*These are the first two of four parts that I am in the process of writing that talk about Complexity Science and it's implications for Economics and Policy-making. I sent this off to Tribune blogs for publishing but was told that it did not "meet their criteria". This on a day where their featured blogs include gems such as "How I became a drunken, cheating swine who destroyed his marriage" and "Pies over fries: Rich and decadent, this dulce de leche Banoffee Pie is instant love!". Clearly, I have fallen short of whatever criteria were applied to getting these published. Anyway, I shall now also claim censorship and count myself among the victims of this closed, regressive, patriarchal setup that stifles controversial themes and ideas! The truth will not be silenced!
On a serious note, is there really no room or interest for pop science or fact-based articles?
Complexity Science is a field that emerged in the latter half of the twentieth century. It investigates problems pertinent to pretty much any field you can think of - from the Natural Sciences to the Social Sciences to Art - and has had contributions from some of the most brilliant minds in these areas, including several Nobel Prize winners. Stephen Hawking suggested that “the next century will be the century of complexity”.
Since the late 1980s the popularity of the field has been on a rollercoaster ride, being called everything from “the new science” to a passing fad. As it stands, Complexity Science does seem to be gaining attention from multiple quarters and you can even see universities offering degrees in Complexity Science.
The most famous place for Complexity is the Santa Fe Institute, a research centre hosting people from diverse traditional fields working in this multidisciplinary area.
I stumbled across this area by accident a few years ago and thought it to be interesting. My interest in this field was rekindled recently. The following is a summary of what I have been able to gather from my investigating this field on my own. Having a primary interest in the Social Sciences, I am fascinated by what Complexity Science has to offer in this area. I think the results, if not important (which I think they are), are at least thoroughly fascinating!
The Story of Chaos
We start with the story of Chaos.
When Isaac Newton presented the laws of motion, this had interesting implications in terms of how the universe was perceived. In effect, by showing that the reason an apple fell to the ground was the force that also kept the moon orbiting the earth and planets orbiting the sun, he had found a unifying principle for how objects in the universe behaved.
An implication of this was that the entire universe was moving like a giant clock – in a pre-determined path in accordance with certain rules. This further meant that, given enough information and processing power, it was potentially possible to predict the exact behaviour of any physical object in the universe to the end of time.
For the more zealous, this could also point towards determinism in other areas. If the universe was a giant clock, then maybe everyone and everything was simply following a predetermined, predictable path. Even if this wasn’t true, one could at least manipulate physical systems by applying the right force at the right place and lead a system from one predictable outcome to another.
Then, along came Henri Poincare. In 1895, the Kind of Norway announced a prize for a definite answer to the question of whether the universe would always go like clockwork, or whether it would break down at some point. At the heart of this question was the Three Body Problem. While Newton had been able to show that two objects would maintain a stable orbit around each other, the question of whether the same principle applied to three objects was still open.
Because the Three-Body Problem involved a complex set of variables, making it difficult to calculate, Poincare devised a system of approximations in order to show the orbits that could be held stable. He was awarded the prize – but there was a problem. On deeper inspection, Poincare discovered that slight changes in the starting positions of the three objects led to completely different outcomes. Thus, his system of simplification could not provide an answer to how the three bodies would interact because even the slightest inaccuracy in stating the starting condition would result in wildly different answers. Poincare had just hit upon Chaos Theory.
Chaos refers to a system that has “sensitive dependence to initial conditions”. The phenomenon was to resurface with the coming of computers in the mid-1900s. Essentially, in a clockwork universe, the physical world should now have become perfectly predictable because computers could now work through the equations that drove each body and provide a path for each body.
The problem with this idea was picked up by meteorologist and mathematician Edward Lorenz when he tried to apply this approach to predicting the weather. He discovered that his predictions varied drastically depending on how many decimal places were being used in the calculations. This was because he was dealing with a system that could be headed in two completely different directions based on whether he started at position where a variable had the value 0.506127 or 0.506000.
This is also famously known as The Butterfly Effect – the idea that a butterfly flapping its wings in one part of the world may set off a chain of events that results in a tornado in another part of the world.
Thus, for certain situations even if one knew exactly where each object in the system was headed based on where it was now, this knowledge could not be used for prediction because it would not be possible to perfectly capture its current position (because, for example, one might need an infinite number of decimal places). Thus, when predicting the weather, it would never be possible to account for all the butterflies flapping their wings and thus it would be impossible to know when a tornado may be about to occur and where. This is an example of deterministic chaos.
This threw a wrench into the idea of a clockwork universe being easy to predict and/or manipulate. The other idea that apparently struck at the heart of the clockwork universe was quantum mechanics. I say ‘apparently’ because I know nothing of quantum mechanics. And if I thought I understood it, I would not have understood it!
If we were to extend the idea of Chaos and use it as an analogy outside the world of Physics, what is the implication for prediction and intervention in other spheres? What might it suggest for policy-making or legislation? Well, it might suggest that while in some non-Chaotic systems the task may entail a simple cause and effect analysis and implementation, in systems susceptible to Chaos we should be very worried about the unintended consequences of every single action.
However, there is some good news. While chaos leads to random behaviour and consequently to issues of predictability, even within that randomness there is the possibility of the existence of some order. Intricate patterns can emerge from seemingly random behaviour. Readers might want to check out something called The Chaos Game, which is a good example of a process that one might assume to have an arbitrary outcome, yet what emerges is a clear pattern of behaviour. The idea here is that while the very next outcome may be unpredictable, the whole serious of outcomes are not so.
Similarly, the concept of attractors means that while we would not be able to specifically predict the behaviour of objects in a system at all times, in certain situations we would have a general idea of the vicinity it might be in.
Thus, moving back to questions about policy in this analogy, this means that any action need not necessarily be completely unpredictable. It might result in emergent patterns of behaviour, however, these patterns would not be immediately obvious. Thus, once again, we have a reason for caution, but not necessarily a forbidding one.
But Chaos is just one small part of Complexity. There are a few reasons for starting off with the story of Chaos. Firstly, it highlights how a particular view of the world was formed, persisted for centuries and then had to be updated. This is a key agenda among Complexity thinkers – the need for a new view of the world. More specifically, Chaos showed that for certain problems, simplifications and approximations simply meant getting the wrong answers. Furthermore, it demonstrates an important theme within Complex systems – order out of randomness.
What is Complexity?
In general, people working with Complexity tend to have difficulty agreeing on what exactly defines a Complex system. In general, they agree that a complex system is one that consists of the following.
- A number of interconnected components – arranged in a mostly non-hierarchical structure, with each following simple rules.
- A system of information transmission and processing between the components and from outside the system.
- The ability to adapt to changing conditions
Thus, an oft-repeated example of a complex system is those of social insects, such as ants or termites. One might imagine that the formulation and functioning of something as intricate as an insect colony might require either a strong central system for coordinating, or considerable intelligence among the members. Neither is true. Each individual insect exists in a non-centralised structure and is fairly unsophisticated. By following simple rules, insect colonies on the whole are able to tackle complex problems such the construction of colonies and optimally foraging for food.
The human brain is also an example of a complex system. It consists of a network of connected neurons sending and receiving simple electrical signal. On the whole, of course, these add up to the execution of very complicated tasks.
Other examples include the internet, biological organisms etc.
Essentially, Complexity scientists are interested in problems and features that are common in such systems. Thus, they are always on the lookout for discoveries and features from diverse fields that may be extended to others.
One general trend Complexity science is interested in is the principle of emergence and ‘order out of randomness’. In order to appreciate this idea, we need to look at what the traditional approach towards problem solving has been over centuries. The principle method of investigating a complicated problem has been to break it up into its constituent parts.
Thus, if a team who’d never seen a car before in their lives wanted to figure out how it worked, they could disassemble it and study each part in isolation. One member could investigate how doors work, another the steering wheel, and another the engine. Another example might be, if economists wanted to study or predict the behaviour of consumers, they might simplify the problem by looking at what a representative consumer might do. This reductionist approach has historically worked well for academics and researchers.
The argument posed by Complexity scientists is that such an approach may not function well in a complex system. If one wanted to understand the functioning of an ant colony, it would not do to simply investigate the behaviour of one ant. Similarly, if one wanted to understand how the human brain works, it might not be sufficient to look at it piecewise. If we take a reductionist approach to such problems, while we might understand how an individual constituent part functions, the overall picture gets lost.
On the flip side, one might expect a large of interconnected components acting without central control or coordination may produce an overall random outcome in the system.
This is where emergence comes in. The idea is that such systems may, in fact exhibit patterns and behaviour at a system level that is not random but could not be predicted by studying the behaviour of the constituent components. Essentially, these are systems where ‘the whole is greater than the sum of the parts’.
So, one feature of Complexity science is the development of techniques to investigate such emergent behaviour from constituent parts.
Another feature Complexity scientists like to investigate focus on is the concept of adaptive agents. The idea is that agents (components) that comprise a system may change their behaviour over time. This could be in response to changes in their environment and to the behaviour of other agents. Complexity science is very interested in outcomes related to a collection of such constantly evolving agents.
Again, there are various systems that exhibit such behaviour. These include biological systems and social systems. Insights for biological systems may provide some hints as to how one might expect organisms to behave in different situations e.g. in response to temperature changes, or in their interaction with other organisms. It might also provide insights into how behaviour in social settings could change or get ossified.
In both cases, the results of emergence within systems comprising adaptive or non-adaptive agents is going to be heavily dependent upon the interactions between them. Thus, connections are important. In many situations it is going to be critical to know which part interacts with which other parts leading to a particular outcome. This is essentially the study of networks.
Networks are all around us - the internet, roads, canals, social networks (both online and offline), our brains, electricity etc. are all good examples of different types of networks. What’s interesting is that many may have similar characteristics due to similarity in their structure. It is, therefore, entirely possible that findings about behaviour in road networks could provide some insight into how telecommunication networks behave, or vice versa. It is these overlapping structures and subsequent characteristics that would be of great interest to Complexity scientists.
From a social science perspective, analyses in these fields can provide great insights in questions of human behaviour and interaction. It can shed light on questions such as why certain norms and behaviours take off and become prevalent, why markets behave the way they do, or why certain groups are able to maintain or lose power.
Why did these questions just come up? A lot of questions being investigated by Complexity sciences could not previously have been tackled as the tools for investigation were simply not there. In the last century, one big change that has come about has been the advent of computers. As a result, a lot of techniques were developed for investigation into problems that would have previously ranged from forbiddingly difficult to outright impossible. Advances in mathematics will have also played a role in bringing about this interest in dealing with Complex systems.
Quite often, a tool developed for one purpose ends up influencing work and approaches in multiple fields. This is very much the case with the coming of computers where, among other things, the notion of modelling and simulation has become a tool of investigation. Simulations have been adopted for several purposes in the field with varying degrees of sophistication and acceptance.
Complex systems use a combination of mathematics and computer modelling techniques for their theoretical work. The most famous for what are known as Agent-Base Models. Models built from ‘the ground up’, wherein multiple ‘agents’ are programmed with behaviour and left to interact in a virtual world to see what emerges. Because such approaches are still relatively new, there is currently some debate as to their validity as evidence of a result. Essentially, what constitutes a valid model or valid in the social sciences is still uncertain.
On the whole, it seems unlikely that the general techniques and areas of investigation of Complexity science can be ignored. In recent times, we have seen such approaches becoming mainstream in economics – Network Economics is one example.