The human brain evidently evolved to commit itself emotionally only to a small piece of geography, a limited band of kinsmen, and two or three generations into the future…We are innately inclined to ignore any distant possibility not yet requiring examination…It is a part of our hard-wired Paleolithic heritage. For hundreds of millennia, those who worked for short-term gain within a small circle of relatives and friends lived longer and left more offspring.
19
So, in today’s small world, we are naturally inclined to behave in a way that diminishes opportunity. When we try to dragoon our friends into helping us, we will often pick people who might be willing but are not best suited for the job. Conversely, if we sift through all our contacts and tap people with different knowledge, we will usually find someone more appropriate. Who is more likely to tell you about a fantastic new holiday destination–your best friend (whose preferred locations you may already have tried) or an intrepid traveller you fall into conversation with on a train? Most of us have far more weak ties than strong ones. We usually have a huge dormant network of contacts, past and present–sociologists estimate that most of us know between five hundred and three thousand people by name. So we can deliberately cultivate many more weak links, windows into new worlds, with little effort. And though the great majority of our casual acquaintances will not uplift our lives, a few just might–and often when we least expect it.
Is your world ordered, random or small?
When we are growing up, our world is ordered. At least until we’re teenagers, our world comprises mainly strong links–parents and friends. We live in a structured world, with few casual acquaintances. We are well connected to the people around us, and largely isolated from the rest of the world.
My world remained structured until I was seventeen. Until then, the influences in my life were nearly all strong links–my family and friends from school and the neighbourhood. After I finished school, like many adolescents at the time, I took four months out to go hitch-hiking around Europe. I wanted to go with my best friend Ray, who’d done it before, but he wisely deflected me, saying he’d seen too many friends who’d travelled together and ended up not speaking to each other. So I went on my own.
The trip was a complete contrast to my previous, structured world. To survive, I had to meet dozens of strangers each day, take the initiative, mix with a motley crew of truck drivers, locals and fellow travellers. This was a truly random world. Each day I set out, I didn’t know where I’d end up, or with whom. I met grizzled Spanish peasants; Italian wide boys; stoned German hippies whose driving was scary; Greeks who told me how much they loved America and Britain; three Australian girls who’d been driving a Mini for two years through endless countries, consuming enough culture and pizza to last a lifetime; an Eastern European refugee who wanted to move to Britain; a Muslim travelling salesman in what was then Yugoslavia; sly American gay men who gave me a lift and asked if I’d been troubled by ‘queers’ and an amorous French lady of a certain age who lived in a huge villa on the Riviera. All the people I met were weak links, strung out in an endless chain, whose structure was tenuous and temporary in the extreme. For sure, it was a network; but for four months hubs and strong links were conspicuous by their total absence.
Soon after, I went to university. Here my life was neither structured nor random, but the best of both worlds. I had very strong links to a small number of people–friends, tutors, lovers–and my life and work with them were rather organised. But I also met a huge number of casual acquaintances, and kept up weak relationships with dozens or perhaps even hundreds of them. I was on the fringe of many disparate groups–poker players, Marxists, Tories, tennis players, student politicians, pot smokers, ravers, debaters, drinkers, serious historians, music lovers, and those who lived by our college head’s famous dictum that ‘fornication is for the afternoon’. University was a small world–I belonged to very few close clusters, and beyond these friends had a very wide range of weak links to different groups.
Within the space of a year I’d moved from a structured world, to a random one, to a small one. These three types of network–pretty much–were later described and analysed by physicists Duncan Watts and Steven Strogatz from Cornell University. Of course, they didn’t know about my early life and they wouldn’t have cared a hoot. But their findings have an uncanny and intriguing affinity to my three network types.
Doing pioneering research starting in 1996, Watts and Strogatz used computer modelling to explore the relationships between different types of networks.20 To simplify, suppose they started with a world of a thousand nodes–think for the moment of these nodes as people–arranged and connected to represent a highly structured and clustered large world. You could do this by setting out the nodes in a large circle with each connected only to its five closest neighbours to the right, and the same to the left, making five thousand connections in total.21 They called this a regular network and it corresponds closely to what I’ve called my structured network. The regular network was highly clustered, since each node was connected only to its closest ten neighbours, not to any distant nodes. In such a configuration each node is a high number of links away, on average, from the other 999 nodes–resulting in a high number of degrees of separation. Watts and Strogatz had modelled a classic, ordered, structured, clustered world where strong links mattered most, similar to the world in which we all grew up.
The researchers then constructed the opposite, random network. Imagine again setting up a thousand nodes in a circle with five thousand connections, but this time with the connections picked at random, cutting across the circle in any direction, and as a result links between neighbouring nodes much less frequent. In other words, the local clustering would be poor: there would be no well-connected groups of nodes next to each other. Also in contrast to regular networks, Watts and Strogatz found that all the nodes in random networks were much better connected on average–not to any local group (unless by chance, which almost never happened), but to any of the other 999 individual nodes through a small number of intermediaries. Each node could typically reach any other node through a small number of steps–and this results in a very low number of degrees of separation. With a bit of poetic licence, this is similar to my random backpacking days–no clusters, but a terrific range of random connections.
Finally, the physicists started with regular networks–where the nodes were isolated from all but immediate neighbours–and then added random links between the nodes. They were looking for the precise moment when the regular networks turned into random ones–they wanted to know how many links it would take. But what they actually produced–and very quickly–was a third type of network with the best of both worlds. Recall, the regular networks we started with had five thousand connections. If this is rewired by adding fifty new random connections–increasing the total number of connections by a mere 1 per cent, this turns out to be enough to make all the nodes well connected to one another on average. In this respect, the new network is as good as the random one. In the regular network the average number of hops necessary to link any two nodes is fifty; by adding just 1 per cent of random connections the hops necessary (the degrees of separation) plummets to seven. (Moreover, in later research Watts and Strogatz found that, regardless of the size of the network, adding just five links at random to a regular network halves the degrees of separation.) Yet, in all cases, the new networks retain the advantage of being highly clustered.