There Are No Attributes


This concept or paradigm or “dangerous idea” is something that grew out of the most basic idea behind Systems Thinking; that systems drive behavior. Let’s take that basic idea and link it two others; Aristotelian thinking and change.


First, Aristotelian style thinking prevents us from solving non-Aristotelian problems. Essentially, abstract thought can lead you away from Systems Thinking entirely. Where we take the wrong turn is thinking that that abstract “attributes” exist in reality. In fact, they do not exist outside of the context of the system they are a part of. This system includes not only what you are measuring, but the measurement mechanism as well.

Second, cause and effect-based problems refer to change over time.  Complex systems contain structures where cause and effect are widely separated in time and space. The connections between the system components are hidden, and the resulting phenomena are emergent; not predictable by knowledge of the initial conditions. This layered complexity drives systems toward phase transitions more than linear change. In these situations, cause and effect, logic, and change are no longer useful tools for problem solving. How can this be? We just removed the foundation for every tool we have right?

Cause and effect and logic are so tightly linked to our world models that it takes a real effort to see  things in a different way. But the fact is that the modern types of problems that we face in our world today are not analyzable or solvable in terms of simple cause and effect or simple deductive  “If – Then” logic. We are talking about problems are layered in complexity. Dynamic variables interacting in feedback loops,  adaptive agents that interact with the environment and that are part of the environment at the same time, systems where every time you make a change the problem definition also changes. When cause and effect are not there to see, simple logic breaks down. When complexity increases to the point where emergence removes our ability to predict the results because of phase transitions, then change is not a measurable factor. This begins to sound a lot like quantum physics, but it is not. They are not classical systems either. Complex systems are a different animal, and they exist all around us. Our own minds and bodies are examples of complex adaptive systems.

Systems Drive Behavior

So Complex Systems science describes solutions like nudging a system toward a phase transition. An example is a vaccine. A vaccine is a “nudge”…a small change with a big effect. Cool.

To explore these ideas a bit more; we usually think that people drive behavior. But if you have ever played the “Beer Game” (see Peter Senge’s “The Fifth Discipline” or your favorite search engine), then you know that it is the systems, the way we think about them, and the way interact with them that drive behavior.

In other words, trying “really hard” to do a “better job” is not the way to better performance of the system. The way to better performance is to redefine performance within the context of, the entire system. Think about it…when we make changes within a system we are often unable to predict the broader consequences of those changes. We might get what we want on the smaller scale of the component or process, but then it has an effect we did not want and we start trying to solve the new problem in the same way. That is if we even realize the connection between the solution and the new problem. More often we do not realize that the new issue is even related. When we define performance in terms of some attribute that we want a system component to possess, we automatically ignore the side effects of trying to optimize that attribute.

So systems drive this kind of behavior, but that is because we use abstract thinking instead of systems thinking.

Author’s Note: I realize this is a bit recursive. Many systems thinkers say “systems drive behavior” and stop there. The fact is, systems only drive behavior via the system and the thinking about that system. This kind of thing can drive one crazy, so let’s stop there. KS.

Aristotelian Thinking

Let’s take an example. Say that researchers find that a health condition is the result of a deficiency in a substance that the body produces under normal circumstances. They determine this statistically, by identifying a correlation between the health condition and the substance.

Then they try to identify the mechanism that relies on this substance to operate. After more study they find that mechanism, and sure enough, adding the substance in the proper quantities brings about a predictable response related to the health condition. Success!

But then a funny thing happens. They find that supplementing this substance causes the body to produce even less of it, thereby causing a rebound effect and an eventual worsening of the health condition.

Worse, the subjects at this stage develop a “side effect”, caused by the body developing an excess of the raw material used to make the original substance. Since the body’s production has decreased, the raw material builds up, impacting another one of the body’s systems. If this situation happens to impact a substance that the body uses in many processes, the overlapping effects can be wide-ranging, even catastrophic.

The mistake is looking at things in a simple cause and effect manner. This was initiated by looking at the substance as an artifact or attribute of the health condition, rather than as part of the larger system.

The health condition only exists because the larger system – the human body – exists. So why would we look at the health condition as if it existed apart from the body? The reason we go down that path is that we use abstract thinking. We see the health condition, and therefore the correlated state, as existing separate from the overall system of the human body.

Further, we look at things abstractly because we are good at analyzing problems. We are not so good at synthesizing the results. To become effective at synthesis, we need to start our analysis in a different manner. Systems tools can help us do that.

So going back to the idea of enhancing performance, we need to be able to predict the consequences of the changes we make, but we find that difficult at best. But we can see the beginnings of the answer now. We can see it is not about developing specific “attributes” in system components, or even in the overall system. There are no attributes to deal with!

We should be able to see this because it is obvious…right out there in the open. “Attributes” are notoriously hard to measure. Why? Because they do not exist!

Instead of realizing that they are abstract and therefore not open for inspection and measurement, we say they are “subjective”. So to try and make it more “objective”, we use baselines to measure them. How “shiny”, something is becomes a simple matter of comparing the current sample to a baseline sample (an archetype…think about that for a moment) that has the proper “shininess”.

So, to determine if someone has a strong “leadership” attribute we compare the current individual to baseline examples of leaders. We make it seem more legitimate by doing analysis of the sub-modalities that our baseline leaders use. First, we pick leaders that seem to have success at leading. Then we look at their “attributes” such as communication, vision, accountability, drive, interpersonal skills, etc. Now we have something concrete right? NOT! We are simply continuing the chain by assigning attributes to an attribute we call leadership. By making it more complex, we think it has more validity, but we have missed a basic aspect of the problem that systems thinking reveals.

Think deeper. If systems drive behavior, it is due to the interactions of the parts. As soon as we start separating attributes from system components, we have made it impossible to see the solution. Our Aristotelian mode of using abstract thinking has already led us astray. Shininess only exists as a part of the system that is the object, its composition, its state, and my eye, the amount of light, the ability of my eye to focus “properly” within the context of what I am trying to “measure” (the shininess). So the same is true of leadership, team-player-ness, motivation, skill, etc

The great Zen philosopher Alan Watts offered a profound example of why we fail to see systems in our environment. He said that we confuse separation with differentiation. The hand and the foot are different, but not separate. A bee and a flower are two aspects of a single organism; different, but not separate.

I like that. By using the term separation, was he referring to abstract thinking? Not sure, but it fits. More importantly, it gives us a way of visualizing systems and a simple way to catch ourselves, or even prevent ourselves from going astray.


Where complexity science comes into play, is that we can start to see that system behavior can change “instantly”  or transform if certain quantities within that system or acting on that system exceed some threshold. We are talking about phase transitions, like water changing its state to ice or vapor as temperature changes. Thresholds could refer to a quantity of something real, like an amount of a substance or amount of energy, or virtual like the level of complexity.

This effectively moves us out of the realm of cause and effect, and therefore out of the realm of time (for all intents and purposes). Without cause and effect, does time come into play at all (quantum physicists take note)?

Applying these concepts to business management, performance management, relationships, the economy, etc, can open up many possibilities for understanding and developing new knowledge.

So essentially, we use Systems Thinking to find the lever points. We use Complexity Science to understand if our nudging can result in a phase transition to another state. Phase transitions could be in either direction, increased complexity, or reduced complexity. For example…at what point does a mom and pop shop, become a hierarchical mega-corporation? Specifically, could a mega-corporation operate as a collection of mom and pop-level entities? Same people, different phase-state.

Attributes (Definition of)

Example: Performance is not an “attribute” that individuals posses, nor is it composed of other attributes. Performance attributes are a metaphor that makes it possible to assign responsibility to the individual for their “level of performance”.

Note that by assigning responsibility to the individual for their performance effectively removes responsibility from others (such as their manager). Although this is useful for “getting things done”, we take thought out of the process.

By removing thought, we can say if someone is performing poorly they simply do not possess enough of a given attribute. If they had more talent, skill, motivation, leadership, followership, aggressiveness, passivity, etc., they would surely show better performance. Remediation then involves coaching, mentoring, training, bullying, coercing, etc., the person into acquiring the needed attribute(s).

But attributes only exist in relation to something else. It could be a perspective, a set of circumstances, an environment, a technology, etc. So they can seem valid within a given system . But definition of a system depends on drawing somewhat arbitrary lines between “it” and everything else. And the individual is only a small component within that system. So everything about an attribute is contextual.

By saying someone has or does not have an attribute, we are working within the Aristotelian framework of linear straight line thinking (see section on non-Aristotelian thinking). Example: Performer A met their financial goal for the measuring period, while performer B exceeded theirs. Is B a better performer? You do not know without the context.

Metrics can fall into two large categories or types. They can be high-level indicators like an index, or they can be lower level detail metrics. Context is when there is a correlated relationship between them, as well as a correlated relationship with something external. This starts to look like an analog of the real system. It should display actual system behavior. But what happens in practice is that we do not use an analog, we use discrete metrics that remove the context.


Sections below are in process. Contributions from readers of LeverPoint are welcome and requested! Add a post to the blog and I will bring the information into the relevant page.


Attributes In Practice

Performance Management

Performance “Improvement”

Attributes in Theory

Complexity Science

Fitness Landscapes
Learning Agents
Simulated Annealing

Systems Thinking

Analysis and Synthesis


Operational Language
Operational Thinking

Non-Aristotelian Thinking

Rational Problems and Solutions
Irrational Problems and Solutions

Theoretical Application

The “No Attributes” approach is preemptive. Problems never become an issue.
What’s My Motivation? (Archetypes and Landscapes)
Common Attribute Sets
Corresponding Performance Types

The Planner


Project Success


Internal (Built in) Causes

External (Special) Causes


The Slacker

Everything is an Emergency

Lack of Recognition

Trust and Teamwork

Punishment and Consequences

Failure to Complete



How can you use the new paradigm to improve performance of individuals and systems?

If not attributes, then what?

Is there “good” VS “Bad” system performance?

Isn’t optimized just another word for good?

How to you identify the lever points within the system?

How can I model the behavior of a change to the system or one of its components?

If emergence is not predictable, then how do I know it will be good?

If I can model system behavior, why can’t I predict emergent behavior with modeling?

Is there a point of diminishing returns with this approach? Is it just a different perspective that does not provide any practical value?

Isn’t this just X repackaged?

What do you call this new paradigm?


© Copyright 2008 Keith Sherwood


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

%d bloggers like this: