I’m a networks person. Some of you may already know that about me. I never got on well with hierarchical institutions. I detested that aspect of school. I had a brief encounter with the military life in the Glasgow Academy Combined Cadet Force; I was the worst cadet in the school’s history. I lost my entire uniform.
My resentment of hierarchical structures of governance naturally sent me to academic life, where, as those of you in academic life will know, there are nominal hierarchies but nobody pays the blindest bit of attention to them, and the higher up you are—say, the Vice Chancellor, or Master of a College—the less power you in fact have. Because as soon as academics get what the Americans call a tenure, they feel themselves to be members of the republic of letters, which is code for an entirely distributed network, where nobody has any power, or indeed responsibility.
That has suited me very well in my life, but it was only recently that I came to appreciate that it was rather odd to be as viscerally attracted to a networked world as I temperamentally am. Most people, I think, quite like being somewhere in the org. chart. And I’d like you to ask yourselves: “Am I an org. chart, hierarchy kind of person? Or am I a networks person?”
And the way to figure this out is to think about your work. Do you know how many people report to you? Do you have someone you report to? Do you, in fact, have a boss? And if you saw the org, chart of the entity you work for, would you know where you were in it, which tier?
If you work for a large corporation, then you are, whether you like it or not, a hierarchy person. But if, like me, networks matter more to you, the graph of your life is not a pyramidal structure at all. It’s a web. And the network of your family, of your friends, matters more to you than any hierarchy.
We could graph the network in this room. Each of you is a node. And the relationships you have with one another are edges. And we could find those relationships out quite quickly, with a series of questions, or if we had your data (perhaps you’ve already given your data away for free to Facebook). If I had access to that data, I could very quickly graph the network and find out a few things about you all.
I could find out who had the highest degree centrality. I could find out who here had the most edges—the most connections to the rest of the nodes. And then I could find out who had the highest betweenness centrality, who was the most influential person, the person who was most worth knowing, who would connect you to the most other people.
In my new book, The Square and the Tower, I’ve attempted to do two rather different things at once. One is to explain history to people who do networks. Now, if you go to Silicon Valley, which I did about a year ago when moved from Harvard to Stanford (which is right next to Silicon Valley), you meet a bunch of people who do networks for a living. And most of them have engineering backgrounds or computer science backgrounds and they basically studied no history. And their assumption—I certainly found this a year ago—was that history was something that began with the Google IPO, or maybe the founding of Facebook, and everything before that was prehistory or the Stone Age. Part of the point of The Square and the Tower is to say to those people: you need to understand what you do in historical context.
The book also tries to explain networks to people who know history. Because most historians, even the ones who say they are social historians, don’t really have a clue about networks. The number of historians who understand the network science that I think you need to understand in order to study networks, is really very small. Most historians over, say, 200 years of historiography, have studied hierarchical institutions.
And that’s partly because hierarchical institutions are the ones that keep records. Archives are the products of hierarchical institutions, like states. And so naturally historians tend to see the world from the vantage point of hierarchical institutions that keep records. Networks do keep records, but by definition they tend to be scattered.
Let me give you an example: the illuminati. Now, if you google “illuminati” you will pass into a wonderful world of conspiracy theories of varying degrees of lunacy. To wit, the illuminati control the world. The illuminati control the Federal Reserve. They control the Rothschild family, which controls the Federal Reserve. They control George Soros, who controls everything else. And they have controlled all institutions of note since the French Revolution, which they started, along with all other historical events of note. And at some appointed time, (this is an option available to the fringe of conspiracy theorists), the illuminati will welcome the aliens from the mothership to the earth to take over the human race entirely.
So, the problem with networks from the vantage point of the respectable historian is that the conspiracy theorists love to write about them. Whether it’s the network of Jewish financiers, the network of Enlightenment thinkers, the network of Freemasons, all of these subjects have attracted the worst kind of attention from fantasists and victims of paranoia.
In order to write seriously about the illuminati, you have to do quite difficult digging. The illuminati existed. They were in fact a 1770s era secret society of radical Enlightenment thinkers, who hatched the scheme of advancing their radical enlightenment ideas by infiltrating masonic lodges in Germany. There were never more than about 1,500 members of the illuminati. They had an enormously elaborate ritual with oaths and codenames, but they certainly didn’t cause the French Revolution. In fact, the illuminati had ceased to exist some years before the French Revolution because the Bavarian authorities clamped down on them, rightly identifying them as a subversive secret society.
I tell you this story to illustrate the fundamental difficulty of writing about networks as a historian: it’s difficult but it’s doable.
Why does this matter? Well, it matters because in our time, networks have become larger, faster and more influential than at any time in history. Let me illustrate the point: Facebook has two billion regular users. In the United Kingdom the latest number is 34 million; just over half the population of the United Kingdom is on Facebook. Twenty million Britons use Twitter. Twenty million, the same number, use YouTube regularly. Just under twenty million are on LinkedIn. I could go on.
We live in a time of network empowerment, a time when hierarchical institutions, like states, kingdoms, and hierarchical leaders like prime ministers, have become figures of fun. You become a figure of fun not because a comedian interrupts your speech, or the letter “F” falls off the sign behind you. Those are symptoms of your loss of legitimacy. Your loss of legitimacy is a result of the radical empowerment of networks that has occurred within 15 years, because every one of the companies I just mentioned, beginning with Facebook, was founded in the last 15 years. My impression is that, while the masters of Silicon Valley overestimate the networks and misunderstand their potential, most people in Europe still struggle to understand the magnitude of the structural shift in the public sphere that has occurred.
For example: 45 per cent of Americans get their news from Facebook. That is nearly half of adult Americans. There has never been a content publisher in the history of the United States so powerful. Not William Randolph Hearst, not Rupert Murdoch, and yet, as I’ll try to explain, it is still the case (and I know this from personal experience) that legislators in Washington DC do not know how urgently they need to change the regulatory framework that currently governs the network platforms such as Facebook and Google. I’ll come back to that point in a minute.
If you’re Mark Zuckerberg, the creator of Facebook, you sincerely believed at the beginning of your adventure, in your Harvard dorm room, that everything would be awesome if everyone was connected. If you’re one of the founders of Google, that was also your assumption: that a planet where everyone is connected would be awesome. And I argue in this book that the disillusionment we have seen in the last 12 months with that utopian vision was entirely predictable if they had known history or paid any serious attention to network science.
The expressions of disillusionment are coming on an almost weekly basis. Sheryl Sandberg was just the other day quoted as saying that it had never dawned on them, no never, that somebody might use the tools of Facebook for nefarious purposes. One of the founders of Twitter said, “I thought it would be awesome if everybody was connected. It didn’t turn out that way.” My response is: “Well, duh.” Because the idea was always delusional that giant, very fast-moving networks would produce what Zuckerberg has called “a global community” in which we’d all happily share cat videos.
There are six reasons why networks don’t work that way, and I want to tell you what they are briefly. And then I’m going to tell you what the implications are, first for the United States, and then for the United Kingdom.
Networks have a number of intriguing characteristics, and the bigger they get, the more obvious this is. One is the phenomenon of homophily, a word you will rarely hear, although if you’d listened to the Today programme on Monday you’d have heard me say it. The presenter looked rather nervous when I said it, but homophily is not an act illegal in southern states of America. It is, in fact, the phenomenon of birds of a feather flocking together. On networks, people tend to form clusters of like-mindedness. It doesn’t necessarily need to be ideological similarity, it can be any kind of similarity. This has been known for years. In American high schools, sociologists have observed for decades the tendency in ethnically or racially mixed schools for the different racial groups to form clusters spontaneously. So self-segregation is a familiar manifestation of homophily.
If you look at some of the recent studies that have been done of Twitter, or the “blogosphere”, you will encounter an astonishing illustration of homophily in action. I’ll give just one example: in the United States, retweeting occurs in two almost entirely separate clusters, a liberal cluster and a conservative cluster. People do not retweet things that they do not ideologically agree with, with very few exceptions. Moreover, the probability of a tweet being retweeted is 20-30 per cent higher if you use strong, emotive language. That is why, if you’re on Twitter, so many people use offensive language. They have figured out intuitively that this is a way to get retweets.
So, number one, the idea that we would form a global community if we all were connected, ignored the phenomenon of homophily, the tendency for clusters to form and indeed, to diverge, for the polarisation to increase. Think about how this works. You’ll have heard people talk figuratively about “echo chambers” in the course of the past year. What they’re talking about is this precise phenomenon where people cluster together and then reinforce their existing assumptions by retweeting or sharing or liking and then adding emotive language as they do so. It’s as if confirmation bias had been institutionalised so that you’re almost guaranteed only to get evidence or opinions that accord with your own preconceptions. That’s part one.
Part two: things go viral. But how do they go viral? Why do some things become tremendously popular? You might assume that it’s because of the inherent quality of the “meme” or idea, just as you might assume that a disease causes an epidemic because of some inherent quality of the virus. That’s not the case. It is as important to understand the structure of the network that it attacks as to understand the idea or virus in itself. Network structure explains why some things go viral. The point of entry into the network can be decisive. So, another unsurprising feature of our network world is that apparently random and indeed crazy stuff goes viral. Untrue things are as likely to go viral as true things, maybe even more likely.
Third point: weak ties are strong. It’s a small world. If we only networked with our close friends and family, the world would be a kind of balkanised network with lots of little unconnected clusters. But that’s not what the world is like. The world is a small world. Roughly 100 years ago, it was observed that there might be six degrees of separation between any two randomly selected individuals on the planet. This gave rise to a whole series of games like “Six Degrees of Monica Lewinsky”. On Facebook, there are now just 3.57 degrees of separation between the two billion people who use that platform. So we are all friends of friends of friends of friends of friends. This is an important insight because it explains why, when things go viral, they go so far. They can go around the world in seconds.
Fourthly, networks never sleep. They’re not static. They’re constantly evolving. They have what the physicists call emergent properties. They’re complex systems with unpredictable dynamics. And that’s one reason that a network world is strangely volatile.
Fifthly, networks network. And as Stan McChrystal famously said, “It takes a network to beat a network.” Interactions between networks are an immensely important part of the way our world works. Think only of what happened last year – more of that in just a moment.
Finally, networks are not egalitarian. If you thought a networked world would be a lattice of equal netizens, all speaking truth to power, as the liberals like to say, you were in for a shock. Because that’s not how social networks work. Because of something called preferential attachment, when new nodes join the network they always want to join the most connected nodes. And so when you look at the structure of a network—if you looked at the structure of Facebook, for example—you’d find it was, to use technical language, scale free, or governed by something close to a power law.
A few nodes have an insane number of edges. Think of the 50 million Twitter followers that Donald Trump has amassed. And a vast number of nodes have hardly any. So, in an unequal world, networking the world amplifies the inequality by creating a new tier of inequality that is, in fact, not uncorrelated with wealth. The networks promote inequality. Ask yourself why it is that so many of the richest men in the world made their money from building the infrastructure and the superstructure of the internet and the social networks that we see in such dominant positions today.
I will not go into the huge analogy which is central to The Square and the Tower, in which I argue that our time more closely resembles the 16th and 17th centuries than it resembles any more recent period of history. The argument I make, in brief, is that the advent of the printing press in Europe and its rapid dissemination throughout most of Europe from the late 15th into the mid-17th centuries, created a very similar set of conditions to the ones that we have in our time — a set of conditions that strengthened networks relative to hierarchical institutions in ways that completely took the established hierarchies by surprise.
Five hundred years ago to the year, Martin Luther launched his campaign for reform of the Catholic Church. Without the printing press he would have been just another chargrilled heretic. But with the printing press, Luther went viral. And Luther thought he was going to create a priesthood of old believers and everything would be awesome. Instead of which, Europe became acutely polarised between those who supported Luther and those who opposed him and, for 130 years, intermittent and bloody, very cruel religious conflict ensued. Crazy stuff went viral: not only the notion of justification by faith alone which was one of Luther’s ideas, but also the notion that there are witches amongst us who should be burnt.
If you’re interested in that analogy, read the book because I think that it’s an incredibly powerful and worrying predictor of what could be a protracted period of polarisation, of manias and panics, and potentially also of conflict. There are similarities and differences, of course, between these two eras, but for the sake of brevity let me leave that to one side and offer some brief reflections on the implications of the networked world for American and British politics.
We saw last year in the US election the ways in which networks can disrupt established hierarchies. The political system, the Republican Party to be precise, was overwhelmed by the Trump network. Trump himself went viral. The curious mixture of corporation and family firm and cult figure that Trump is simply swept the established Republicans, beginning with Jeb Bush, to one side and then proceeded to win the election with some help from two other networks. First, the network of Islamic State, because its well- or ill-timed terrorist atrocity in Orlando last June did a much larger amount of damage to the Democratic cause than was realised at the time. Nothing polled more strongly last year in the course of the campaign than Trump’s response to that, with his call for a Muslim ban.
Then there was the third network. Remember I said that networks can attack other networks? Well the Russian intelligence network successfully attacked not only the Democratic Party’s email servers, but also the Trump organisation itself. The full extent of the penetration and collusion has yet to be revealed. The networks used the social networks. Trump’s campaign used Facebook for surgical strikes on key swing states because Facebook’s data allowed them very precisely to target advertisements, and hone advertisements until they had got the message absolutely right. The Clinton campaign outspent the Trump campaign by two to one. But it blew the money on old-school campaigning: local TV advertising.
What does all this imply for Britain? We live in a time when neither the United States nor the United Kingdom really knows how to manage the advent of network platforms as the dominant sources of information in the public sphere. In the United States there’s a frenzied debate on the Left about whether to bring anti-trust actions against the big platforms. This is a doomed venture. These are natural monopolies. You’re not going to break up Google and have Google Arkansas.
The Right is asleep at the wheel. It has still not dawned on many Republicans that 2016 will never happen again, because Silicon Valley, with a handful of exceptions like Peter Thiel, leans Left. The algorithms have already been tweaked to the disadvantage of the far Right, and increasingly not just the far Right. Who decides what is hate speech? After Charlottesville there was a great outbreak of virtue signalling in Silicon Valley. White supremacists had to be banished from the internet, and while we’re at it, why not Islamophobes? Who decides what hate speech is? Mark Zuckerberg? European regulators say yes, it’s up to Facebook. It’s Facebook’s job to do the censorship for the German government. This seems to me a crazy line of approach because it empowers the social networks even more.
In the United States, uncertainty reigns. Facebook says it’s a network platform, not a content publisher. Under 1996 legislation, it is free from any liability for content that appears on the platform, with a handful of exceptions. Rupert Murdoch argues, as does Lionel Barber, the editor of the Financial Times, that this is unsustainable and I wholly agree. But will it change? If it doesn’t change soon, 2018 is going to look a lot different from 2016. You can make Sean Hannity vaporise from the Facebook news feed by just tweaking the algorithm.
In this country, it seems to me there is even greater cluelessness about the implications of the social network phenomenon. And let me illustrate this point with some numbers. Jeremy Corbyn today has four times as many Facebook and Twitter followers as Theresa May. And four times as many, too, as Boris Johnson. You may be a fan of Ruth Davidson, but Nicola Sturgeon dominates Ruth Davidson on social media. Sadiq Khan has more followers on social media than any Conservative politician. I have more followers on Twitter than Jacob Rees-Mogg. Four times as many, to be precise.
You laugh. This matters. Most pundits who followed the 2016 election in the United States were wrong. If they’d looked at data like those, they’d have seen that on Twitter and Facebook Donald Trump dominated Hillary Clinton. Dominated her, but not by the same margins that Jeremy Corbyn today dominates the leaders of the Conservative Party on social media.
My sense is reinforced by a week here—and the pitiful performances one witnessed on television at the Manchester Tory Conference—that the social networks are coming, and they are coming to town to deliver a Corbyn government.
This should trouble anybody, inside this room and outside this room, who thinks seriously about what that would imply for the future of this country. There are a lot of towers in this town, some very beautiful ones. I think of one in particular with a big clock that the world hears regularly ringing on the BBC World Service. But what is happening in the new squares, the squares of the social networks? That will turn out to matter a very great deal more than what has been happening just lately in the shadow of that tower.
Taken from a speech delivered to the Legatum Institute