.jpg)
Networks, Weak Ties, and Thresholds
August 8th, 2019
Few living scholars have had the influence of Mark Granovetter. In a career spanning almost 50 years, his seminal contributions to his own field of sociology have spread to shape research in economics, computer science, and even epidemiology.
Granovetter is most widely known for his early contributions to social network analysis—in particular his 1973 article, “The Strength of Weak Ties.” In that paper, Granovetter demonstrated that, because of the way social networks evolve, “weak ties” between people often form bridges between clusters of more strongly connected individuals and thus serve as important conduits of novel information. This surprising finding has proven to have important and enduring implications for a diverse range of fields. The paper remains one of the most cited social science articles of all time.
Granovetter was also key to the development of the “New Economic Sociology,” who argued—against the disciplinary divide separating economics and sociology—that economic action is another form of social action. This implies that individuals engage in economic activity to pursue non-economic goals like status and power, and it also means that economic activity occurs within a web of social relations: people do business with existing family, friends and neighbors, and these roles and their associated meanings shape behavior. Granovetter aptly describes the research value of this framework in his widely influential 1985 work, “Economic Action and Social Structure: The Problem of Embeddedness.”
Finally, Granovetter laid the foundations for what is now known as “threshold analysis.” By examining differences in individuals' willingness to act before others have acted similarly, threshold analysis can determine whether such actions generate behavior cascades that lead to, for example, riots and effective collective action. Equipped with modern computing power, a new generation of scholars is continuing his work on threshold modeling, some of whom are working to wed it to social network analysis.
We met Granovetter at his home in California, where we spoke about his career, the state of interdisciplinarity, and the legacy of his research.
An Interview with Mark Granovetter
Stephen Nuñez: It would be great to start with the story of how you gained an interest in sociology.
Mark Granovetter: When I applied to college, my admissions essays said I wanted to go to law school. As soon as I started freshman year I lost that ambition. I took a world history course and, as you do when you’re a freshman, I fell behind by about 1000 years’ worth of readings. So I had to do 1000 years of reading in about a week, which I found to be an exhilarating experience. It was as if I could see the march of history unfolding before my eyes. I became a history major very early on, and I loved being a history major. I decided that I wanted to stay in college forever, and in order to do that I had to become a professor.
It was very simple and logical, but to become a professor you need to get a PhD, and to get a PhD in History would have meant that I had to become a historian of 18th century France, or 17th century Russia, or some very restricted subject like that. I realized that I was interested not so much in why there was a French Revolution, or why there was a Russian Revolution, but why there were revolutions in general. I was interested in these big questions which were not the kinds of things that historians asked, but they were the kinds of things that sociologists found out.
So I found myself reading sociology in my spare time, and I imagined that if I became a sociologist I would have a lifetime of fun. It doesn’t work like that, because when you’re doing it professionally it becomes more of a job. I applied to graduate school in sociology, though I only applied to Berkeley, where Seymour Martin Lipset was, and Harvard, where the department was at that time called “social relations.” It had been designed by Talcott Parsons in the ‘40s, and it included sociology, social psychology, developmental psychology, clinical psychology, and social anthropology. Ellen Granovetter, my wife, worked in Skinner’s lab, where she worked on the pigeon experiments.
But social anthropology was distinct from the so called “stones and bones”—the archeological part. Parsons’ hope was that you could bring all of the social studies from different disciplines and they would unify in the department of social relations. By the time I was a graduate student, the faculty and the different departments were no longer interested in talking to one another. But for the students it was great taking courses from separate disciplines.
Sara Constantino: I’m curious to hear where you think the trajectory of interdisciplinarity has gone.
MG: As I mentioned, there was this social relations department at the time, which was dying. Today, we have a number of interdisciplinary institutes, but they’re usually focused on some particular set of issues, like international security or public policy. They bring together people from different disciplines that have a unified interest, but the departments themselves are not getting particularly close to one another.
The reason for that is that disciplines are the unit by which resources are allocated in universities, and the way that disciplines get resources is by saying that they have special expertise, and the more they work with people from other disciplines the less they can claim that they have special expertise.
SN: Going back to your intellectual development, I know that there was a sort of anti-functionalist revolution against Parsons, which you were a part of. What else influenced you?
MG: With respect to the debate between functionalism and anti-functionalism, I can say that as an undergraduate student of Carl Hempel I engaged very deeply with Parsons and was conscious of the limits of his work by the time I entered graduate school.
The other thing I came to graduate school with was an interest in the details of what happened on the ground. There’s a book called The Great Fear of 1789, by the great French historian Georges Lefebvre, in which he draws maps linking the spread of the revolutionary panic to postal routes. That was fascinating to me—I gained an interest in causal mechanisms from that book, and I also gained an interest in what I didn’t know at the time was called networks.
My dissertation advisor was Harrison White, who was particularly interested in social networks, and listening to him speak, I realized that the thing I was interested in had a name, and that there were all kinds of new theories about social networks being developed. So I intuitively situated myself in that scholarship, and for my dissertation I wanted to do something that demonstrated the importance of social networks. I had two ideas for my dissertation: one was to study how people found spouses, and the other was to look at how people found jobs. I settled on jobs because I was also interested in inequality and this gave me an opportunity to study the labor market. I tried to find already collected data that I could use, but it wasn’t there. I could have either changed my topic, or collected my own data, and I chose the latter. I created a survey and interviewed people in Newton, Massachusetts. And that’s how I almost accidentally wandered into what came to be called—but was not yet called—economic sociology.
SC: Was empirical work in sociology common at that time, or was this a departure from the norm?
MG: It was a departure. Ethnographers would of course immerse themselves in settings and talk to people, but actual random sample surveys in which the researcher chooses a sample and interviews them are still pretty rare. In retrospect, I realize that’s because it’s very hard!
The people I chose to interview had recently found a new job in Newton, and I had narrowed my focus to white collar men per my committee’s recommendation. I sent them letters which said “We’ve drawn your name in a random sample. We’re really interested in the mobility of people like you, and I’m coming to your house to interview you in the next few weeks.” The actual letter and my methods for finding the subjects are all in the appendix to my book.
The amazing part for me looking back is that I was what, 25 years old? I just assumed you could do anything if you wanted to, and you could tell people you’re coming, and they would open up and talk to you. Amazingly, that’s what happened! I interviewed 100 people, and of all of the letters I sent out only 1 person wrote back saying he was not interested. For all the rest, I just went to their houses in the evening, after dinner but before sunset, and said “I’ve come to interview you.” Nobody threw me out, many people said “sure come in, we’ll talk,” and we would talk.
It turned out that networks were the most important factor in determining what kind of jobs people found. While writing the dissertation I had read a lot of what economists were saying about all of this, and found that their theories had no relationship to the decisions that real people make. So I was sucked into a critique of labor market economics, which gradually pulled me deeper into a critique of economics in general from a sociological perspective.
In 1937, Talcott Parsons wrote that economists talk about how people achieve their goals, while sociologists talk about where they get them. However, he later tried to integrate the two together again. I don’t think he realized that the book, with its diagrams, boxes, and lines, ended up convincing people that the two disciplines had nothing to say to each other. In the 70s, we took a different approach and tried to use sociology to see more of what was missing in economics and how you could actually reassemble it in a more plausible way. I published a paper in ‘85, called “Economic Action and Social Structure: the Problem of Embeddedness”, which argued that it doesn’t make sense to talk about economics as if people didn’t know each other.
SN: Your most famous paper is, of course, “The Strength of Weak Ties,” which obviously draws from your dissertation research. But what else inspired it? Does it, for example, draw from cognitive balance theory, which argues that it’s unusual to not be friends with your friends’ friends?
MG: I put the theory of weak ties together from a number of things. I learned about hydrogen bonding in AP Chemistry in high school and that image always stuck with me—these weak hydrogen bonds were holding together huge molecules precisely because they were so weak. That was still in my head when I started thinking about networks.
In tandem with that was the paper by Anatol Rapoport and Willian Horvath, in which they asked junior high students in Ann Arbor Michigan to name their best friends, all the way up to their 8th best friends. They found that the networks of first best friends were much smaller than those of 7th or 8th best friends, because responses to the first and second best friends tend to fold back. Basically, Rapoport and his colleagues already knew that weak ties were strong, but they didn’t understand why that was interesting.
So I had hydrogen bonds in my head, I had these friend graphs in my head. I’m interviewing people about how they found their jobs and they tell me, “Joe helped me find this job,” and I would respond, “your friend Joe?”, and they would say “oh no, he’s only an acquaintance.” I heard this over and over again: acquaintances, not friends. I was just getting hit in the face from all directions about how important weak ties are. Then Stanley Milgram did the chain letter experiment, and he figured out that people are connected to other people in distant places in their social network through a chain of acquaintances. If you try to connect people through close friends it bounces back and never gets there.
There were so many ways in which these dynamics were being discovered, but no one had pieced them together yet. So I wrote “The Strength of Weak Ties,” which initially was rejected vigorously by the American Sociological Review. The original draft was titled “Alienation Reconsidered” and the journal sent it to European alienation theorists, who thought it was terrible because it had nothing to do with Marx.
SN: Didn’t you actually scan and upload the rejection letter a few years ago?
MG: I did, I like to show it to graduate students. I was 25 when I submitted it, and I was crushed by the rejection. I didn’t get back to it until a few years later, and I got rid of alienation in the title and rewrote the text. I sent it to the American Journal of Sociology and got a revise and resubmit. Once it was accepted it just took off, I was shocked. Today it’s one of the most cited articles in sociology, with upwards of 52,000 citations.
SN: Can you tell us a little bit about your work on threshold models, and the renewed interest in that approach?
MG: I ended up becoming an economic sociologist, but I’ve always had an interest in politics. In 1971, I came across a paper by Thomas Schelling on residential segregation. In the paper, he developed models that indicated that if people had even just a slight preference to live in neighborhoods which were mostly home to people of their own skin color, each of these slight preferences feed on each other and generate huge phenomena.
This idea was really interesting to me, and it related to my intuition that sociologists often wrongly assumed a linear model, which holds that the size of outcomes is roughly proportional to the size of their causes. In 1978 I published a paper called “Threshold Models of Collective Behavior” in which I talked about models of collective behaviour in which slight differences in the distribution of people’s individual thresholds would lead to huge differences in outcome.
Riots are my simple empirical example for this idea: if people’s threshold for joining a riot is determined by how many people have joined the riot before them, such that there might be someone with threshold 0, someone with a threshold of 1, and so on and so forth in a uniform distribution from 0-99, then the 0 person does something which activates the next person and so on until all 100 people are in on the activity. But if you took out that person with threshold 1, or changed it to threshold 2, then the first person breaks the window but the action goes no further—the outcome is completely different.
SC: What are your thoughts on recent papers that apply thresholds and tipping points in policy spheres to influence people’s behaviors? The idea is that we can influence the likelihood of adopting birth control, or norms around female genital mutilation, by targeting nodes in social networks and trying to change values.
MG: I don’t think we’re at the point yet where we can confidently determine things like that. There’s a lot of discussion about whether we can intervene in social networks and make a difference: are there certain nodes which might encourage adolescents to smoke less, or engage in less risky sexual behavior? Can we identify the right nodes in the network to target?
If you target people with high centrality in a network, you might be able to have more influence. You also have to integrate it with thresholds, which complicates things. I don’t think we know how to integrate research on networks and thresholds yet. People like Damon Cintola, Nicholas Christakis and James Fowler are making some progress on that, I think, and some of it might derive from the work I did earlier. But I’m not really engaged in doing this anymore.
SC: How do you think new data and computational capacities are changing the nature of research in social networks?
MG: Today, we can do social network research with multiple datasets, each of which has many millions of elements. We never imagined that that was possible. One of the propositions in my original paper was that a tie that connects people who otherwise wouldn’t be connected without a long chain must be a weak tie. There was an interesting paper in Science Magazine last year which showed that although the principle of weak ties is valid over networks of most sizes, when you look at networks which are really enormous and take really long chains those bridging ties are actually strong ties. What does that mean? No one knows yet. There are always new surprises.
Source: phenomenalworld.org
Comments