What are Complex Systems and Why Should We Care?

In a January 2000 newspaper interview, renowned Cambridge University physicist and best-selling author of the book A Brief History of Time Stephen Hawking said, “the 21st century will be the century of complexity”.?

What is complexity? Why does one of the most intelligent people alive believe it will occupy the minds and energies of scientists for the present century?

To better appreciate complexity science, which is the study of complex systems, let us go back in time to 1948, when American scientist Warren Weaver reflected on the state of science at that time. In his essay, Weaver explained that over the 17th, 18th, and 19th centuries, scientists like Newton, Faraday, and Maxwell worked on and solved many problems of simplicity. These are problems like the orbit of planets around the Sun, the link between electric and magnetic fields with light and radio waves; problems that can be written mathematically in terms of a small number of variables.

From the late 1800s onwards, scientists like Boltzmann and Gibbs also started attacking problems of disorganized complexity. The problems they confronted, like the pressure of a gas, or the magnetization of a bar of iron, must be written in terms of 1023 variables. Finding individual solutions for this many variables is mathematically impossible, so Boltzmann and Gibbs switched from a mechanistic picture to a statistical picture, to understand macroscopic properties in terms of the average behaviors of microscopic variables.

With these two successes under our belt, Weaver then called on scientists to move on to problems of organized complexity. These include problems in biology, social science, and earth systems. Like problems of disorganized complexity, these involve a large number of variables, and are therefore not simple. However, unlike problems of disorganized complexity, a purely statistical description in terms of averages falls far short of explaining the diverse phenomena observed in such systems.?

After Weaver’s call, the response was tentative at first. The field really took off in the 1980s, when complexity scientists started to get themselves organized, founding the Santa Fe Institute in 1984. After more than three decades of intense study, we now understand that a complex system typically consists of a large number of parts, like cars in traffic, individuals in crowd, biological molecules in a human body, traders and stocks in a financial market, producers and consumers in an economy. When we put these parts together, we get complex behaviors that none of the parts have, as a result of strong interactions between them. Describing this emergence or self-organization phenomenon that can occur on multiple levels, Nobel Laureate in Physics Philip W. Anderson explained that “the whole is more than and also different from the sum of parts”.

Complex systems are robust. When their environment changes, their behaviors frequently do not, until the changes go past a critical point, where we see abrupt switches in the behaviors of these complex systems. We call robust behaviors regimes, and when a complex system switches from one regime to another, we say that it has undergone a regime shift. Over a long time, a complex system will wander from regime to regime. This sequence of transitions, and the times spent in each regime, are hard to predict. More importantly, when several outcomes are possible in a given transition, the actual outcome depends on previous transitions. This is called history dependence or contingency.

All this is nice and well, but why should we care about complex systems??

I can think of three reasons.

First, complex systems are clearly the next frontier of scientific knowledge. While we already know defining characteristics of complex systems, we do not really know why complex systems are so common in nature and human societies. In fact, most complex systems seemingly defy the Second Law of Thermodynamics, which states that the Universe must become more disordered over time. Nobel Laureate in Chemistry Ilya Prigogine resolved this ‘paradox’ in living organisms, by suggesting that these dissipative structures can continue to grow in complexity using energy supplied by the Sun. We suspect the behaviors of complex systems are intermediate between simple deterministic and completely random, due to some kind of compromise between the ability to tell different situations apart and the ability to recognize the same situation each and every time. We believe this information processing aspect is key to building a theory of complex systems, one that allows us to explain the evolutionary advantages complex systems have over simple ones, but much work is needed to develop such a theory.?

Second, engineers are also increasingly interested in complex systems. Engineers are good at building complicated structures like cars and road networks, but struggle to understand why it is so difficult to tame vehicular traffic. A car is a complicated system, because it is made up of more than 30,000 parts. However, a car is not complex, because we can make it go faster by stepping on the accelerator, make it go slower and eventually come to a stop by stepping on the brake, and change directions by turning the steering wheel. Ultimately, vehicular traffic is complex because of the human drivers, and the choices they make navigating the road network.

When a highway jams frequently, the standard engineering reaction to the problem would be to add lanes to the highway. This is what the Land Transport Authority did, when traffic jams persisted along the Central Expressway (CTE) even after electronic road pricing (ERP) was introduced. They undertook a 4-year widening of the CTE, adding one more lane to the section between Bukit Timah and Yio Chu Kang. After the widening, traffic was smooth for a while, before jams returned, ever larger in size. The wider CTE attracted drivers who previously used other southerly roads to get into the city area.

We see from the CTE example, when we do not recognize a problem as complex, and go for the ‘obvious’ solution, we end up with unintended consequences. Engineers should learn complexity thinking, and understand that complex systems cannot be controlled, but can be steered. For many complex problems, the complexity solutions also may not lie entirely within the engineering domain. We may need people to change their behaviors, or make sacrifices. So lastly, once people become part of the complexity solution, we need to talk about public policies.

Complex systems and complex problems are not entirely new to public policy makers. As early as the late 1960s and early 1970s, Horst Rittel, Melvin M. Webber, and C. West Churchman talked about wicked problems. Using pre-complexity management terminologies, Rittel and Webber came up with 10 defining characteristics of wicked problems. If we recast these characteristics into complexity terms, then wicked problems are essentially problems that are strongly history dependent, where variables are strongly interdependent, and eliminating symptoms in one set of variables lead to the emergence of symptoms in another set of variables. In other words, ‘obvious’ solutions to wicked problems are merely treatments that suppress symptoms.

As our world gets increasingly global, we will encounter more and more of these complex wicked problems. We should expect complexity solutions to not only cross domains (from technological to sociological), but also cross political and cultural borders. This is perhaps what went through the minds of Heinz R. Pagels, when he said famously, “The great unexplored frontier is complexity…I am convinced that nations and people that master the new science of complexity will become the economic, cultural, and political superpowers of the next century.”

The Singapore Government has always impressed other world leaders with its foresight and long-term vision. It seems that our leaders also understand the importance of complexity science in public policy. As early as 2004, Prime Minister Lee Hsien Loong spoke at the opening of the Commonwealth Association of Public Administration and Management on how globalization makes governance more complex. He then explained that to cope with this trend, the Singapore government has undertaken to “function in a more networked fashion, to cope with new issues that are complex and multifaceted.”?

More recently, Prime Minister Lee mentioned at his swearing-in ceremony on 1 October 2015 that Singapore will “[be] entering a new phase of our nationhood. We face more complex challenges and new issues that cut across multiple domains”. Therefore, he appointed three Coordinating Ministers, in National Security, in Economic and Social Policies, and in Infrastructure. Speaking at The Straits Times Global Outlook Forum on 20 November 2015, Deputy Prime Minister and Coordinating Minister for Economic and Social Policies Tharman Shanmugaratnam spoke about the challenges posed by global economic complexity that arise as a result of restructuring in the United States, and reforms in China.?

Transport Minister and Coordinating Minister for Infrastructure Khaw Boon Wan, also urged the public to be “realistic about such a complex system - it's not rocket science but also not straightforward. There will be fires big and small but I hope for Singaporeans' patience and we will do our best” as his ministry strives to improve the transportation system’s reliability.

To sum it all up, Singapore’s leaders understand the importance of complexity and complex systems to public policy since the early 2000s. This recognition is important, but we must also be honest: there are not yet complexity solutions to complex policy problems. In fact, there is not yet a scientific theory of complex systems within which we can work these solutions out, or engage in any semblance of complex systems engineering. There is much work to be done, by scientists, engineers, and policy makers interested in complex systems. This is therefore a research area where public funding can be put to good use. As early as the late 1940s, Weaver felt the urgency of complexity research, and suggested that to solve problems of organized complexity, we need to harness the power of computer simulations, and to form multidisciplinary teams. Weaver was indeed prophetic.

[Editorial article I wrote for the Jan 2016 newsletter of the Nanyang Technological University's Complexity Institute]