable, and why is it so difficult to refute daft ideas?
For a few days that August, the normally reserved English rioted and sections of their cities burned. This was probably the first time that Twitter and Blackberry Messenger were used to co-ordinate civil disorder.
Professor Rob Procter of Manchester University (see Procter et al, 2013) and his team looked at the Twitter stream on the night of the riots and produced, with the aid of the Guardian newspaper, a fantasticdata visualisation of 2.6 million tweets. It repays study, but here is a taster:
Twitter is a perfect incubator for simple memes. Imitating the transmission is as simple as clicking a button, and (as will become evident) that very little mental filtering is applied.
There is a posh residential district close to London Zoo called Primrose Hill, and I was amused (and bemused) by the suggestion that there had been tigers roaming its streets. So I put my memetic hat and tried to see how this had arisen.
The following graphs show the rate of retweets of four specific ideas. The first, in orange at the back, proposed that rioters had hurt or released animals from the zoo. There was then a set of tweets, close to the front in brown, questioning whether this was true.
Then we had a suggestion, which absolutely took off, that tigers were loose on Primrose Hill. This is the huge peak in pale yellow. And finally, at purple in the front, we had the anti-meme stating that this was just not true.
Origin of the specious
So here we can see the start of the Primrose Hill meme, and the anti-memes that killed it. Let's take a look at the timeline of the preceding 48 hours in a little more detail, pointing out when the zoo rumour mutated.
|Time / Date||Rumour (variants existed)|
'those rioters are monkeys' (ethnic slur)
|02:00 07/08...||'those rioters are like zoo animals'|
|'we should lock those rioters up in the zoo'|
|22:00 07/08...||'we should set the zoo animals loose on the rioters'|
|02:00 08/08...||'let's loot the zoo'|
|'hoodies are looting the zoo'|
|21:05 08/08...||'looters have set the animals free from the zoo'|
'tigers are loose on Primrose Hill'
This looks like a game of Chinese Whispers, doesn't it? There is clear evidence of accidental evolution for the first 24 hours, but the reason that the Primrose Hill rumour spread like wildfire (see the yellow graph) was very specific - photographic 'evidence'. This was certainly spread (and may have been created) by a young woman in London with a large Twitter following, and is clear evidence of directed evolution.
As we saw in the section on the Arithmetic of Belief, interesting rumours spread quickly through a population, regardless of whether they are true. And boy, was this one interesting! It was furiously retweeted by about a thousand people, perhaps in horror or perhaps as a warning, followed equally quickly by other tweets questioning it.
So our tiger meme (the yellow curve) is up and running, but with a great deal of skepticism following it (the brown curve). Shortly after that, the anti-memes - basically saying “this isn't true, don't spread it” - emerged, and they in turn took off when someone noticed that the photograph had been plucked from a 2008 newspaper article about a tiger that had escaped from a circus in Maddaloni, Italy.
That anti-meme effectively killed the rumour.
So what can we learn from this? Firstly, that people tend to believe photos regardless of how improbable the story (tigers on Primrose Hill? Really?); dull truths spread much more slowly than interesting but untrue statements (well, we knew that anyway), scaring others is fun; and one person with a good story can sway a population. We may also have learned something about the gullibility of Twitter users.
But there is a serious point here: in this age of social media, it is frighteningly easy to spread disinformation and to manipulate populations. Of course, the 'quality' of a meme has nothing whatsoever to do with its truthfulness. This has such profound implications that the European Union is funding the Pheme Project to “model, identify and verify rumours as they spread in social media and on-line networks”. The intent is to be able automatically to verify whether an online rumour is speculation, controversy (e.g. the scare stories over the safety of the MMR vaccine in the UK), misinformation (where something untrue is spread unwittingly, or disinformation (where it is spread deliberately). The intention is to improve the quality of journalism.