Social Physics by Alex Pentland is a valuable book. That does not mean it is a good book. It would be a decent at translating academia to the general public if Dr. Pentland weren’t a compulsive self-promoter who seems to use “I” or “me” in every other sentence. (Un)fortunately there is a lot of value buried underneath the poorly organized book culminating in the final third of the book, which is almost well written and definitely thought provoking.
To make the book more bearable I have created an alternative reading order that highlights the most important threads and concepts while cutting out most of the fluff. If you want a review of the book, I have one of those too. All page numbers are from the 2014 hardcover version published by The Penguin Press.
Special topic: Language; pages 19-21 (this can be bookmarked to return to later when unfamiliar terms appear)
Idea Flow; pages 43-46, 58-61, 79-80
Exploration and Engagement; pages 39-42, 77-78, 96-104
Chapters 6; pages 105-107, 116-119
Chapter 7: All of it except the social signals box
Chapters 8 & 9
For the interested, bellow is a series of summaries and explanations for my choices.
Like a classic college essay, the most valuable sentence of the entire book is on the penultimate page:
Throughout this book I have argued that we need to think about society as a network of individual interactions rather than as markets or classes. To accomplish this, I have presented a social physics framework that outlines how the flow of ideas from person to person shapes the norms, productivity, and creative output of our companies, cities, and societies.
Nowhere else in the entire book do I believe Dr. Pentland so clearly articulates what he seeks to, and has, accomplished with his research. Keeping this in mind will make the rest of the book far more intelligible. Pentland is trying to build a theory of society contradictory to both a Marxian class analysis and a neoclassical atomistic (market based) approach.
Why does he have an introduction with this preface? The preface situates the book and its author socially and somewhat intellectually. He is the head of MIT’s Media Lab, which gives him access to amazing data sets, a lot of intelligent people, and the resources to actually use these resources. The introduction is largely Pentland trying to summarize the products of his theory while acting like almost no social science in the last three hundred years produced valuable knowledge.
Special Topic: Language (page 19-21)
Pentland does something that I believe should be mandatory for any text that is even semi-academic: he provides a glossary. In it he defines a variety of terms he will use throughout the book, and as far as I can tell he is pretty consistent with his definitions and meanings. You don’t necessarily need to read through this, but you should bookmark it for translating jargon later.
Chapter 3, selected parts (Pages 43-46, 58-61, 79-80)
There are interesting parts of this chapter, but they are better read later. The sections I note have Pentland summarizing his thoughts on collective rationality and how idea flow determines, well, everything in society.
Idea flow is a central component of Pentland’s theory. It is what explains why some organizations flourish and others, with seemingly more talented and intelligent members, struggle. He defines it as:
The propagation of behaviors and beliefs through a social network by means of social learning and social pressure. Idea flow takes into account the social network structure, the strength of the social influence between each pair of people, as well as individual susceptibility to new ideas.
In other words, the way in which people’s minds are expanded, shrunk, and otherwise change within and because of a social setting. This is a useful heuristic for thinking about human interactions, but Social Physics aims to provide more than a heuristic, it aims to give a general theory of how idea flow works. The main drivers of idea flow is what Pentland calls social learning and social pressure;
Social learning consists of either: (1) learning new strategies (e.g., context, action, outcome) by observation of other people’s behavior, including learning from memorable stories; or (2) learning new beliefs through experience or observation.
Social pressure is the negotiating leverage one person can exert upon another, which is limited by the exchange value between them.
If you want more on social learning right now, check out pages 25-32. For more on social pressure, see pages 62-65, 69-70 I think it is more useful here to move to exploration and engagement, the actual mechanisms behind idea flow.
Exploration and Engagement pages 39-42; 77-78; 96-104
Engagement is social learning, usually within a peer group, that typically leads to the development of behavioral norms and social pressure to enforce those norms. In companies, work groups with a high rate of idea flow among the members of the work group tend to be more productive
Exploration is the process of searching out new, potentially valuable ideas by building and mining diverse social networks. In companies, work groups that have a high rate of idea flow from outside the work group tend to be more innovative.
Together, exploration and engagement are the processes by which a group becomes more creative and, in so doing, more productive.
These two processes are somewhat at odds, however. One demands increased idea flow from outside of the group—which assumes members of the group will be step away from the group and explore other thought; but for those ideas to help the group the members of the group need to engage with one another or else the ideas flowing into the group won’t be able to flow amongst the members. Flipping it around, without new ideas entering the group it is liable to become an echo chamber of like-minded and inadvertently myopic thinkers. Thus productive groups oscillate between exploration and engagement.
Chapters 6: 105-107, 116-119
In this chapter Pentalnd applies his theories in an integrated way, and I think these passages are a good way to think about the theory in a less abstract fashion. They also work as introductions to social intelligence, a key product of Pentland’s work (even if these findings are quantifying things social scientists already know). In essence, social intelligence is the intelligence of a group and in Pentland’s research it is more predictive of group success than any other single factor, including the talent and intelligence of individual group members.
Chapter 7: All of it (except the social signals box)
This is where the book gets fun. In this chapter Pentland demonstrates the power of social pressure to rapidly create cooperative organizations. He also does a good job of explaining the nuanced, but vital, differences between his approach to social organization and motivation and the “old school” atomistic thinking that is more present in the world of business and economics.
Conclusion time! Now that Pentland thinks you have a solid grasp of social physics he provides a variety of new ways to approach social interaction as products of his theory. Most notably, he talks about three things to maximize in a society: social efficiency, operational efficiency, and resilience. Social efficiency is the extent to which goods and resources are distributed optimally throughout a society. Operational efficiency is how well the infrastructure of the society works to maintain social efficiency. That is, outside of times of crisis resources should be distributed optimally (social efficiency) and society should be stable (operational efficiency). Resilience is how well society responds to, and adjusts in, crises. Resilience is what allows a society to reestablish high social and operational efficiency, though this likely requires changes in the infrastructure of the society to resolve the tensions that generated the crisis.
Chapters 8 & 9
These chapters outline a variety of challenges facing cities, why cities can be so innovative, and some fascinating conclusions from the plethora of data MIT Media Lab can analyze. Their structure is similar; Pentland will use his datasets to learn something about cities, and then he will use that data and social physics to provide interventions and conclusions. The goal of this entire section is to imagine a way to design better cities. If you worry about some of the big data implications in this section, don’t worry, so does Pentland. He addresses these concerns in Chapter 10.
There are a lot of big data implications in some of the things Pentland proposes and, like all tools, there is no guarantee the tools for monitoring and influencing people produced by social physics are beneficial. Here he tries to grapple with the downside of this project.