Story Nine

　In this series of Episodes, we have discussed the inequality that causes division in modern society (Episode 1), the problems of greedy capitalism that imposes monopolistic dominance as its root cause (Episode 2), the future potential of cooperative enterprises that have a long history of mutual support (Episode 3-Episode 5) and have a proven track record in Europe (Episode 6-Episode 7), and the need to re-examine the values of corporate organizations (Episode 8). It has become increasingly clear that the key lies in the autonomous realization of the construction and operation of networks, such as connections between trustworthy people and publicly owned social infrastructure like electricity, telecommunications, and logistics. Therefore, we will first provide an overview of network science (also called complex networks), which emerged around the beginning of the 21st century, derived from the science of complex systems (known for chaos and fractals). Cutting-edge scientific insights will help to promote a fundamental understanding of these problems in modern society and guide us toward solutions.

"This book has a very simple purpose: to get readers to think about networks. This book will explain how networks come into being, and describes what they look like and how they evolve. It also shows how nature, society, and business appear from a network perspective. Networks are a new framework for understanding a variety of issues, from online democracy to the vulnerabilities of the internet and the terrifying spread of viruses. (omitted) You will see that the internet, often considered entirely man-made, is actually more like living organisms and ecosystems, and follows the fundamental laws that govern all networks. (omitted) As you apply your understanding of one network to the next, you will be amazed by the similarities between diverse systems such as economics, cells, and the internet. It will be an awakening journey spanning multiple academic disciplines." (Quoted from book [9-1], pages 16-17).

"Now, in the 21st century, a fundamental and universal concept has emerged with the deep relevance to real life. The concept is 'network'. (omitted) Basic knowledge of networks is essential for responding to a fluid society, avoiding destructive crises, it can be said to be an essential tool for thinking about the society of tomorrow." (Quoted from the Japanese version of book [9-1], page 326).

　At the end of the 20th century, large-scale data analyses were conducted to understand the nature of the many networks that exist around us—that is, how they are connected. These analyses revealed common structural patterns in social networks such as personal relationships, inter-company transactions, and academic citation relationships; technological networks such as power grids, communication networks, and air networks; and biological networks such as gene and energy metabolism reaction systems and food chains. Of course, each of social, technological, and biological networks has different components and purposes. For example, individuals and companies act as nodes, with friendships and business relationships as links for hobbies and economic activities; power plants, substations, and communication hubs act as nodes, with power lines and communication links for transmitting electricity and information; and proteins and species act as nodes, with biochemical reactions and predation relationships as links for involvement in cancer development and ecosystem maintenance. However, when we focus on the way they are connected, there are commonalities. This is because there is a significant inequality in the number of links between nodes, resulting in a structure where links are concentrated in a very small number of hubs (similar to the unequal distribution of wealth mentioned at the beginning of Episode 1). However, no matter how you choose two nodes within the network, the shortest path (by connecting links) is short and efficient. In other words, even without direct connections, the network structure allows for quick communication with minimal intermediaries.

"On the web, a small number of hubs attract the majority of links. This discovery triggered vigorous research in various fields, yielding surprising results. Today, we know that Hollywood, the web, and our society are not unique in any sense. For example, even in networks of molecules linked by chemical reactions, there are hubs within cells. A small number of molecules, like water and adenosine triphosphate (ATP), act as the cellular rod-steigers, participating in a vast number of reactions. In the internet, that is, the network of physical wiring connecting the world's computers, a small number of hubs play a crucial role in protecting the internet from failure. (omitted) Most of the large and complex networks that scientists have studied have hubs. Hubs are components of this intricately interwoven world. It exists quite naturally." (Quote from book [9-1], page 93).

"The 'small world' is a property inherent in networks in general. Small distances are not unique to any particular group in our society, nor are they a strange property unique to the web. It's a property that most networks around us possess. And this property is rooted in the network structure. That is, there's a structure that allows you to reach a vast number of web pages and friends by following just a few links." (Quote from book [9-1], page 62).

　Furthermore, it was found that this common structure follows a power law when viewed in terms of the frequency distribution of the number of links connected to a node (the number of links connected to a node is called the degree, and its frequency distribution is called the degree distribution). In fact, the power law indicates that even very large numbers can occur rarely, as seen in the frequency of major earthquakes, stock price fluctuations, the diameter of meteorites impacting Earth, the number of copies of best-selling books, the number of telephone calls, etc. Frequency distributions of coastlines and river lengths are familiar concepts in 20th-century fractal physics and are related to the term 'scale-free' network."

"Scientists have learned over the past few decades that in nature, there are distributions that follow a power law, not a bell-shaped curve. These power law distributions differ significantly from the bell-shaped distributions seen in height distributions, for example. (omitted) While studying the network of Hollywood actors, we noticed that it was governed by a simple mathematical relationship. The number of actors with k links decreased according to a power law. Later, we found that Erdős's co-authorship network also followed this law. It also became clear that in intracellular webs, the number of molecules interacting with k molecules decreases according to a power law. (omitted) This heterogeneity is characteristic of networks whose frequency distribution follows a power law. Most real-world networks have the characteristic of having a large majority of nodes with few links and a handful of hubs with a vast number of links. The power law is a mathematical expression of this." (Quoted from book [9-1], pages 98-102).

　There must be a scientific basis for the regularity of following a power law. In fact, mathematical analysis has revealed that networks are formed by a principle of connection called 'preferential attachment', similar to the principle of universal gravitation. Following this principle, links to hubs are added consciously or unconsciously, prioritizing efficiency (because routes become shorter and more efficient via hubs). For example, in Japanese domestic air travel, airlines want to increase the number of flights to Haneda Airport. This is because Haneda flights allow passengers to reduce the number of transfers during their travels, leading to a higher expected ridership and increased profits for an airline company. Furthermore, it seems that connections aren't made under someone's orders, but rather based on immediate profits or a general sense of efficiency, in suggesting that the network is self-organizing by the selective connections to hub nodes as proportional to these numbers of links. This is essentially the same principle as the wealth inequality mentioned at the beginning of Episode 1. To be more precise (by replacing people with business partners, etc.), 'preferential attachment' is also effectively carried out through link copying, such as 'being introduced to a friend of a friend you happened to meet, and becoming friends yourself'."

"The important thing is that the web pages we are drawn to link to are not ordinary nodes; they are hubs. The more famous they are, the more links they acquire. The more links they acquire, the easier it becomes to find them on the web, the more famous they become, and eventually, unconsciously, we begin to link to nodes we know. These known nodes are none other than those that have acquired a large number of links on the web. In short, we are selecting hubs. (omitted) Hollywood is also governed by preferential attachment. Producers whose job it is to make money from movies know that using stars will sell movies. (omitted) Manipulated by this law, we link to nodes that already have many links with a high probability. (omitted) Thus, preferential attachment leads to the phenomenon of 'the rich get richer'. Nodes that already have many links acquire an unfair number of links at the expense of newcomers. From this 'the rich get richer' phenomenon, the laws that should be seen in real-world networks are naturally derived." (Quoted from book [9-1], pages 124-129).

"A router with wider bandwidth will have more links. Therefore, it is natural that network engineers will be drawn to access points with more links when selecting link destinations. This simple fact may give rise to preferential attachment. While it is unclear if this is the only factor creating preferential attachment, there is no doubt that preferential attachment exists on the internet." (Quoted from book [9-1], page 219).

　However, unfortunately, many real-world networks are extremely vulnerable; even a small number of targeted hubs can cause the entire network to lose connectivity and quickly fall apart. When it falls apart, the essential function of the network—transmitting and transporting information and goods—is lost, leading to system collapse. This vulnerability exists not only in technological infrastructure such as power, communications, and logistics, but also in socio-economic systems connecting human relationships and businesses, and even in ecosystems. Thus, the fundamental problem of extreme concentration has been scientifically revealed. In other words, the existence of hubs (which unfairly concentrate and monopolize links) is the root cause of fragility and fragmentation. In economic systems, of course, hubs are the wealthiest.

"Inspired by hackers targeting the largest hubs on the internet, we embarked on a new experiment. (omitted) First, we removed the largest hub, then the second largest, and so on. The impact of this attack was very clear. Removing the first hub did not cause the system to collapse, because the remaining hubs held the network together. However, removing just a few hubs had a clear impact. A large group of nodes was cut off from the main cluster and left the network. When more hubs were removed, the network spectacularly collapsed at a certain point. (omitted) This vulnerability to targeted attacks is an inherent property of scale-free networks. In fact, our group's study of yeast cell protein interaction networks found that removing several high-link proteins resulted in a spectacular collapse. Ecologists have also observed similar collapses in food chain networks when high-link nodes are removed." (Quoted from book [9-1], pages 168-169).

　The following seems to illustrate the dangers of modern society, where a select few giant corporations control everything from wealth to resources.

"In recent years, stimulated by the network renaissance in physics and mathematics, the power of networks has become apparent in all areas, from corporate structures to markets. For example, It has been found that a thin network of a handful of powerful individuals controls a significant portion of the important personnel decisions at the 'Fortune 1000' (the top 1000 large American companies selected by Fortune magazine). Furthermore, in the biotechnology industry, a company's success depends on its collaborative network, and its ability to adapt to rapidly changing markets depends on its internal network structure. The success of marketing also depends on whether it can leverage the network nature of the consumer base." (Quoted from book [9-1], page 286).

　On the other hand, apart from scale-free structures with a few hubs, many real-world networks have several cluster structures (also called modular or community structures) where nodes are tightly coupled to one another in a cluster. Unfortunately, the internet, for example, is increasingly modularized along with the rise of massive hubs. Apart from the advantages of obtaining information through loosely connected, weak ties, we will discuss the problem of 'as long as we're okay' inherent in modular structures with strong, tightly connected bonds in Episode 11.

"In 'the Strength of Weak Ties,' Granovetter argues something seemingly absurd: that as far as finding a job, obtaining information, opening a restaurant, or creating trends is concerned, weak social ties are more important than strong friendships. (omitted) Society is structured to have highly interconnected cluster structures—that is, tightly connected friendships—where everyone knows each other. These clusters are linked to other clusters by a few external links, preventing them from becoming isolated from the outside world. (omitted) Weak ties, or acquaintances, are bridges to the outside world. This is because mere acquaintances live in different worlds than us and therefore have different sources of information than close friends." (Quoted from book [9-1], pages 64-66).

　Incidentally, independently of network science, the concept of 'resilience' has been gaining attention in the fields of systems engineering and environmental ecology for about a decade. Resilience refers to the flexible recovery power, like bamboo, that is essential for maintaining the connectivity necessary for the transmission and transport of information and goods, particularly how to revive systems from major disasters and malicious attacks.

"Scientists, politicians, engineers, business leaders, and activists around the world, across seemingly unrelated fields such as economics, ecology, political science, and digital networking, are posing a common fundamental question: Why do some systems fail while others recover? To what extent can a system absorb change and maintain its integrity and purpose? What characteristics enable systems to adapt to change? In these turbulent times, how can we build better buffers for ourselves, and for our communities, businesses, economies, societies, and the planet? (omitted) Ultimately, intervention strategies, regardless of the system, suggest a strategy of resilience. (omitted) Throughout this book, we will examine the resilience of both systems and individuals. Therefore, borrowing terminology from ecology and sociology, we define resilience as the ability of systems, businesses, and individuals to maintain their fundamental purpose and integrity when faced with extreme circumstances." (Quoted from book [9-2], pages 8-10).

　One way to enhance resilience is through feedback mechanisms to maintain functionality. For a network to withstand major disturbances (such as catastrophic events or malicious hub attacks), decentralization, or autonomous decentralization, is an effective means.

"Another way to enhance resilience is to decentralize or separate the material requirements underlying the system, that is, to diversify the resources needed to complete a specific task." (Quoted from book [9-2], page 15).

　Furthermore, while 'tolerance', 'reliability', 'redundancy', and 'response and recovery' are listed as four elements constituting the resilience of social infrastructure, these remain conceptual discussions and do not provide concrete design guidelines. On the other hand, (at least in a broad sense) resilience does not necessarily mean returning to the original state, but rather aims to improve it if possible. While the importance of 'tolerance', 'reliability', and 'response and recovery' is clear, regarding redundancy, it is better to have some redundancy than to have no margin for error due to an overemphasis on efficiency (such as cost reduction). This refers to the idea that a system with some redundancy, even if it means tolerating some inefficiency, is better because it allows for partial substitution. However, this is different from backup switching. Specifically, we will discuss in Episode 11 what a network should look like to achieve optimal resilience and how to improve it. At the very least, it is clear that reverting to the current network after a disaster or attack is undesirable because it inherits the vulnerabilities of a scale-free structure.

"Perhaps most counterintuitively, but resilience does not necessarily mean 'recovery' to the original state." (Quote from book [9-2], page 19).

　To reiterate somewhat, although not anticipated during its initial development around 1970, the internet's weakness lies in its hubs. On the other hand, the internet is robust against unexpected node failures and is an excellent distributed system that autonomously finds alternative routes through a mechanism called routing.

"The internet is a classic example of a world where robust yet fragile dynamics are at work. From its inception in the 1960s, funded by the United States Department of Defense, the internet aimed to solve a single problem: the establishment of a communication network that could withstand catastrophic situations. (omitted) Therefore, the engineers who developed the internet embarked on developing a system that could detect equipment failures inevitably caused by external attacks and automatically reroute communications. (omitted) However, the current internet is extremely vulnerable to attacks not anticipated at the time of its development. That is, while it can evade attacks that disrupt communication paths, it cannot cope with actions that exploit the network's open structure to flood it with unnecessary information. (omitted) When massive amounts of data place a significant load on the system, individual computers, hub mechanisms, and even the entire network could shut down." (Quoted from book [9-2], pages 37-39).

　In economic networks, hubs are also a weak point. The concentration and monopolization of not only wealth but also links must also be avoided.

"Because financial markets have a densely connected network structure like the internet, the probability of a systemic problem developing even if one of the countless financial institutions worldwide becomes dysfunctional is extremely low. This is because the vast majority of financial institutions are merely one of countless spokes connected to a very small number of hubs. However, if one of the financial institutions acting as a hub fails (which is extremely unlikely but very dangerous), not only thousands of directly connected financial institutions but also other hub mechanisms will be affected, putting thousands of financial institutions connected to them in dire straits." (Quoted from book [9-2], page 55).

　The following point is considered particularly important. This is because, in real-world scale-free networks, everyone (people, companies, etc.) is connected homogeneously, through efficiency-driven preferential attachment, resulting in the creation of excessively concentrated hubs and extreme vulnerability. Therefore, from the perspective of resilience, preferential attachment is undesirable. Also, even if the shape of a large and intertwined network appears complex, the operating principle must be relatively simple, such as self-organizing.

"Complexity, centralization, and homogeneity amplify the vulnerability of a system, while appropriate simplicity, locality, and diversity enhance its resilience." (Quoted from book [9-2], page 76).

As an example of a self-organizing organization, the advantages of decentralization can be considered as the reason why centralized militaries like the United States have difficulty defeating terrorist organizations like AL-Qaeda.

"The dynamics of autonomous and independent elements within loosely interconnected networks are replicated in the relationships between terrorist organizations and other non-governmental combat groups, and also within individual groups. These small organizations are bound together not by traditional, rigid command and control, but by flexible, redundant, and informal social relationships. They are closer to a ragtag basketball team than the Marine Corps. The network maintains its agility through its inherent small groups, and the many-to-many connections of the larger network ensure that the entire network continues to function even if 10 or 20 percent of its members are eliminated." (Quoted from book [9-2], pp. 85-86).

"The fundamental idea of a resilient system is decentralization and shared control. In a collectively organized system, there is no single entity wielding absolute power. Nor is it completely chaotic. It strikes a balance between the advantages of central command and control, and appropriate local authority and autonomy." (Quoted from book [9-2], page 120).

　To demonstrate resilience in a group organization, the first step seems to be to build a decentralized network of trustworthy people, rather than creating a centralized hub. However, the ideal decentralized mechanism has not yet been fully scientifically clarified. On the other hand, the following profile of a leader who acts autonomously rather than by title is not inconsistent with the Teal Organization described in Episode 8. For example, in relation to the latter half of Episode 7, B. Taylor, the technical manager at Xerox Palo Alto Research Center (PARC) in the United States, who invented and developed almost all of the basic technologies for networked PCs (personal computers) in the 1970s, seems to have been such a leader. He was someone who could connect communities with weak ties, manage global trade effectively, get along with people of diverse ethnicities, religions, and cultures, and also likely to be a key person in creating a circular economy in a region, as mentioned at the end of Episode 6

"Beliefs and values, habits of thinking, trust and cooperation, diversity of perception, strong community, interpretive leader, The ability to adapt to crises—these elements form the fertile ground for fostering social resilience. When these come together, new ways of enhancing the resilience of communities, organizations, and the people who live within them emerge." (Quoted from book [9-2], page 22).

"The fact that communities exhibiting superior resilience have a specific type of leader supporting them. (omitted) They differ from the typical image of a leader. They are neither the strong-willed CEO with a clear vision nor the politician who makes bold decisions and takes charge. Nor are they the grassroots activists who gather the opinions of the general public. They exert leadership from the middle ground. They are a new type of leader that has previously gone unnoticed. They flexibly work across organizational hierarchies, drawing in groups that are often left out, and acting as interpreters for mutual understanding among the stakeholders. Their influence, which could be called that of a 'leader as an interpreter,' is more about their actions than their formal titles. It is rooted in informal authority and cultural norms. (omitted) Interpreter-type leaders must play the roles of bridge-builder, flag-bearer, leader, behavioral economist, and social engineer in a variety of situations. They must do so fairly, generously, and with transparency and dedication. (omitted) They must create the groundwork for collaborative action, connect people, gain their approval, and arrange for people to come together in the gaps. Leaders govern not by command or domination, but by influence and coordination." (Quoted from book [9-2], pages 319-356).

　In Episode 10, we will explain how the situation in many real-world networks, where links are concentrated around a very small number of hubs and the vast majority of other nodes have no choice but to rely on them for connectivity, applies broadly to the concept of centralized monopolies and subordination in the world.

Series: Learning about cutting-edge science and human history to prevent social division