Complexity Chaos and All That Jazz- John Watkins
Summary of a talk given in Cardiff 13th June 2012
The basis of this talk is to bring together some of the underlying principles surrounding the emerging field of complexity. I will seek to define and develop ideas surround the terms chaos, and complexity and show how a greater understanding of these can lead to a better approach to health and healthcare.
The macroscopic complex nature and beauty we see in snowflakes, flows naturally out of the interplay between the underling structure of solid ice and the transition that water undergoes as it passes from the liquid to solid state. One might ask what this has to do with complexity, but all will be revealed.
One overwhelming difference between chaos and complexity is that the former is deterministic, given the right knowledge of the starting conditions of a system, and the latter which is not, due to inherent randomness built into complex systems.
The well known ‘manifold’ used to look at the relationship between 3 parameters in a chaotic system, reinforces the image that systems held far away from equilibrium, e.g. super cooled liquids, some aspects of human behaviour, etc., can suddenly switch from one state to another, liquid to ice, stable societies to unstable. These parallels between physical and human behaviour have often, mainly erroneously, led people to believe that the reductionist methods of the natural sciences, such as physics, can be used to focus on very human endeavours, such as economics. Whilst it is true, that many aspects of our daily lives undergo sudden shifts from chaos to order, e.g. people walking in corridors, traffic meeting at junctions, traffic jams, this shift of state, this phase transition, can be unpredictable due to the complexity inherent in human behaviour. Phase transitions, such as those seen when ice melts or water boils, or when magnets cool past their critical point, are well studied in nature, however, when these changes are studied by physicists a number of assumptions are made.
Physics, by its very nature, is reductionist and most of the states explored by it are inequilibrium. Physicists assume that the particles in a gas are non interacting and indistinguishable and so can be grouped together as a whole system. By contrast, many aspects of our society are interconnected in complex ways leading to great difficulties in knowing what effects small perturbations may have. Economics as a discipline has struggled under this misconception for well over the last 100 years.
Some of the greatest minds in economic theory, for example, have made the assumption that ‘the market sets the price’ and any perturbations about this price are purely random, a bit like theBrownian movement of dust particles suspended in still air. Nothing could be further from the truth, market fluctuations are the product of the complex interactions of human kind,interacting in unpredictable ways. One only has to travel down the M4 to London and experience the ‘phantom’ road block to see that this human interaction is anything but random.
Detailed examination of the daily fluctuations in the Dow Jones will show that the financial markets are very different from those predicted by the randomness of Brownian motion and, at times,undergo tremors almost indistinguishable from an earthquake’s seismic trace. We can also see that using standard models to predict complex behaviour, based on the assumptions of randomness and non interaction, rarely gives an accurate forecast c.f. the barrel price of oil over the last 20 yrs. The spike in the oil price in the late 2000’s was not due to a shortage of oil, or social unrest, but driven by speculators, the coordinated efforts of a small sector of society.
So what is going on? Can we make sense of this?
Rather than systems such as; the world financial markets, healthcare demand, flooding, cloud formation, earthquakes etc., being due to purely random fluctuations, they are each infact complex systems that interact in ways that are non linear.
To understand what we mean by non-linear we need to dip into the world of fractals, which is dominated by power laws. The defining feature of power laws are what physicists call scale free distributions, in which there is no typical distribution (unlike the binomial/normal distribution) and in such circumstances, it is not possible to state a mean and standard deviation with any authority, other than to say that such systems are biased to smaller events and that large deviations are less likely. These scale free non linear events can be represented both mathematically and graphically and underpin much of our understanding of the world we live in e.g. the laws of electrostatics, gravitation, black body radiation and Pareto’s distribution, well known to social scientists. In such systems log/log plots will lead to straight line relationships in which the slope will define the power of the relationship between two parameters,e.g. size of flood v its probability. Scale invariance leads to the fractal nature of; second to second, day to day, week to week etc. variations in water levels, seismic activity, share prices etc.
Systems at the ‘edge of chaos’ do not have the same orderly behaviours of gases or liquids in equilibrium, which makes predictions uncertain. Complex systems are generally made up of many parts, with potentially non linear interactions and interdependence. Complex system exhibit structure over several scales, the emerging behaviour of which might not be deduced from examination of the parts, e.g. human behaviour is not predictable by mere anatomy and physiology alone.
Complex systems are self organising as we have seen from traffic jams, people in crowded corridors or the beauty of snowflakes. Complex systems have multiple feedback/control loops with interplays between cooperation and competition. If we want to understand and shape healthcare we need to understand and copy the most complex system we know, ourselves.
The human body, with its emergent behaviour that allows us to contemplate our own existence, employs multiple feedback systems at all levels, we should accept, that in order to shape healthcare and healthcare demand, we need to work at all levels and not,unidirectional, from top down.
Its not rocket science it’s much more complex that that!!