Skip to content
Who's in the Video
Nicholas A. Christakis is a physician, sociologist, and director of the Human Nature Lab at Yale University, where he is the Sterling Professor of Social and Natural Science. His most[…]

A conversation with the Harvard physician and social scientist.

Nicholas Christakis: My name is Nicholas Christakis. I'm a physician and a social scientist at Harvard University.

Question: What do you do on a day-to-day basis?

Nicholas Christakis: I’m a professor of medicine and professor of medical sociology at Harvard Medical School and a professor of sociology in the Faculty of Arts and Sciences and I do what I would call network science, so I, for the last ten years, have been studying how and why human beings come to be embedded in social networks. Not the kind of online kind that people might think about all the time nowadays, but the kind of ancient kind that we have formed for hundreds of thousands of years. And I study how we humans come to form these very elaborate networks and what these networks come to mean for our lives, and the sort of the field as it were of medical sociology is concerned with all sorts of phenomena, social processes and social phenomena that influence health and health care. But I’m focused primarily on I would say a subset of that or not a subset, but a different field. Let’s say network science.

Question: What kind of research went into your book “Connected”? 

Nicholas Christakis: We have done work on the social, psychological, mathematical and biological rules that govern how human beings come to form social networks—the structure of networks—and then we’ve also examined the kind of social and psychological rules or attributes of how social networks function. So how do we form social networks and how do they affect our lives, and it’s what we would consider to be the anatomy and the physiology of a kind of human superorganism. In a very fundamental way we are like ants or actually kind of like fungi too, where individual human beings assemble themselves into these elaborate complex structures and we’re… James Fowler and I, my coauthor, are deeply concerned with how and why we form these structures and what they mean for our lives. So in the book we present about… We talk a lot about our own research, but we also pull in the research of many other scientists who have been looking at a variety of phenomena, and we talk about the role of social networks in human emotions. We talk about the role of social networks in human romantic and sexual behavior, in health, in politics and in economics, and then we also talk a little bit about the genetics of human social networks and the sort of modern online variety of social interactions, and then we close with an argument in the book about why we form social networks and what in a very deep sense they mean for our lives.

Question: How have human social networks changed over the millennia?

Nicholas Christakis: Well I mean for thousands of years people have been concerned with very basic questions about how and why humans form societies, how and why people form groups, but the social networks are different than groups in that in addition to a collection of individuals a network has a specific set of ties that you add to the individuals. And not just ties, but a specific pattern of ties. So, for example... the simplest example of this would be if you take a group of a hundred people that are waiting in line to get into a theater, for example. That’s a group of people, but it’s not a network. If you assemble those people into the simplest possible network, a linear network, like a bucket brigade to put out a fire for example now you have these hundred people and you’ve added 99 ties between these people, so and a specific pattern of ties. Each individual is connected to one individual on the left and one on the right. And now this network is capable of doing something which the group was not capable of doing, namely, putting out a fire rather more efficiently than a group of people. Or you could take the same group of people and take the same 99 ties, but organize them totally differently in where each individual now... in the form of a telephone tree for example, so each person calls two people, so you take the first person. They call two people. Each of those two people call two people and then you would get a completely different sort of branching pattern. Now instead of a linear network you have a more complicated network. 

In fact, the same kind of structure, archetypical structure was used by Bernie Madoff in a kind of a Ponzi scheme, but instead of distributing information outwards, money was sucked up and drawn inwards towards the center, so these would be artificial human networks. They have constituent individuals and they also have a specific pattern of ties. You add something more to the individuals, these ties. And in fact, as we argue in the book it’s the addition of these ties that makes the whole greater than the sum of its parts. It’s the addition of the ties that makes the population of people within it, the network, capable of doing things like putting out a fire or distributing information rapidly in the telephone tree example, that it wasn’t previously able to do. So a network of people is a collection of individuals and a collection of ties between them and a specific pattern of ties at that.

One of the key ideas about human social networks is that in the addition of ties between people and specific patterns of ties that obey particular mathematical rules the whole becomes greater than the sum of its parts. The collection of human beings have properties that do not reside within the individuals, and this collection of human beings is now able to do things that they previously were not able to do. And one of the illustrations or examples that I most like to give about this is something that most people are familiar with from high school or college chemistry and that is the example of carbon. So you can take carbon atoms and you can assemble the carbon atoms into graphite and here we put particular hexagonal pattern of ties and you get sheets of graphite and this graphite is soft and dark. Or we can take the same carbon atoms and assemble them differently into a kind of a perimetral structure with the ties between them, the bonds between the carbon atoms and we get diamond, which is hard and clear and these properties of softness and darkness or hardness and clearness first of all differ dramatically, not because the carbon is different. The carbon is the same in both, but rather because of the ties between the carbon atoms. And second these properties are not properties of the carbon atoms. They’re properties of the group, properties of the collection of carbon atoms. Therefore, when we take constituent elements and assemble them to a larger whole, this larger whole can have properties that we could not have foreseen merely by studying the individual elements and properties which do not reside within the individual elements.

And the same thing happens with human beings. We can take human beings and assemble them in different patterns and depending on the pattern in which we assemble human beings they have properties that we could not have understood just by studying humans. For instance, individual human psychology is not enough to understand some of these bigger properties and second these individuals depending on how they’re assembled can have different properties, so you take the same group of people and you assemble them one way and you get a bucket brigade, which has particular properties or you assemble them a different way and you get a telephone tree, which has yet again different properties. And so the pattern of ties between individual people is actually a kind of a resource that we all can use. It’s actually a reservoir of value. It’s a kind of social capital, actually.

And it’s not just the pattern of ties between people that matters.  It’s also what is flowing across those ties, so if you inhabit a network with a particular structure of ties, but it’s a trusting network versus a mistrustful network it has different implications for your life, or if you’re inhabiting a network where a pathogen is spreading versus where a pathogen is not spreading—different implications for your life. Something is spreading through the network. You’re connected to others and it affects you.

And we have been looking, James Fowler and I, at a variety of sort of counter intuitive examples of these kinds of phenomenon. For instance, we’ve looked at how things like obesity or your emotions, like happiness, spread through human networks. And we find that a lot of deeply personal things, things that people might not think of as being under the influence of others are affected, not just by their friends, but by their friends’ friends and even their friend's friend's friends. So people are used to think of things like fashions, for example. Like their taste in music or clothes might be affected by their friends or perhaps even they have this image that fashions can spread through the network or people might be used to thinking that germs, that right now they’re not sick, but their friend's friend's friend has a germ and that germ is going to spread to their friend’s friend and then to their friend and then eventually inexorably to them, but what they may not realize is that other sorts of phenomenon like who they vote for or how big their body is or even how happy they are also can behave in similar ways, and that is what James and I have been working on trying to understand over the last few years.

Question: Are there key differences in the way different phenomena spread through networks?

Nicholas Christakis: I think it’s important to emphasize that not everything spreads in networks and not everything that spreads, spreads by the same mechanism. So germs spread differently than money, which spread differently than behaviors, which spread differently than ideas, which spread differently than emotions, and so on. So just because something spreads doesn’t mean it spreads the same way.  And what is interesting is that we’re used to or we’re…most people would accept that claim or would be used to thinking of it in any case if we spoke about germs for example. So we know that sexually transmitted diseases spread differently than the flu. Not all germs spread the same way. Similarly we shouldn’t expect all other phenomena to spread the same way, so mechanistically there are a variety of social, psychological and biological processes, which can undergird these spreading phenomenon, how things might spread within the network and, for example, in the case of obesity there are at least two possible explanations for why your friend’s friend’s friend’s weight gain might ripple through the network and cause your weight gain. Or why well just to stick with the obesity example for a moment, so there are at least a couple of possible… So there are at least a couple of possible explanations for why your friend’s friend’s weight gain might spread through the network and affect you. One is, is that there is a spread of a behavior through the network, so your friend for example says let’s go have muffins and beer. That’s a terrible combination, but you pick up this bad habit from your friend and in so doing you come to adopt the weight status of your friend. Another possibility is that your friend starts gaining weight and it changes your ideas about what an acceptable body size is, so now it’s spreading from your friend to you as not a behavior, but rather a norm: a specific idea about acceptable body size. And both these behaviors and these norms could in principle…not in principle, actually do spread not just from one person to another person, but from person to person to person and person to person to person to person. So, for example, your friend’s friend’s friend can gain weight and that can ripple through the network and affect you either because of the spread of behaviors or because of the spread of norms or both and we find suggestive evidence for both.

Question: How can social networks multiply positive health effects?

Nicholas Christakis: Well we’ve looked at a variety of phenomena. I mean we’ve also looked at, for example, the spread of smoking cessation, the spread of sleep behaviors, a variety of other health related phenomenon... the spread of adherence to medications, proper self management if you have various chronic diseases. And we find similar influence processes for these other phenomena.

So what's crucial to understand about networks…So what is crucial to understand about networks is that they are kind of agnostic. Networks magnify whatever they are seeded with, good or bad. They will magnify fascism. They will magnify germs. They will magnify sadness, but so too will they magnify love and altruism and smoking cessation and ideas. In fact, we just did some work that was just published a couple of weeks ago on…doing some experiments where we studied a kind of "pay it forward" phenomenon where we took random individuals and they played something called the Public Goods Game. And we found that if one person was kind to another person that person then went on to be kind to a total stranger and then that stranger in turn went on to be kind to yet another stranger, so we could see the signature of my kindness. If A is kind to B and B goes on to be kind to C and C goes on to be kind to D we can see in the interactions between C and D a signature of the interactions between individuals A and B, even though A never interacted with C or D, never saw C or D. So you have a kind of rippling through the network of this kind of pro-social positive kind of kind or behavior we found in these experiments that we did. 

So and actually interestingly that also spread to three degrees of separation like some of the observational we’ve done with other phenomena. But here is the point: networks magnify whatever they are seeded with, good or bad, but on balance the reason human beings form social networks is that the benefits of a connected life outweigh the costs, and the magnification of good things outstrips the magnification of bad things. So even though our connection to others puts us at risk for instance of getting germs, it also puts us a risk for getting ideas. And the benefits of getting information from other people in this simple example more than outweighs the cost that we pay from our connection to others. Look, if you wanted to avoid ever getting the flu or sexually transmitted disease or any other kind of interpersonal pathogen, be a hermit. Don’t talk to anybody, but if you did that you would lose all the benefits of social interaction and so it’s because of those benefits that we assemble ourselves into social networks.

Question: How can we manipulate social networks to spread desirable effects?

Nicholas Christakis: Well there are two distinct issues here. One is how do networks naturally come to evince particular spreading processes and the different one is how might we manipulate these ideas to seed networks affirmatively with desirable properties and the way we think about this in terms of policy interventions is there are two broad classes of things we might do. We might manipulate the pattern of connection between people or we might manipulate the pattern of contagion between people.

So simple examples of connection are ones most people are familiar with. For example Alcoholics Anonymous or Weight Watchers are artificial social networks. You take a group of people and you form a set of ties between them and having done so you achieve or exploit a kind of power of social networks to magnify whatever they are seeded with, so you get this positive reinforcement between these people and just like the bucket brigade example we discussed earlier, these people now assembled into a little artificial miniature network that is focused on alcohol you know cessation or weight loss are now able to do something or at least do something better than the individuals acting alone. And it’s not just that the weight loss like if you had 10 people and you gave them each an intervention and they each lost a pound they would lose a total of 10 pounds, but you put these 10 people together in a room and you have them interact with each other in order to lose weight and they might lose a total of 20 pounds. So the group, now having been assembled and this interaction having been fostered you get more bang for the same efforts, same people, but now with the ties between them you can get more benefit.

So one set of interventions would be focused on manipulating the connections between people, creating connections, cutting connections, particular patterns of connections forming the kinds of networks, the prototypical networks that we described earlier. A completely different set of interventions might be focused on manipulating patterns of contagion within the network. Here, for example, you might map the network of a set of school kids in a classroom and target specific kids to receive a smoking cessation message or a seatbelt use message. With the idea that if you were able to persuade that one individual you would get, sort of, cascade benefits to other people. So it’s not just that one person that benefits, but their change in behavior influences others and so forth, so that you get… you can collect all of that benefit just for the one cost that you bore in changing the behavior of one individual. So here the idea might be that there might be particular individuals who are more influential within the network, but this is actually a very complicated area because it’s not enough to have influential individuals. You also need influencable individuals. And it’s not enough to just have shepherds and sheep, but you also need them to be connected in very particular ways, so I don’t want people to think that "It’s just very simple. You just map the network. You find somehow the most influential person and you just target them and then everything is terrific." First of all, behavior change is difficult regardless, but second it’s not always obvious how best to intervene and manipulate networks.

So I’ll give you one simple example. We mapped a network of college students and we looked at their tastes in things. We used actually Facebook in this particular case to get our data. Most of our data is real life, face-to-face networks or experimental networks that we create. This was an online network example. And we looked at what these students listed as their favorite taste in movies and we found that of the top 10 movies that they listed not all of the movie tastes spread. So for instance, if I listed "Lord of the Rings"as my favorite movie, it didn’t affect my friends to also be interested in "Lord of the Rings." Maybe they were interested in "Lord of the Rings" anyway from advertising or they heard about it or God knows what, but it wasn’t because I chose "Lord of the Rings"that it affected them. But certain movies if I listed a taste in this movie it did affect my friend’s interest in the movie. For instance, we found that if people picked "Love Actually" as their favorite movie it affected whether their friends liked "Love Actually" or if they picked "Pulp Fiction" as their favorite movie that it would influence their friends to like "Pulp Fiction" as well. Okay, fair enough. Well then when we mapped the network what we found was that the seemingly most influential individuals or at least individuals you might superficially think of as influential, those in the center of the network they were all "Pulp Fiction" fans and the "Love Actually" fans were located on the edge of the network. So if you’re someone that is trying to sell "Love Actually" and you think I’ll just map the network and find the central individuals it probably wouldn’t work because those are the "Pulp Fiction" fans. Trying to get them to watch "Love Actually" isn’t going to be so effective. So in this example the point is that a simpleminded kind of sense that "Oh, the central individuals—all I need to know is the structure of the network, I know who is central, I’ll just give them my message," isn’t likely to be effective because you need to know more about the system, more about the pattern of ties, the individuals within them and the processes by which things flow before you can really have an intervention.

Question: How can startups or other new organizations design social networks for maximum effectiveness?

Nicholas Christakis: I want to emphasize that although I am obsessed with networks and a zealot for networks I do not think they’re a panacea. That is to say I don’t think that every social problem or every social challenge can readily or easily be mapped with some kind of network intervention. And you know we have all kinds of other longstanding policy levers at our disposal—taxation policy, legal policy, advertising—I mean there are lots of ways in which we try to change our society for the better let’s say. So I don’t think that for instance anything can be addressed with a network approach.

Now with respect to organizational networks, for example, startups that you asked about, I don’t have like some kind of general rule, but I’ll give you a little nugget of an idea, which is that in the network of organizations you don’t want to be overly networked or overly densely interconnected with entities that resemble you, nor do you wish to be too diffusely linked to entities that don’t resemble you. There is a kind of a sweet spot in the middle and my favorite example of this at the individual level is some work that was done by a colleague of mine, Brian Uzzi at Northwestern University. And what he did is he was very interested in the success of Broadway musicals and this is a very famous paper of his and he assembled a dataset of about 350 Broadway musical production companies—so the director, the producer, the actors and costume designer, so forth—and he mapped what the network of those people were, how interconnected they were and at one extreme you could have individuals who never had interacted before and that would be what we would call in network science very low transitivity or very low density in the ties between the people. So people didn’t really know each other, they hadn’t worked together before. At the other extreme you had a group of people all of whom had all worked together before, so very high density of ties. Everyone knew everyone else within this little network because they had all worked together before. And in between you had a mixed bag. Some people who had worked together before on the team and some were new to the team and had never worked with anyone on the team before. And when he plotted financial success and commercial success of these Broadway musicals on the Y axis and the density of ties or transitivity on the X axis, he found that if everyone had worked together before the show was a flop. And if nobody had worked with anyone else before the show was a flop. And the optimal success was in the middle, when you had the middle number of interactions or ties between people. Something between let’s say graphite... you know, coal, which is totally disconnected, and diamond, which is everyone rigidly connected, something in between was optimal... to pick up the example of carbon we discussed earlier. 

So I suppose then if one wanted to extrapolate and I’m not saying this is true, but I’m saying one could imagine this to be true. If you’re an organization and you’re embedded, a startup as you asked, embedded in a network of other organizations you want to have some relationships with firms that resemble you, but you also want relationships with firms that are very different to maximize your chances of success. And actually these ideas relate to some other ideas that James Fowler and I have explored on the genetics of human social networks and what we find is that people vary in their transitivity. People vary in the extent to which they’re friends know each other. So most people who are familiar with the idea that people vary in how many friends they have, some of us are born shy, some of us are born gregarious, some people have no friends or two friends, some have 80 friends or eight friends. People vary in how many social intimates they have and people are familiar of thinking about this as a kind of a genetic... maybe it’s partially genetic if you’re born shy or not. And in fact in our work we found that about 46% of the variation in how many friends you have can be ascribed in part to your genes.

Well we also find that whether or not your friends know each other also can depend in part on your genes, so we find that if you have Tom, Dick and Harry in a room, whether Dick knows Harry depends not just on Dick and Harry’s genes, but also on Tom’s genes. This is a very bizarre result. What I’m saying to you is whether your friends know each other has to do with something to do with your own genetic heritage. And we think the reason for this is that people vary in their tendency to introduce their friends to each other. So some of us you know knit our networks together and introduce all our friends to each other and other of us in a kind of worlds collide theory, you know, don’t introduce our friends to each other, keep them separate from each other. And it turns out that you can construct arguments for why people might vary in how many friends they have. Sometimes it’s to your advantage to have many friends. Sometimes it’s a disadvantage. Ad you can also construct arguments about why it’s to your…why you might introduce your friends to each other or not. Sometimes it’s to your advantage to have all your friends know each other and sometimes it’s not and so the example that we give in this particular paper and that we discuss in the book is that your friends knowing each other makes it easier for you to achieve an objective. 

For example, in evolutionary time, you know, maybe to bring down a big animal, so if you want to hunt a mastodon you want a group of people who all know each other very well, everybody knows everybody else. Let’s go kill the thing. On the other hand if you want to find a mastodon that’s not the group you want because if everyone is connected to everyone else your friend’s friend is right back again a friend of your own, so do you know where the mastodon is? No. Do you know where the mastodon is? Each guy comes right back to the first guy whereas in a network that has low transitivity your friend’s friend is not your friend. That person is now able to get information from a more distant location within the network, let’s say about where the mastodon is. So the point is, is that different microstructures of the network confer different advantages depending on what the challenge is. If the challenge is to acquire information you might want low transitivity. If the challenge is to work together you might want high transitivity. And therefore it’s not hard to imagine that these tendencies as well might be deeply rooted within our genetic heritage.  So we find this too and so what we find therefore is that human beings assemble themselves into particular kinds of networks. We do this naturally. We vary one person to the other in what kinds of networks we construct for ourselves and whatever network we construct for ourselves, turns out has rather profound implications for our experience in the world.

Question: Are the effects of social networks changing or accelerating in the digital age? 

Nicholas Christakis: I don’t think so, and I’ll give you the example that I usually give: The size of a military company in the Roman Army was about 100 men. Centurion, right, led 10 groups of 10 men. They were Decurions that reported to him. Each one led a squad of 10 men and there was a company of 100 men.  It's the Roman Army. In the modern American Army it’s the same, 120, 150 men in a company, little squads of 10 men. Well now why is that? I mean we have invented huge advances in communication. We have telegraphy and telephony and radar and radio communication and the Internet... and yet the size of a working unit in the military is the same as it was. We’ve grown in our ability to communicate with each other by many orders of magnitude and yet the company size hasn’t changed. And the reason is that what limits the ability of human beings to interact with each other, or what permits it even, is not the communication technology, but the fundamental capacity of the human brain to form social interactions and to process social information. So I would argue that actually online networks and modern telecommunications don’t fundamentally modify the capacity of the human brain for social relationships and therefore, while they might affect the efficiency of information transmission and might have other sort of affects at the margins, at the core it hasn’t changed.

And in fact, if you think about this other example different from the Roman and modern army example: talk to your grandparents and ask them how many best friends they had when they were young. They’ll say they had one best friend and two close friends. When you talk to your own kids now it’s the same. So even though my kids might you know Sebastian and Lysandra and Lena might you know have hundreds of Facebook friend, those really aren’t real friends. Those are acquaintances. The core social group hasn’t fundamentally changed despite this incredible new technology that is available. So I guess the way I would answer your question is to say that while it is the case that these advances in communications are astonishing I don’t think they change our fundamental humanity. I don’t think they change our desire to interact with others or our ability to interact with others, because I think those desires and abilities are dictated more by our brain and our humanity, which hasn’t changed even as the technology has. 

Question: Do Web media increase the efficiency of information transfer? 

Nicholas Christakis: No, I don’t think so. I mean think about Twitter. Like, I mean I don’t know how many Twitter followers you have, but you know if they’re all tweeting three or four times a day and you have a thousand followers you’re getting 3,000 tweets. You are not reading those tweets. They’re not affecting you. And in fact, when we looked at like when we looked at some of our Facebook networks we looked at the spread of or the diffusion of tastes in books, movies and music in Facebook networks. We found that if anyone of these random acquaintances of yours expresses a new cultural taste, it didn’t affect you. But people that you appeared in photographs with that you tagged and posted onto your Facebook page. What we call your "picture friends" who might be your real friends. So on average you might have 110 Facebook friends in our data, but you only had about six-and-a-half real friends, picture friends, when one of your picture friends expressed a taste in a certain movie or music or book it did affect you, for some of them. 

Online interactions have enormous scale. You could have thousands of people you interact with. Have tremendous specificity. You can target specific individuals and find them in a way that was very difficult in the olden days. They have a kind of communality, a kind of collaborative feature that was difficult to achieve in olden days, like Wikipedia would be a great example. And they have what we call a virtuality, so for instance, if you’re a man you could have a female avatar. Now you could always pass as a woman in real life too, but it’s much harder in the real life than it is in a virtual world or you can be a disabled person with an enabled body avatar for example, so there is a kind of way in which you can have social interactions online, which was very difficult, if not impossible to have face-to-face. But despite all of that fundamentally I don’t think that these new technologies change human interaction. 

Question: Will the Web ever fundamentally reshape our social interactions? 

Nicholas Christakis: Well I mean ask yourself, I mean do you think that we are different in our sociality after the invention of the telephone as before? You know we’ve only had a telephone for a hundred years. Are we a completely different species than the Victorians? You know, do we have different desires? Are we… No. And now can we communicate more efficiently, transact business and so forth? Yes. And what is also ironic and we discuss this in the book is the ways in which when the telephone was invented and introduced all the same kind of concerns about you know the spread of inappropriate information or people predating on other people or the intrusiveness of the technology. You know the quietness of the dinner hour would be interrupted by the phone ringing off the hook. All of those concerns were articulated just like we talk about the Internet now. And so completely acknowledge that it’s astonishing what the Internet offers. I just am not convinced that the Internet fundamentally is changing how we think or how we really interact with each other. It’s giving us additional ways of interacting, but I don’t think it fundamentally changes friendship, romance, love, violence, all these very deeply human traits.

Recorded March 31, 2010

rnInterviewed by Austin Allen