Howard Rheingold has in mind this idea of emergent behaviour when he writes about “Smart Mobs.” Rheingold is a veteran of the first stage of online community development, and in Smart Mobs he travels the connected hotspots of the world to survey emergent networked behaviour, his curiosity aroused by observing Japanese youth absorbed in text messaging. His thesis in Smart Mobs is essentially that mobile communication devices, pervasive computing and new networked communication technologies have the potential to amplify human talents for cooperation; this creates the conditions for smart mob behaviour where, under certain conditions, people can exhibit signs of collective intelligence as they “swarm” or “flock” together. He cites as evidence the 1999 anti-WTO protests, where protestors used dynamically updated websites, cell-phones, and swarming tactics during the “battle of Seattle,” and the text messaging networks that helped Filipinos organise demonstrations to topple President Estrada.

The behaviour of smart mobs, and in some respects online blogging communities, exhibit many of the features of what scientists call complex systems : they are emergent, highly connected with intricate inter-relationships, self-organising and simple on the micro level but create effects that appear complex and unpredictable on the macro level; plus, they tend to evolve through rapid collaboration or feedback loops. In the physical world, such systems (arguably human evolution is one example) are not necessarily the most efficient, but they are highly adaptive to changing conditions and they generally get the job done. Complexity theory has become a very popular way of thinking about the behaviour of biological, meteorological, financial, social and other systems, and it has helped us understand that apparently turbulent and chaotic systems actually have a tendency to create their own non-linear order.

The classic pop-science example that illustrates the point is the way in which ants forage for food. Ants display a kind of collective intelligence (described by some as a “hive mind” ) that is based on apparently dumb rules, repetitively followed by thousands of individual insects. Each ant forages for food in an apparently random manner, but when it finds food it marks a pheromone trail back to its colony. Trails fade over time, but positive feedback means that well-travelled paths will attract more and more ants until the particular food source is exhausted. The system works because there are enough ants each following the same rules to ensure comprehensive coverage of any given area.

This technique, whereby simple drone-like behaviour can create a physical piece of shared knowledge, has been referred to as generative psychogeography . Parallels have been drawn between this behaviour and “warchalking ,” where individuals mark “free” wireless internet hotspots in cities for other users to come along and use. Complexity theory shows us that from the seeds of such small inter-connected actions, large trees of system behaviour can grow. These physical phenomena are reflected online as well, where the emergence of the Wiki movement and the growing cult of Google both display a simple form of collective intelligence. Wiki web sites – named after the Hawaiian word for “quick”-   open themselves up to editing by anybody who cares to contribute, albeit in many cases with some moderation, such as the surprisingly good Wiki encyclopaedia . Google’s page rank systems works by using link popularity as a major measure of relevance, which means that user behaviour contributes to the selection of search results. The popularity of weblogs (measured in site traffic) is perhaps even closer to the ants model than Google. Examination of site statistics shows that a tiny minority of sites get the vast majority of blog traffic because of the nature of the positive feedback loop created by the interconnections between blogs. This is similar to the ants’ pheromone trails, where some trails rapidly emerge as dominant as others fade to almost nothing. This phenomenon has been described in terms of power laws , and some bloggers have responded to the analysis by recoiling from what they see as the anti-democratic conclusions of the research, seeking to find a way to equalise traffic in some way.

Complexity theory is a type of so-called “systems thinking” applied to the natural world. Systems thinking essentially involves treating a system (e.g. a large organisation) as a unified whole that is greater than the sum of its parts, rather than reducing it down to its component parts and analysing each in isolation. Pioneered by the Sante Fe Institute , these days complexity theory is an important part of how we understand the behaviour of organisations and social systems, and the subject of a major research initiative of the European Commission .

The case for applying systems thinking to organisational behaviour was made most famously by Peter Senge in “ The Fifth Discipline: The Art and Practice of the Learning Organization,”first published in 1990. In his book, Senge sets out five disciplines- Systems Thinking, Personal Mastery, Mental Models, Shared Vision and Team Learning – that underpin "learning organizations" with the capacity to meet the challenge of doing business in complex, dynamic, and globally competitivemarkets. According to Senge, systems thinking is the fifth and overarching “leadership discipline” because it integrates the others into a holistic body of theory and practice.

The application of systems thinking can help explain why some small changes to a system can have a huge impact, whilst in other cases apparently major change can have almost no influence on existing system or organisational behaviour. For example, Peter Fryer, a consultant specialising in the application of complexity theory to companies and public bodies, uses the term Trojan Mice to describe actions taken to affect change that are “small enough to be understood and owned by all concerned but their effects can be far-reaching.” He poses these as a counterpoint to large, highly-conspicuous events (Trojan Horses) that organisations often set up to force through change, but which are soon forgotten and ignored. This analogy will certainly resonate with anybody who has tried to affect top-down change within a large organisation.

Highly connected social networks behave much like complex systems if the right conditions exist, and the rise of ubiquitous computing and networked communications is taking us further in this direction. Among companies and governments alike there is a sense that the old “command and control” system of top-down management is not working. In both cases, there is a growing fear of asymmetric threats – a sense that large, hierarchical systems are vulnerable to external threats such as, for example, terrorism in the case of governments and anti-corporate culture-jamming and hand-to-brand activism in the case of consumer brands. For many organisations, the problem is a lot more mundane: they are throwing huge resources at enterprise software, intranets and other systems that are attempting to affect organisational change with very little to show in terms of results; meanwhile, they watch frustrated as their employees, customers and stakeholders flock to cheap, simple web sites that engage them directly and retain their interest.

If we recognise that organisations and communities are not simply reducible to their component parts – if they behave like complex systems – then surely the applications we use to connect people must reflect this? Some of these emerging theoretical approaches to the analysis of online social networks and behaviour can potentially help us build better, more social online applications for real, existing groups of people.