3.3 Reductionist thinking

Logical and causal ways of thinking aren't so good at helping us to think about systems for four reasons.

First, the tendency of much logical and causal thinking is to observe the characteristics of specific situations and try to derive general principles or recognise general patterns about the chosen class of activities. In contrast, when dealing with complex situations we are often looking to take specific actions to improve them.

Second, logical and causal thinking attempts to be rational and objective and ignore what are seen as subjective, emotional factors. Certainly it is possible to be objective about the workings of the ‘car engine system’, but a lot harder to be objective about a ‘system for choosing a whole car’ – people have distinct preferences. The first example is of a ‘simpler’ mechanical system, the last example a ‘complex’ human activity system. The difference is that in the latter the views of particular people, and the views of groups, are almost certain to conflict (rather than just may conflict), and these conflicts will often be over what they want from the systems they design, run or use.

Third, we cannot always predict the behaviour of ‘complex’ systems – any changes can lead to unintended consequences. Most of the time we have to say ‘it all depends’. In the case of installing a new information system, whether the software was well specified for the task and whether the people involved adapt well to using computers, and how well they adapt, will all influence how easy the people find it to use the software – which should have been considered as part of the original specification.

Fourth, systems are characterised by interconnectedness, and in particular by feedback loops. So thinking of separate, simple cause and effects isn't going to help us to consider the many different interacting factors and feedback loops in, say, a public transport system or a family system for shopping and cooking.

These features of looking for general principles from particular instances, ignoring subjective elements, concentrating on ‘simpler’ systems and breaking situations down into smaller parts where single cause and effects are likely, are typical of the scientific method. Such reductionism artificially restricts the components in a system to make it possible to observe repeatable experiments. In spite of this, reductionism has proved so effective in practice and produced such outstanding results that it has become embedded in our language, literature and thought. Another reason for restricting what is looked at is the scale of calculating any quantitative changes but the advent of computers means that scientists are also beginning to look at more complex situations that are characterised by non-linear, dynamic interactions rather than the simple, linear relationships.

However, systems thinking tries to take account of these particular factors by adopting a holistic approach that complements reductionist activity and/or by tackling situations where scientific thinking and the scientific method are inappropriate.