It used to be called “dying of a broken heart” but modern scientists prefer the term “the widowhood effect”. After their spouse’s death, surviving partners have an increased risk of dying themselves – up to 66 per cent in the first three months, according to research last year from the Harvard School of Public Health in Boston.
The effect has been noted in dozens of studies and seems to apply to both genders. The possible reasons include the stress of caring for a sick person; grief leading to bad lifestyle decisions; and social isolation, because their partner used to be their link to friends and family. As the gerontologist Kenneth Doka put it after the Harvard study: “It’s not unusual that widowers will often say no one ever stops over any more, because they didn’t realise someone else was calling and inviting them.”
But the widowhood effect doesn’t stop with the surviving partner, as a young hospice doctor called Nicholas Christakis noticed several decades ago when he first began to practise medicine. A seriously ill patient’s condition affects not just their immediate family and friends but the next circle out, their children’s spouses, say, and the one after that, of those spouses’ colleagues and friends.
As he got older, Christakis became more and more interested in these “social networks” – a phrase we’ve come to associate with online environments such as Twitter and Facebook, but which are deeply embedded in our offline world, too. He now works at Yale as a sociologist, collaborating with a colleague called James Fowler, championing an obscure but fascinating discipline called “network science”.
Their research uncovered some astonishing results: obesity really is an epidemic, in the sense that it is infectious. If someone you know well is obese, you are more likely to be obese. The effect can be seen – albeit at progressively dwindling intensities – as you look at friends of friends, up to four degrees of separation from the original subject. (Christakis says the possible causes are induction – your friends get fat because you are fat; homophily – you are friends because you share a lively interest in eating biscuits; and confounding – you both live in a society where fast food is cheap and many jobs involve sitting down for eight hours a day.)
When I saw Christakis explain these ideas in a TED talk, the political and policy implications immediately struck me. And so, while he was on a visit to London to lecture at the Cass Business School, I met him to ask the obvious question: doesn’t this vanquish libertarianism once and for all? How can anyone argue “There’s no such thing as society”? He agrees. “My favourite example is the analogy of carbon – it has a number of forms, allotropes; one of them is graphite and one of them is diamond. If you take the carbon atoms and assemble them one way, you get carbon. Assemble the same atoms another way, you get diamond. The graphite is soft and dark and the diamond is hard and clear. Softness or hardness are not properties of the carbon atoms; they are properties of the collection of carbon atoms.”
Society is the same: it is just as much about the connections between individuals as the individuals themselves.
In person, Christakis is bombastically erudite and many of his answers are delivered as mini-lectures. His research has led him to think about the perils of a laissez-faire system of government, he says. “The function of the state is, at minimum, to address collective defence, market failure – things that individuals are not equipped to do or things that individuals, acting rationally, subvert. That’s the tragedy of the commons. Everyone is grazing their cattle on the commons and the commons collapses.”
Network theory has uses far beyond giving me ammunition to argue with people with Twitter handles such as @AynRand93, however. Christakis and Fowler create complex maps of connections between individuals in a group and use algorithms to identify the most structurally influential.
The professor gives me a concrete example: in India, women are particularly badly affected by the lack of toilet provision for 500 million of the country’s citizens. Open defecation is risky, because a lone woman is easy prey for sexual assault. In rural Bihar, where about 85 per cent of households have no sanitation, one victim this summer was an 11-year-old girl, raped when she visited the field at night. Many women try to avoid going to the loo at all for hours on end, increasing their risk of urinary tract infections.
So how do you convince a community that toilet provision is necessary? Even when local government or an NGO provides the funds, there is often still a cultural resistance. “If you ask people, ‘Why don’t you construct a loo?’ they’ll say, ‘But all my neighbours do the same thing,’” Christakis observes. He says the same model explains widespread tax avoidance in Greece – “If you ask people, ‘Why don’t you pay your tax?’ they say, ‘My neighbours don’t pay their taxes. What am I, an idiot?’” – and risky sexual behaviour in countries with endemic HIV.
Traditionally, you might try to win over the highest status member of the community: but he and Fowler believe you should instead target the most influential as identified by their algorithm. “That way, you can target 5 per cent of the people and 80 per cent of the people change,” he says – making scarce development funds go far further.
Christakis’s work gives him many insights into human nature, sometimes with unexpected results. Take the way that his work on obesity was reported. In America, he says, the headlines read: “Are you packing it on? Blame your fat friends.” In Europe, however, the message was different. “Are your friends fat? You are to blame.”