5: More Is Different
The example of bees is used to consider how decisions are made in complex organizations: when faced with a task such as relocating to a new hive, how does the colony choose among options? While we perceive insect colonies as being highly organized, there is no hierarchical structure. The "queen bee" is largely a misnomer - her status as the only breeding female in the hive is not accompanied by any authority, and there does not seem to be any decision-making apparatus.
Researchers (Seely and Visscher) observed that a few hundred bees will scout locations, then return to the hive and perform a "waggle dance" to indicate its location, much as they do for finding nectar. The decision to select the nest is made at the nesting site, not in the old hive: when a sufficient number of the scouts assemble at a given location (fifteen in this experiment), they return to the hive and stimulate the swarm to move. Essentially, it's a vote in which the site that is the first to gain 15 votes wins.
That is to say: even without a formal hierarchy and only the most primitive means of communication, swarms have all the equipment they need to make decisions and coordinate the activities of a large group.
Complex Adaptive Systems
Complex adaptive systems have three parts:
- A group of heterogeneous agents (independent and different to one another)
- Interactions among the agents have the ability to create structure (emergence of patterns)
- Finally the structure itself becomes a system that is distinct from the individual agents.
Return to the notion of swarms: the individual insects are completely inept, but the colony as a whole has considerable organization and power as a unit. In such a system, understanding any of the parts does not give you an understanding of the whole - and those participants in such a system often have no perception of the whole. This can be disorienting, because people have a deep desire to simply complex systems and understand the whole as debilitative of its parts - but a complex system defies such analysis, and must be accepted on a systemic level.
The author mentions stock markets as complex systems, in which the interactions of many parts and many people drive the prices of securities. In spite of tremendous effort put into market analysis, no-one has yet been able to identify the factors that drive the behavior of the system. Moreover, the typical method of analysis - considering the behavior of specific individuals and assuming that what is true of one applies to all - also fail.
This is a situation in which the wisdom of the crowd is clearly more accurate and relevant than the opinions of experts: the experts can describe how people should act based on certain known principles, but this does not suffice to explain or predict the behavior of the market because not everyone in the crowd acts by the same principles.
In a complex adaptive system, the behavior of the system aggregates the behavior of the components - but does not derive from it. A colony of stupid bees can show significant intelligence. A market of smart investors can show significant stupidity. It is not a simple averaging of the intelligence of the members, with smarter ones cancelling out the dumber ones. The collective behavior is a thing unto itself.
Unintended Consequences
When changes are made to a system that has many interconnected parts, attempting to make improvements in one area can have unforeseen consequences for the system.
Consider that virtually every attempt mankind has made in the area of ecology has been disastrous. For example, feeding the elk in Yellowstone caused their population to increase. The elk ate the vegetation and then began tearing up the aspen trees. For lack of trees the beavers could not dam the rivers ad the waterways that remained were clouded by erosion, such that the trout population dwindled. The loss of beavers was incorrectly assumed to be the result of predators, so the rangers killed off wolves, mountain lions, and coyotes, who no longer thinned the herds of prey animals, so the elk population increased ... and repeat.
Another example is the way in which medicine has become specialized, with an increasing number of doctors who are highly focused on one part of the body or another and fewer focused on the patient in a holistic manner. Patients who begin with a minor condition can be bounced from one office to another, none of whom seeming to recognizing a treatment for their area of focuses causes side effects in other areas, and medicines are prescribed to deal with the side effects of other medicines, until the minor issue spins up into a serious systemic problem.
We generally recognize that even good intentions can cause catastrophe because complex systems do not obey simplistic cause-and-effect rules ... but then, we go ahead and tinker with them anyway. Consider the recent economic crisis and the efforts of governments around the world - each thing they do to mitigate the consequences of their actions seems only to make matters worse.
The Myth of Star Power
Many organizations seem to seek the same solution: they hire a star player, thinking that such a person can turn their operation around - but they often find out that stars fail to transform an organization. This is also a mistake related to complex systems, assuming that the individual parts, rather than the system itself, are responsible for performance.
The author mentions an observational study in which the performance of financial analysts were tracked for a period of ten years - and the conclusion was that those who achieve "star" status and are hired away do not perform as well, and some do quite miserably, after changing companies.
First, consider that the individual had been a star in the context of a different organization. He was able to flourish in that environment because of the environment itself because his performance depends on the people, structures, and norms around him. If this new organization doesn't provide him the same kinds of support, he will not achieve the same level of performance. To draw an analogy, this is a failed organ transplant, in which the body rejects the new organ.
Then, consider the organizational perspective: the star player from another environment is a foreign body in the system, and one who is inclined to work at odds with the system rather than collaboratively with it. He may not be able to contribute, and his efforts may even be counterproductive. Back to the organ transplant analogy, this is a case where the organ sickens its new body rather than strengthening it.
The conclusion is that "the constellation matters more than the brightest star."
Suggestions
- Attribute the behavior of systems to the systems themselves, rather than assuming it to be a mere aggregation of individual agents.
- Consider the arrangement of elements within a system - each component interacts with others, and some have strong dependencies on one another, such that a cascade failure can result when one unit fails.
- Create models that will simulate of complex systems, using them to consider the net effect of any change. The simulation should consider how a change in one element might impact others - including the intentional and unintended consequences.