How does the way information is organized in networks affect what we can achieve as human beings?
In this issue of d4e we ask, “Where are the old borders and boundaries of our countries, firms and financial systems reshaping? How are the lines of our communities and even our ideologies shifting? How does cooperation work?” Network science applies research from information theory, game theory, complexity theory and physics to understand all of these things, which boil down to how information organizes itself into networks to shape the world we live in.
César A. Hidalgo, head of the Macro Connections group at MIT Media Lab and author of Why Information Grows, puts it a little more simply: No matter what the world is doing, it is organizing information. It’s this order we’re trying to understand when we want to build the systems we need to make our way in the world.
d4e: Why is something like information theory so important in understanding how the world works?
César Hidalgo: As a kid I wanted to study physics. I thought everything was about matter and energy and equations or formulas or some underlying laws. As I came of age and began to look at things that were more complex, I started to realize in some ways the universe is made of more or less three things: Some things are what we can actually call things — such as matter — and what we need to move around — energy. But the thing that was less obvious but more important was the way that these things were arranged, what we call order. And I realized that that order was information, and that was what it was all about, because the difference between the society we live in now and the society in which people lived thousands of years ago was not in the atoms that are on earth or the amount of energy that we’re receiving from the sun. The difference is in the way in which we have ordered everything.
I started to realize that the only way in which I could understand all of things from a unified perspective was to think about them in terms of information.
d4e: In your book, you talk about a perpetual battle between order and randomness. Must randomness be battled?
CH: Absolutely. In some way if we don’t battle randomness, we die. Every day we’re consuming energy in the form of food. Our body uses that food to ensure that all our cells remain organized and our organs continue to function.
The universe is not very friendly toward order. It has the tendency to average things out. If you averaged yourself out, you wouldn’t work that well, would you? You don’t want the neurons in your head averaging out with your toenails. You want all of those things to be separate, because the separation of those functions is what allows you to have the complexity that you have as a human being.
“We live in a world in which everything that’s worth doing is done in the context of a social network — a team.”
d4e: This must be where boundaries matter, because information — life, even — tends to be a distributed system.
CH: It’s a very distributed system, that despite being distributed, only works as long as the parts in it can cooperate effectively. It is distributed but you have very clear boundaries between the parts of the system.
d4e: How do you see those boundaries shifting and changing how cooperation works, particularly for the firm, as we move into the network age?
CH: That’s a great question because in some ways we live in a world in which everything that’s worth doing is done in the context of a social network — a team. We need teams to play sports, but we also need teams to manufacture cars, manage logistic chains across the world — we need teams for everything. And the reason we need them is quite simple: We each have a finite capacity to accumulate knowledge and know-how, and the only way we can accumulate the large amount of knowledge and know-how we need is by distributing it in a network of people that collaborate and reconstitute it into this super-organism that is the team.
The problem is that creating teams that are increasingly larger is hard, and if you think about it, the history of modern civilization has been about developing tricks that allow us to make larger and larger teams. One of those tricks, for instance, is to speak the same language.
Other tricks involve the development of relationships and trust between people — having interactions with people, having expectations about their behavior. Or the development of things like currency, that allow you to interact with strangers in exchanges that are more or less standardized. Then modern institutions like police and the rule of law would also let you collaborate on a larger scale, by having cities in which people are not afraid their neighbors are going to come and club them to death and take their things away just because they were in a bad mood.
But there is this idea that there is some sort of collaborative equilibrium that is not just spontaneous, but that is also being enforced by the people in that system. So, the way the that world works is that we’re individually kind of limited, and the way we overcome that limitation [is] by coming together in groups — and to do that we have to develop increasingly more sophisticated tricks which go from something as basic as language to the level of firms. Nowadays, we live in a time when we’re finally getting to achieve cooperation at a global scale, which is something I find very fascinating and very recent.
d4e: What happens to trust and information flow when organizations become really large?
CH: When organizations grow really large, what I find that ends up happening is that there’s a lot of people just meeting other people all the time, and at that point you hit a ceiling: The organization cannot process more information because people are at capacity. The people who are overcapacity — because they’re getting more inputs than they’re able to process — become roadblocks to the flow of information.
So when you have very large organizations, especially large bureaucracies like the government, you have systems that are so laden by their size that they’re unable to process vast amounts of information. Then networks keep growing, but their ability to process information doesn’t, because now, just for the system to remain in communication requires the full time and attention of everyone involved in it.
d4e: If knowledge scales through networks, what do we do? The internet is a global tool and bureaucracies become bogged down in connecting with themselves. Large governments are just too large. Is that right? Or are there companies just as large as governments that are highly dynamic, so size doesn’t necessarily mean stasis…Are there things one can do to keep one’s network agile and adaptive as it grows?
CH: Knowledge is embodied in networks of people, and the reason for that is simple: People’s capacity to accumulate knowledge is finite, so our only escape to our limitations is to form networks that have a larger capacity to accumulate knowledge than individuals. The problem is that our capacity to interact with others is also finite. We can only meet a few people every day, and we speak at about 100 to 150 words a minute, making a one-hour meeting equivalent to just one or two book chapters. So the networks we can form and effectively manage are also constrained in their capacity to accumulate knowledge by our finite capacities to interact.
So what are the solutions we have available to help transcend these fundamental limitations? The last 100,000 years of social, economic, and technological development can be understood — in part — as an uphill battle to develop institutions that can help us cooperate with a larger groups. These are institutions that help reduce the cost of interacting with others. Think of the development and spread of languages, or religions. Languages clearly help us communicate better with those who share the same language, and religions — as Yuval Harari explains in his book Sapiens — help reduce the cost of interacting with strangers that share the same faith, as a common religion implies shared moral codes, rituals, and expectations. More recent inventions involve money, roads, and even more recently, communication technologies such as the printing press, film, radio, television, and the internet.
Yet, inventions — from religion to the internet — are not the only innovations that help us build the networks we need to accumulate the volumes of knowledge that individuals cannot accumulate. Think of trust, and that feeling of familiarity and confidence you have in others, and how trust makes communication easier and more direct. Think of liberal and inclusive institutions, that focus on the sameness of different people, regardless of religion, ethnicity, or sexual preferences. These institutions help us more effectively build the networks we need to accumulate knowledge.
So to your point. I don’t think there is a silver bullet to make an organizational network more adaptive and capable, but there are many factors that can affect the capacity of a network — from the technology they use, to the institutions that they endorse.
d4e: What insights has Pantheon given you about shaping collective memory and culture?
CH: So far with Pantheon — a dataset on globally famous biographies that we use to quantify collective memory — we have uncovered two main findings. The first one was a very strong correlation between the number of globally famous biographies produced by the speakers of a language, and the centrality of that language in the global network of translations (paper here, video here). The second one–which is in a paper under review–is that we find that all major changes in communication technologies, like the invention of movable type printing, the introduction of journals and newspapers, the introduction of film and radio, and the adoption of television, are accompanied by changes in the number of people that we remember from a time period and the occupations of those people (pre-print here).
d4e: What has Immersion taught you about the formation of successful networks within companies?
CH: Immersion is a tool that allows people to visualize the networks they weave while communicating in email. So far, we have only deployed immersion for individuals, but we have plans to create a new version focused on groups. At the individual level, however, what Immersion has taught me, is that the best way to maintain relationships on the long run is to include people into groups that you already have. This was something that was somehow to be expected. In 2007 I published a paper that looked at a year’s worth of mobile phone data and found a strong connection between the longevity of links and their position in the social network. Nevertheless, as someone who has been using Immersion for more than three years to monitor my email based social network, I am now very aware that if I want to keep someone in my network for long, I need to connect that person to others in my circle.
d4e: Do the mappings you have done of people’s perception of urban environments (PlacePulse and Streetscore) offer insights for managers in creating spaces that encourage high trust and collaboration both virtually and in the world? Given that your work predicts crime, can it also predict hotbeds of success?
CH: Our work on Place Pulse and Streetscore focuses more on the street and neighborhood scale, so we do not have much to say about the spaces within buildings. Nevertheless, we have found evidence suggesting that the aesthetics of urban environments affects human behavior. In a recent collaboration with Bruno Lepri, Marco de Nadai, and others from the Fondazione Bruno Kessler in Italy, we found that places that are perceived as safer are more active (as proxied by mobile phone activity traces), than what would be expected from those places’ population density, employment density, and distance to the center. Although these results are observational, rather than causal, they suggest that people may have enough of a preference for the aesthetics of urban environments for these to modify their walking patterns within a city.
d4e: If economic complexity is key to long term growth for countries, is it the same for firms? If so, how should we think about measuring it?
“Trying to get everyone to know how to do everything is probably not the best strategy for a firm.”
CH: This is a good question because it goes to the heart of the problem of multiple scales, or what in statistics is known as the ecological fallacy. An ecological fallacy occurs when one assumes that what is good for the group is good for the individual, when that is not always the case. In the case of economic complexity — understood as the ability of an economy to make a diverse set of sophisticated goods — what is good for the economy as a whole, is not necessarily good for a firm or an individual. Consider the US economy as an example. It is a very diverse economy whose exports range from soybeans to industrial machines. This complexity gives the US economy a global strength, but it does not mean that firms in the soybean industry should try to diversify into textiles, machinery, and online search. Diversification at a large scale is a consequence of specialization at smaller scales. The same is true in a firm. In a company there are many people that know how to do many different things. Trying to get everyone to know how to do everything is probably not the best strategy for a firm.