MGMT 8035 – Holland’s Model For Complex Adaptive Systems


John Holland (2014) defines complexity as a “noun” describing “objects” with many interconnected parts (p. 1) that exhibit “emergence” (p. 2).  In his Ted Talk video entitled “Building Blocks to Innovation”, Holland (2010) describes four properties and three mechanisms present in complex adaptive systems.  This discussion will provide my perspective of Holland’s mental model for complex adaptive systems and evaluate the strengths and weaknesses of this approach to complexity.  I will conclude by explaining how Holland’s ideas might inform my understanding of the structure of organizations and how this understanding might influence my ability to manage effectively.

Holland’s First Property -Aggregation

Aggregation refers to the organization and collection of agents or objects into a structure that can be used to categorize performance.  Within the aggregate, the individual agents employ a degree of parallelism.  Parallelism in a complex adaptive system involves a large number of agents that interact simultaneously by sending and receiving signals. Borges, and Menegon, (2012) describe how complex systems demonstrate a high degree of interactivity between agents, and that the interactivity is present at all times, leading to highly dynamic conditions.  In Holland’s (2010, 2014) view of aggregation, there is no “best agent”, because all agents are required to perform the exchange of materials, information or energy across aggregate.  Schwandt, and Szabla, (2013) describe how fine-grained interactions at the individual nodes between people lead to coarse-grained behavior of social systems, and the heterogeneous detail is required for the system to operate as a whole.  An example of aggregation is an automobile, which is ostensibly an aggregation of parts designed to perform the function of human transportation.  Within the aggregation of “car” are further aggregates of parts and functions, including the steering, transmission, engine, fuel delivery, exhaust, braking, suspension, cooling and human containment systems.  A single gear that does not perform its function can result in the aggregate function of a transmission to fail, further causing the function of the “Car” to cease functioning.

Hollands First Mechanism – Tagging

Tagging refers to the ability to characterize the behavior or function of agents performing a collective task that can be recognized by other agents.  In terms of a complex adaptive system (CAS), tagging often involves the identification of a discriminating feature or function that allows for differentiation between that function and other functions within the CAS. Tagging is commonly associated with the functions of an aggregate. For example, the tag “automobile” differentiates the mode of transportation from other forms of transportation (train, bus, walking, etc), similarly, “transmission” provides a mental abstraction for the various abstracted functions associated with the transfer of energy from the engine system to the drivetrain system.

Holland’s Second Property – Nonlinearity

Papenhausen, and Parayitam, (2015) point out that the aggregate of linear functions often results in complex nonlinear behavior.  They further suggest that human beings tend to make poor decisions in the presence of nonlinear behavior, resulting in amplification and phase lags between stimulus and response.  Saurin, Rook and Koskela (2013) describe nonlinearity in complex systems as a significant number of dynamically interacting elements which are interdependent, leading to nonlinearity.  An example of nonlinearity in an automobile is ride stability.  When a car moves at relatively slow speeds, the tires react predictably with the road surface, and the suspension absorbs inconsistencies in the road, and the car drives straight.  As the car accelerates to high speeds, the damping system of the suspension is no longer capable of reacting linearly with the road conditions, and the force of the wheels on the road change.  Small perturbations in the road, the wheel system, and torque applied to the wheels makes the car wobble and swerve, forcing dynamic changes into the stability of the car.

Hollands Second Mechanism – Internal models

Internal models are used by agents and functions within a complex system and allows for rapid identification of nominal and extraordinary circumstances.  Internal models act as a filter to allow for the identification of patterns that the agents can use to adapt to internal and exogenous stimuli.  Repeated patterns of performance allow an agent to recognize the pattern and adapt rapidly.  Nyamsuren, and Taatgen, (2013) studied how people respond to complex games, and found that they used internal models to adjust and improve future play.  Internal models are useful in driving, because the driver does not need to analyze all possible interactions at a stop sign. Instead, the action is abstracted to a model which ascribes a red octagon as “stop / give the right of way”.

Hollands Third Property –Flows,

Stocks and flows are a fundamental concept in system dynamics, and an integral part of complex adaptive systems.  Materials, energy, and information are stocks that can be quantified, but change over time.  The relative change of a stock over time is described as a flow.  Materials, information, resources, and energy tend to flow as edges in a network of nodes (agents) connecting them by the flow activity. A car is connected to its destination by the road.  The drivers are connected to the road through traffic.  Adding new nodes increases the complexity of the network, just like adding a new road provides drivers with multiple new combinations of path choices to get to their destination by including that path in every possible path choice available to the driver.

Holland’s Fourth Property – Diversity.

Agents in a complex adaptive system provide a function that serves the aggregate.  As new agents arrive, or are created, they adapt to complex behavior and the interactions between the agents affects other agents.  The way I drive my car influences the way others near me drive.  If I hit the brakes, the agent behind me must adjust to ensure the continued function of “driving” proceeds.  As agents spread through a system, they diversify, and create new patterns of behavior that are unique.  Holland (2014) describes these as a “niche” that affects the overall capability of the system.  Disruptions in the system force the niche to adapt or persevere in response to the disruption.  The evolving behavior results in a new niche. And multiple adaptations lead to diversity in the system.  Diversity is a key factor in determining resilience, which measures the ability of a system to successfully adapt to change. Xiao, and Drucker, (2013) studied the economic diversity of the Midwest United States, and found that economic diversity provided the ability for the area to respond and adapt to disastrous flooding, hastening relief and eventual economic recovery.

Holland’s Third Mechanism – Building Blocks

The mechanism of building blocks is perhaps best understood by Lego construction toys.  A Lego construction set is made up of simple bricks that snap together to form shapes. But the shapes arising from the combination are practically infinite.  For example, six pieces of eight-stud Lego bricks can be combined in over 915 million ways (source:  Complex adaptive systems can be broken down into simple parts. These parts can be combined in any numerous ways, resulting in adaptation and change.  Holland (2010) suggests that adaptation is the key to change through the combinatorial effect of building blocks, and not through mutation.  If all the nodes and their tagged functions in an organization represent building blocks, the number of permutations from the potential recombination of those tagged functions would be significant. Holland (2014) points out that with agents, building blocks could be formed by groups of rules that combine to form subroutines for behavior.  The use of building blocks ensures that innovation part of the adaptive culture, and true equilibrium in a CAS is rare and exists only temporarily.

Share with your friends
Order Unique Answer Now

Add a Comment