There are upsides and downsides to being a pioneer. You start off excited by a new idea or concept, continue to develop it in multiple collaborations often covertly, make it practical, get the first few organisations to take a risk and do something with you. That period is frustrating but enjoyable. Then the idea starts to creep into the collective consciousness and selling becomes easier. This is the sweet period, vindication is great, even without acknowledgment. However almost inevitably a sour note creeps in as the ignorant and the opportunistic (and sometimes the opportunistically ignorant) jump on the band wagon and trivialise the subject. The other major problem is where another field purloins the new language or idea, hijacking it to vindicate a now tired concept, the classic example of this is relabeling Information Management as Knowledge Management. In effect the characteristics of my title are, in effect the four horsemen of the apocalypse for new ideas and concepts. Hijacking is the main topic of this post having come across a fairly blatant example yesterday which I want to share. I was also irritated enough the avoid the usual circumspection and name names, so read on if interested.

Now I have been through this cycle for new concepts and technologies several times in my life with personal computing, decision support, object orientation, data warehousing, JAD/RAD and most recently knowledge management. I have a body of examples in which all four horseman are clearly evidenced, normally in isolation or in pairs. In the vast majority of cases I respond with a display of mildly irritated stoicism. We are now entering that stage of adoption in respect of complexity science (CAS) and I was mentally prepared, or at least thought I was until yesterday.

The occasion was a seminar on Innovation and I will freely admit that I was not in the best of moods. A bout of gastroenteritis had meant that I had to cancel a there and back again trip to Mexico. I could struggle up to London, but not survive two flights in three days. I turned up to give the closing keynote on the application of complexity science to innovation. I knew that the programme also had Elizabeth McMillan offering an Edge of Chaos Assessment Model. Now I met Elizabeth some years ago when I was in IBM and we only had one meeting. I came to the conclusion fairly quickly, and with cause, that she was confusing systems dynamics with complex adaptive systems theory. In effect taking the language of complexity but hijacking it to fit an older and more familiar model. I am never very happy with the edge of chaos phrase either, preferring far from equilibrium which is I think more accurate. She is not alone in either of these sins by the way and stoic indifference would be my normal reaction although I had warned the conference organiser that harmony might not prevail.

So I get to the conference, and I am having a cup of coffee waiting for my turn. I thumb through the slide handouts to check what had gone before and discovered Elizabeth’s slides. At this point mildly irritated stoicism started to mutate into an oscillation between weary resignation and righteous fury with the latter winning out. So what caused this? Well to understand I will need to go through some basics first.

There is a fairly standard distinction made between ordered, complex and chaotic systems. Most writers in the field accept this with variations. For example I split order into simple and complicated as does Zimmerman, although who came first is an open issue! The point is that these are three types of system. I normally summarise it as follows:

  • Order in which the system constrains the agent. As a result they are characterised by repeating relationships between cause and effect. Reductionist techniques work and they can be controlled by process. They are deterministic and observer independent.
  • Chaos, here independent agents operate without constraint. They can be studied and understood through statistics & probability which create a degree of predictability
  • Complex, and here while the system constrains the agent, the agent also constrains the system creating inherent uncertainty. These systems operate in far from equilibrium situations and are subject to the principle of “locality” and constant adaptation. They are highly sensitive to starting conditions and system level effects are emergent &non aggregative

Now note, these are descriptions of systems not value statements about their nature. They just are, they exist and they behave and are interpreted in different ways. I also know that the above are simplifications and I owe a particular debt to Cilliers in his book and various papers for those definitions.

Now this threefold distinction had been picked up by Elizabeth but with a completely different spin. In effect order (or stability in her language) and chaos had been used to designate undesirable states, with complexity (now labeled as edge of chaos) representing a desirable and idealistic outcome. Her assessment tool went further in claiming to allow an individual or an organisation to designate its status with action plans to live on the edge of chaos in the event they were either stable or chaotic. Not surprising the first stage of any such action plan is to take the Edge of Chaos Assessment Model. Now I am not allowed to reproduce her slides without permission, something I think I am unlikely to get! However I can summarise them in the table below, and then use that as a basis for criticism and a justification of my earlier assertions.

Totally Stable Edge of Chaos Chaotic
1 Solid IceTwo rigid no novelty Water Gaseous StreamToo disorderly, novelty overload
2 Ultimate Couch Potato Moving around/exploring Ultimate Headless Chicken
3 Stuck in the past. Repeating past behaviours to detriment of present and future Values past, contemplates the future, lives in the present Obsessed with the future to the detriment of the present
4 No innovation Constant flow of appropriate delivered innovations Radical innovations conceived but not delivered or inappropriate
5 Tight, rigid management controls. Change can be organised but does not occur Self organising principles, bottom up, shared processes – constantly adapting Change cannot be co-ordinated. THere is confusion & no coherence
6 Inadequately connected to all parts of the system. Little or no flow of relevant information Well connected to all parts of the system. Flow of relevant good quality, important information that is useful & manageable An overwealming overload of information relevant & irrelevant & of differeing quality such that it is impossible to handle & make sense of
7 Single loop learning only, static mental models Lots of double loop learning & single loop learning too Learning is disconnected from reality & frantic – lack of sense-making
8 Ossification certain Survival chances are high Disintegration inevitable

Now I need to make three statements here before proceeding:

  1. Elizabeth’s full assessment model introduces two intermediate stages between the extremes and the centre, to wit Stable Aspects and Chaotic Aspects. These add little to the model and are irrelevant to my argument so I have left them out. In addition row 1 is a description, possible metaphor used by way of general explanation. Rows 2-3 are from the Individual Assessment and the remain rows are the Group/Organisation Assessment. I have picked off some illustrative rows for my argument, but I was spoilt for choice, no need to be that selective.
  2. Critically, my argument is not with the table above per see, but rather with the column headings and the hijacking of CAS to validate an approach and model to which the science bears no relation, the opposite if anything. The model is a fairly classic here are the bad things, here are the good things, where are you, where do you want to be approach of which there are a thousand examples. Most consultants have their pet set of platitudes and this one is no exception. The addition of a self assessed qualitative approach is not especially original either. I would have ignored it if it was not for the hijacking.
  3. I know several people at the Open University who are engaged with issues of Systems and Complexity. I would like to make it clear that I exempt them from any guilt by association. I have no idea if her fellow directors of the Open University Complexity Science Research Centre are tarred with the same brush, although the web site does not indicate much activity. There are other centres in the UK, notably Stacy’s at Hertford and Middleton-Kelly’s at LSE as well as others. I have my disagreements with both of these, but I would not challenge their understanding of the subject. This is not an issue of competition it is an issue of abuse of language.

So, what are my objections? Well many and various.

  • The solid-liquid-gas distinction is a useful one if you use it strictly as a metaphor. The phase shifts of latent heat help people understand that order, complexity and chaos are distinct. However it is critical whenever you use the metaphor to say that it is a metaphor only. It is not clear if this is being used as a metaphor or not, but lets give it the benefit of the doubt as there are more serious issues to deal with.
  • Order is not necessarily the same thing as stability, nor is it true to say that stability is always wrong. There are whole aspects of an organisations work which need to be ordered. Getting payments out on time, validation of security transactions, control process on swabs in an operating theatre. Order or stability does not preclude (row 6) timely and relevant information flow. In fact it may enable it. Rigid management control, sometimes with zero tolerance for failure are necessary.
  • Chaos is not necessarily the same thing as anarchy. Used properly the controlled creation of chaotic environments is a classic innovation move. Exploitation may require us to create interdependency with the system (in effect a shift to complex) but the exploration itself takes place in a chaotic domain. To all intents and purpose markets may at times be considered chaotic if agent decisions are independent allowing statistical and other mathematical tools to come into play.
  • The descriptions of a complex domain (or edge of chaos) bears little or no resemblance to complex systems per se. Living in the present, valuing the past and contemplating the future is plain common sense. There are no special features of such behaviour that would mark this out as exclusively complex.
  • Connectivity to ALL parts of the system (row 6) is only partially true. A complex system is largely determined by proximity of agents, with agents adapting to local contacts not to the system as a whole (other than through constraints and those in term are proximate when encountered). Good quality, important and manageable information flow is a characteristic or ordered system not complex ones. Complex systems are always messy, they allow seemingly irrelevant information to flow which may or may not be useful.
  • A constant flow of innovation is not guaranteed by complexity, niche adaptation, strange attractors all induce system stabilities while preserving the possibility of adaption. It may not even be desirable in all circumstances.
  • Row 7 indicates the origins in systems thinking with the reference to double and single loop learning. This is not the place to consider this in any depth and the concept was a valuable insight when it was created. However, like mental models (Row 7 again) this is concept from which we have moved on in terms of our understanding of learning in the context of the pattern based nature of human intelligence.
  • Self-organisation is a key aspect of natural systems, and complexity theory gives an explanation of how this operates. It also opens up a whole new range of intervention techniques for management science. However it is not the be all and end all of human systems. While self-organisation can be good it does not follow that all goodness is to be found in self-organisation. Human complex systems are distinguished from those in nature by issues of intentionality and identity, we can create order and do so to good purpose.
  • A human system based solely on principles of self-organisation would be a hunter-gather culture of small clans in a resource rich world. With current populations, technology dependence and other factors self-organisation will always be necessarily bounded, and rightly so.

Enough, its simple really, the table above and the assessment tool has got nothing whatsoever to do with complexity theory, chaos theory or any science for that matter. It is a list of desirable states with opposing negative extremes. It is platitudinous, contains little that has not been said elsewhere and more effectively, but it is not harmful per se. Some of the case studies she uses (Gore, the Eden Project) are useful if somewhat naively described. The actions proposed in part, but not in whole make sense. What is wrong is grabbing some of the language of complexity theory to justify it. When you let loose the horse of hijacking, those of ignorance, opportunism and trivialisation are never far behind.

The really very, very scary thing is that she has started to use the phrase sense-making …….

< Prev

Aboriginal staff induction

Large government agencies can be daunting places for new starters especially for those people who ...

Category:

Further Posts

Next >

The Balle-Argentee method of business improvement.

In a comment to my post of yesterday, Brian Sherwood-Jones referenced this wonderful story. It ...

Category:

Further Posts