Self-organisation is in no way a bottom up process. All humans are involved in interaction, and the results that emerge are the results of self-organising processes, whether you are a top manager or lower in the hierarchical ranks.
It is not very accurate to call self-organising a "bottom up" approach or process. To me, a bottom up process is where there is activity amongst those lower in the hierarchy or closer to the front line who feed ideas or a new process up through the hierarchy. A bottom up process can happen because it’s been designed by the top managers e.g. a series of facilitated focus groups, or it can happen more informally e.g. where a problem is solved or a new process is established at one site and then the results are seen by those higher in the hierarchy and the new process is implemented at other sites.
Either way, central to the idea of a bottom up process is that it is starting at the lower end of the hierarchy and bubbling up through the layers. Things happen low in the organisation first and then are fed upwards.
When we say that human interaction is self-organising, what is meant is that ALL human interaction occurs through local interaction in the absence of a central blueprint or plan. The term "local" is used here in a particular sense, which is that all humans interact with a relatively small number of others, not with the whole population.
The top managers are engaged in local interaction and so are those lower in the hierarchy. It is all "local" in this specific technical sense. The main difference between interaction amongst top managers and that of those lower is that the top managers have more power, although the power balance is never static, it is always shifting and changing.
So, self-organisation is not about the top managers letting something happen at lower levels and then feeding the good ideas or lessons learnt up through the ranks. Top managers and lower ranks alike are engaged in self-organising interaction and there is no overall plan or blueprint for the patterns that emerge from that interaction.
Managers cannot "allow" something to emerge from the lower parts of the organisation by relaxing constraints. They can relax constraints, but still do not know what will emerge from the myriad interactions taking place in the organisation. This is because senior managers cannot control how people will respond to that they say and do.
This does not mean that what occurs is random. What occurs is the result of myriad interactions amongst people with their own intentions, goals, psychodynamics, backgrounds and aspirations amidst shifting power relationships. That is why the results are emergent or self-organising – these different intentions, goals, psychodynamics are not controlled by even the most powerful manager.
As the intentions, goals, power relationships change, the sense people are making of what is happening changes, and so new or different results become possible.
So, instead of trying to manage what emerges from this, managers can only participate in interaction (or conversation) with others, and notice what responses they are getting and then adjust and respond accordingly.
This is why I say that self-organisation is not a "bottom up" phenomenon.
Stephen, I think you make an important statement in this post. Bottom-up processes sometimes get mixed up with other definitions. I’m the first to admit that I make this mistake as well.
Self-organizing and bottom-up are two different things. The starting point of these processes are different. With self-organization, it starts from the self (hence self) and goes outwards. It’s more a cyclic orientation, whereas bottom-up inclines a one-way process, initiated from the bottom.
I would say that bottom-up is more democratic or more as self-management. Also, bottom-up can be a result of self-organization, but that’s just one of the permutations.
Comment by Bas Reus — August 24, 2009 @ 10:03 pm
Thanks for your comment Bas. It is interesting that you use the word democratic, perhaps in an approving sense. But of course organisations are not democratic, nobody votes for the manager they want.
Self-organisation can only be said to start from the self if you are talking about human interaction, in which those interacting, i.e. humans, have a sense of self. Non-human things can also be self-organising, like ants, bees, birds and so on. The “self” part of self-organising means that there is no overall blue print or plan that is organising the activities that are going on. I suppose that self-organising patterns can be said to be cyclic in the sense that there are recognisable patterns. But if you think about birds flocking, it would be quite difficult to say these patterns are cyclic, because they are never repeated exactly the same, the way you would in a cycle. I think this is one of the things that makes complexity so interesting, is trying to figure out how complex patterns are different from other patterns we are familiar with.
Thanks once again for your post that made me think about another aspect of self-organisation.
Comment by Stephen — August 24, 2009 @ 10:13 pm
Yes by democracy I mean that in an approving sense. However, democracy in organizations does happen now and then. For example, look at Semco in Brazil. Some employees already determine their own salary (the idea is that eventually everyone there can), and yes, a group at Semco have to decide how will be in the lead. Management is often circulated in terms of 6 months for example. But of course, the organization as a whole is not 100% democratic.
The distinction between humans with self-awareness and creatures like ants is self-organization versus stigmergic behavior. The latter behavior is completely without a plan (or blueprint) where we as humans are always aware of our actions (however, that can be questioned as well
)
But the concept of stigmery remains interesting in both cases because of all actions changes the environment which in turns influences behavior. That is true for ant behavior, and also for human behavior. And both have a social element in there as well.
Comment by Bas Reus — August 25, 2009 @ 12:44 am
Once again Stephen great post. Those of us not immersed in complexity, tend to use words lightly.
Indeed as you explain bottom-up is just an instance of self-organisation, not a criteria.
And thanks for clearing up the meaning that the word “self” is in absence of a plan/blueprint…does this apply to non-humans as well eg. ants
Are you saying that emergence in self-organisation exists in org conversational interaction due to our unpredictable nature(ie our different intentions, goals, psychodynamics, background)…we cannot be controlled by even the most powerful manager, as self-organisation is who we are?
Then what about ant colonies, you say they self-organise, yet they don’t have intentions and goals like humans…(sense of self)
Maybe intentions, goals, etc (sense of self)…is not the only criteria for self-organisation.
Stephen, I thought you were saying that this sense of self is what caused self-organisation, and also the fact that we converse in local interactions rather than a blueprint.
What about ants, yes they have local interactions too, but don’t they differ as they actually are living off some kind of blueprint in their DNA, or do they not live off a blueprint, but differ because they are not self-aware?
So perhaps, if I get you right, self-aware (sense of self) or not, as long as we interact locally and not according to a bluepint we can self-organise.
So if two criteria for self-org are:local interactions, no blueprint, and of course a number of agents
We can say both org conversations and ant colonies have these aspects of self-org.
The difference is humans self-org is due to self-awareness (sense of self)and the ability to have conversations…and ants self-org due to stigmergy (tracemarks that allow coordination with indirect communication)
Not sure if “conversations” is a defining difference…both humans and ants communicate, but in different ways
Comment by John Tropea — August 25, 2009 @ 6:05 pm
Bas, we say that self-organisation occurs through interaction in the absence of an overall plan or blueprint for the patterns that emerge. When it comes to humans interacting though, the individuals who are interacting DO have plans and intentions. The overall patterns that emerge though, occur in a self-organising way through the interweaving of the plans and intentions of the humans.
The self-organising of ants or termites, whether through stigmergy or through interaction (e.g. birds flocking) and the self-organisation of human interactions both result in overall patterns that are not in the control of any one of the ants, termites, birds or people who are interacting. Even though in the case of humans they may be aware of their own plans and intentions.
Comment by Stephen — August 25, 2009 @ 7:58 pm
John, you are really developing your thinking and getting to grips with self-organisation in your comment. You are right that the term ’self-organising’ does apply to ants as well as human interaction. Originally the term referred to natural phenomena such as birds flocking, and also computer modeling of these phenomena where “agents” i.e. bits and bytes would interact with each other myriad times and scientists could simulate many many interactions in a short space of time, that would take years to occur in the natural world.
Intentions and goals are not a requirement for self-organisation. However humans have intentions and goals, whereas the natural phenomena like birds that were studied and simulated by complexity scientists do not have human consciousness. The models and conclusions from the study of complexity in the natural world (such as self-organisation, simple rules, edge of chaos and so on) cannot be directly imported into the study of human phenomena. They can be used as source analogies, but have to be adapted to make allowance for human consciousness, or they will not reflect human reality very accurately.
Self-organisation is not due to our unpredictable nature. It is due to the fact that we interact. If the agents interacting are similar to each other, then the patterns that emerge will be relatively stable. If the agents interacting are different from each other (i.e. diverse) then there is the possibility for novelty in those patterns that emerge. So, the patterns of birds flocking, while always unique and never reproduced exactly the same twice, always look similar, always look like bird flocking patterns, because the agents interacting, i.e. the birds, are similar to each other.
It’s not that we can self-organise, as though we have the choice to do so or not. It is that our interactions are always self-organising. The interactions of managers and powerful people are also self-organising. And so when you are thinking about change or knowledge management or any other aspect of organisations you also have to have an explanation for power so that you can factor that in as well. Complex responsive processes theory which was developed by Ralph Stacey and his colleagues draws on Elias and Arendt’s view of power as a relational phenomenon. Many other schools of thought draw on Foucault with his very different view of power.
You are right that self-organisation requires local interactions between a number of agents. The term “local” has a specific technical meaning in the context of complexity – it means that agents are interacting with a relatively small number of other agents (i.e. “locally”) – they are not interacting with every other member of the total population.
From these myriad local interactions, population-wide patterns emerge. There is no blueprint or overall plan for the population-wide patterns. Although managers may desire certain patterns to emerge, they cannot directly bring about their desired population-wide patterns.
Conversation is not a defining criteria, interaction is. Conversation is one of the ways humans interact and hence the emphasis on conversation in complexity based approaches to organisations.
Human self-organisation is not due to self-awareness and the ability to have conversations. Human self-organisation is simply due to our ability to interact with each other. Ants and birds also have this ability. But, humans have a far greater range of ways of interacting and also have consciousness and will. So you cannot just take insights from the natural world and import them into the social world – you have to make allowances for these aspects of human relating that are different from the relating of birds or ants.
Comment by Stephen — August 25, 2009 @ 8:17 pm
I couldn’t agree more with your post Stephen. Nice one.
Comment by Chris Mowles — August 26, 2009 @ 2:32 am