Monday, April 13, 2015

The Holy Roman Empire was a complex adaptive system

So a few weeks ago my son and I were watching the Crash Course World History series. In it John Green mentioned Voltaire's famous statement that the Holy Roman Empire was neither holy, nor Roman, nor an empire.

A few days later, I found that the phrase "Holy Roman Empire" was still bouncing around in my mind. It seemed like it was trying to tell me something. The pattern of the name, its rhythm, felt similar to something I was familiar with. After a while I realized that "Holy Roman Empire" sounds a lot like "complex adaptive system." Two seconds later, I realized that a complex adaptive system is neither complex, nor adaptive, nor a system.

I wondered if there might be some connection between these things, so I began to explore.

Magic words of power

First I read about why Voltaire said what he said. Here's a very brief summary:
  1. Pope Leo III crowned Charlemagne as emperor in 800, but this was a purely political move, a quid pro quo. The word "Holy" was not used to refer to the empire until 1157, when it was added to remove any implied dependence on the papacy (as if to say, we don't need the Church to make us holy).
  2. The Holy Roman Empire did include Italy at first, but most of it was located in Germany and France. In fact, in 1512 the name of the empire was changed to "The Holy Roman Empire of the German Nation."
  3. The Holy Roman Empire was for the most part a symbolic institution with little centralized power. It was more like a loose confederation among independent states than a true empire. 
So why did people call a loose secular confederation of independent European states the Holy Roman Empire?

I'm no historian, but my guess is that the name makes use of what I like to call "magic words of power." People have been using such words for thousands of years because they tap into deep currents in society. Think of a word that, if you walked into a room full of people and said it loudly and clearly, would start a fight or a cheer, or would clear the room, and you've found a magic word of power. Curse words are magic words of power, because they attack things we hold sacred -- religion, health, sex, family. Their power comes not from their attacks but from the things they attack.

So, I thought, the three words of the Holy Roman Empire might act as a sort of incantation, a way to draw power from deep sources.

What we want to know

The next question I asked myself was: how did each of these words contribute to the power of the whole phrase? As I thought about each word, I ran into a familiar construct: Robert F. Bales' empirical work on how strangers talk to each other in task groups.

Bales found that the people he observed constantly evaluated others in three dimensions:
  1. Is this person dominant or submissive? (power)
  2. Is this person friendly or hostile? (safety)
  3. Is this person useful or useless (to me)? (effectiveness)
So, calling a loose secular confederation of independent European states the Holy Roman Empire says:
  1. We are powerful. Who's in charge? (Remember: it's the middle ages.) God. We are in good with the big guy.
  2. We are safe. Remember the Romans? They did things right. They had roads and plumbing and great parties. We are the tried and true solution. You can trust us.
  3. We are effective. What unifies peoples, brings stability, and provides protection? That's right, an empire. Don't worry, folks, we have everything under control.
Thus, in the context in which they were used, these three words hit all the "what we want to know" points that would lead people (that is, princes and other local powers) to support the Holy Roman Empire. So it was both an incantation and an advertisement.

By the way, does "safe and effective" remind you of anything? Perhaps something medical? Maybe you've heard the phrase "potent, safe, and effective?" That phrase hits the same three evaluation points.

Now let's consider the term "complex adaptive system." I should probably explain first why it's none of those things.

Complex or complexicated?

Complexity refers to self-organization, where the interactions of (at least somewhat) independent agents lead to emergent properties, patterns that are more than the sum of their parts. Complex systems exist in theory and in simulation, but nothing that actually happens in nature or in human life is purely complex. In reality, organization and self-organization intermingle and interact to create a complex-complicated tangle of aspects and influences.

Consider an ant colony. Each ant follows simple rules, and complex patterns appear. But that's not the whole story. In most if not all social insects, the queen exerts pheromonal control over other members of the colony. Centralized organization and complex self-organization intermingle and interact in their effects on what happens in the ant hill.

The same is true for everything people say "is" complex: cities, traffic, the weather, the mind, the body, and so on. Every real situation has both complex and complicated aspects that intermingle and interact. I call it "complexicated."

Adaptive or reactive?

Is a complex adaptive system adaptive? Sort of. There is adaptation in the world, but maladaptation is just as important a force. Having complexity in a system does not always make it work well; sometimes it can lead to disaster. If you don't believe me, google "ant mill." A circular ant mill, in which ants follow their simple rules into a spiral of death, is a complex maladaptive pattern. (In a cruel example of complexication, you can find instructions on the internet for creating ant mills.)

I don't think it works to use the word "adaptive" to mean "sometimes better, sometimes worse," because that's not what people think when you say it. The popular meaning of "adaptation" is "getting better all the time." An example: Blogger's spell checker knows the word "adaptation" but not the word "maladaptation." Why is that?

A better word might be "reactive." Swarms and flocks react to perturbations. Sometimes they adapt, but sometimes they crash, or they dissipate into something that is no longer a flock or a swarm. The result might be better, or it might be far worse. At least the word "reactive" doesn't imply a "better and better" value statement. Even chemicals react, and nobody thinks every chemical reaction turns out well.

System or frame?

Now to the final word: system. This one's easy. All boundaries are decisions. Ask anyone what is included in the "health care system" or the "educational system" or the "legislative system," and you'll get as many answers as people.

The online etymology dictionary says "system" comes from the Latin systema, meaning "an arrangement," and from the Greek synistanai, meaning "to place together." Wikipedia says that the word originally meant "something to look at."

So the word "system" originally meant the choice of components and the creation of a frame or perspective. It still means that to me. Every time I hear the word "system," I can't help thinking of the archetypal image of a movie director making little frames with their hands as they look around. That's all a system really is, a frame, with some things included and some excluded.

But we run into problems when we try to use the word "system" to mean "frame," as it originally meant. That's because today, "system" no longer means a thing we create. It means a thing that exists, a thing we can rely on. Something solid. Like an empire.

A modern-day incantation

If the term "complex adaptive system" is, like "Holy Roman Empire," an incantation, what is it meant to invoke? If it is an advertisement, how does it persuade? Here is my guess.
  1. We are powerful. Complexity is like magic: it's order for free. Look! All of the awesome (and awful) things in the world run on complexity: hurricanes, the internet, snowflakes. This is powerful stuff, and you're going to want to get yourself some of it.
  2. We are safe. Remember evolution? That's rock-solid science. Adaptation is one of the things that makes evolution great. Evolution led to us, didn't it? You want things to get better and better, don't you?
  3. We are effective. Stop worrying. Complexity is not crazy or unpredictable. It's a system. We have everything under control.
Thus, in the context in which they were used, these three words hit all the "what we want to know" points that would lead people (that is, academic institutions, corporations, and other local powers) to support the Santa Fe Institute (who coined the term).

I don't mean to blame the Santa Fe Institute for coining a term that probably made perfect sense to them. I just wish they could have come up with a name that wasn't so easily misunderstood, given the popular connotations of the terms they used.

How we use it

After thinking about this for a while, I began to wonder how people actually use the term "complex adaptive system." I wondered: If these words are an incantation, to what purpose is its power applied? If it's an advertisement, what does it sell?

So I did one of my tiny Google research projects. I googled "is a complex adaptive system," then tallied up what came before the word "is." I got up to 252 mentions before Google cut me off, and this is what I found.

Let me ask you: is language more complex than knowledge? Is health care more complex than the internet? I don't think so. I don't think this graph reflects what is complex and what is not. It represents in what areas people most feel a need for incantations and advertisements. Evidently for language and for health care, there is great need. What does that say about language and health care? Are these areas in which we need to create power, achieve safety, and gain control? Given the fact that language underpins education, social society, and international relations: probably.

Actually, language is a perfect example of organization and self-organization intermingling and interacting. Yes, every human being has an impact on language, and there are emergent properties involved. But there are also clear organizers in the world of language: dictionaries, thesauri, style manuals, publishers, universities, governments, Google. All of these organizers exert centralized control over the evolution of language. It's complexicated.

Side note: I also searched for "was a complex adaptive system" and "will be a complex adaptive system." I found too few matches to draw any graphs (8 for "was" and 5 for "will be"). What does it mean that we only use the CAS term to describe things in the present tense? I don't know. Something.

You can't whisper with a megaphone

At this point, if I were you, I would be asking: What is your point? Do you think we should stop using the term "complex adaptive system?" 

Of course not. Every public-facing campaign needs a slogan, a tasty sound bite. The Holy Roman Empire needed one, and the Santa Fe Institute needed one. "Complexicated reactive frames," or some other more accurate term, doesn't have the same punch.

The danger is not in the fact that people need to create slogans to sell ideas. The danger is in the fact that people come to believe that the slogans are accurate descriptions of reality. They're not. Slogans are like megaphones. They carry messages fast and far, but they also distort.

From what I've seen, popular usage of the term "complex adaptive system" is distorted. I'm not sure how usage of the term started out, but today it appears that anything with even a hint of complexity in it qualifies to be called a complex adaptive system. The term has degenerated into meaning nothing more than "something I'd like to talk to you about."

I've been listening to people say "X is a complex adaptive system" for a while now. I haven't looked into this systematically, but my sense is that when people use the term, they tend to mean one of three things:
  1. You need what I'm selling. 
  2. We need to change the way we do things.
  3. I know a lot, so you should let me handle this.
If you've been reading carefully you will have noticed that these statements match Bales' evaluation points: power, safety, effectiveness. In other words, the slogan is being used as a slogan. It is not being used as an explanation or an exploration. People are using it to get other people to sign on to things. If people were using the term to explain complexity, my graph above would have a different shape, because everything on that list has some complexity in it (and lots of things that are not on the list as well).

Here's what I think. Anyone who uses a slogan to sell an idea bears a responsibility to understand when it's time to pick up the megaphone and when it's time to put it down and speak carefully and quietly. At some point we need to move past slogans to explanations that are useful and informative. If we never move past our slogans, our efforts risk becoming caricatures of themselves: full of sound and fury, signifying nothing.

Paired images

I hate it when people write critical, demanding essays that don't include positive solutions. So I sat on this blog post for a while, waiting for it to become more hopeful. One evening I noticed a book my husband had bought, called Electronics for Biologists, by Timothy J. Gawne. I know this sounds unrelated, but bear with me.

The first chapter of the book starts with a paired set of images: a photograph of a battery connected to a resistor, and a circuit diagram representing the elements in the photograph.

First Gawne describes the items in the photograph:
Voltage varies continuously at all points, both through the wires and inside the different parts of the battery and resistor. Currents flow through the wires and components in complex patterns, much as water flow varies with position in a river, and some current leaks out of the wires to flow through the air and table-top. The circuit elements act as antennas, and so external electromagnetic sources like radio transmissions or electric lights affect the operation of the circuit. The currents flowing in the wires create external magnetic fields, which can interact with other objects. Different parts of the circuit that are not directly connected to each other can nonetheless affect each other via the electric field.
I love this description. It captures the messy, fascinating details of real electromagnetism. This is the kind of thing people should be learning about complexity. That it's rarely the only phenomenon taking place. That it can be miraculous and disastrous. That what is seen depends on who is looking. That what looks like self-organization can turn out to be organization, and vice versa. That the state of affairs can switch unexpectedly between placid stability and wild turmoil -- or not switch, for far longer than you thought possible. Some of the most exciting things about complexity, to me, are the messy, fascinating parts that never make it into a simple term like "complex adaptive system." Maybe that's why I don't like the term, because it's complexity with all the fun parts sucked out.

The second image in the electronics book is of a simplified circuit diagram. Here's how Gawne introduces it.
However, the simplified model circuit in panel B is easy to analyze. We say that this is a lumped quasi-static model. Lumped, because the full complexity of the 3D geometry is reduced to simple lumped elements connected by uni-dimensional nodes. Quasi-static because, even though such models can handle time-varying signals, they do not model the true dynamics of how electric and magnetic fields interact through space.
In panel A [the photograph], voltage is a continuous function of space. In panel B [the circuit diagram], there are only two voltages: the voltage at node 1 and the voltage at node 2. In panel A, the resistor is a physical object with real size and where current and voltage can vary in complex ways inside it. The abstract resistor in panel B has no internal structure at all, and is completely specified by its value of 1000 ohms, and by the fact that it connects node 1 with node 2. The model resistor has no other properties.
Likewise, the term "complex adaptive system" is a lumped quasi-static model for the ways in which complexity manifests itself in the world. It's like a circuit diagram: abstract, with no internal structure and only two states (complex or not; in the system or not).

I have a question for you. What would happen if people laid the messy, fascinating reality of complexity next to its simplified model and showed how they were similar and different? Would that give the model more power or less? More, I think, because people would then see it for what it is: a stepping stone, a way point on the path towards greater understanding, a tool.

Gawne explains that the circuit diagram is useful in certain contexts.
For all its simplification, the model can often provide very accurate predictions of what the real circuit will do. When is the model good enough, and when does it fail to give a sufficiently accurate model of the world? This is a tough question with no precise answer. 
He goes on to examine specific cases where a circuit diagram does fail to provide accurate predictions, such as when frequencies are high, or changes in voltage are rapid, or the circuit is physically large, or when external magnetic fields are strong. I would love to see people do that with the term "complex adaptive system." The places where the term breaks down are well known and easy to summarize (and fascinating).

I'm not saying an electronics textbook holds the key to explaining complexity. I might have been able to find a similar pairing of reality and model in dozens of other books about any number of topics. I just happened to pick up this book. But what it says to me is: It's not necessary to give up on slogans, and it's not necessary to give up on helping people understand what's really going on. The way past slogans lies in juxtaposition, in paired images. We shouldn't give people only slogans, but we shouldn't give them only the messy, fascinating details either. When it's time to put down the slogan megaphone, we can give people both views, side by side. We can help people use slogans when they are useful and put them aside when they stop being useful.

The ideas of complexity were once confined to a small number of scientists, but now they are roaming in the wide world, and they are being widely misunderstood and misapplied. I don't know if this is bothering anybody else, but it's bothering me. I think this could be a way out. What do you think?


Victor Newman said...

Perhaps the problem with modelling complex adaptive systems is that they tell us more about the individual using the term than the system being modelled?

Cynthia Kurtz said...

Hello Victor and thanks for the comment! Yes, I do think CASs tell us more about the individuals doing the modeling than the system being modeled. But I don't think that's a problem, any more than, say, the fact that we (usually) have ten fingers is a problem. Like my father-in-law used to say: wherever you go, you take yourself along. The problem only arrives when we FORGET what we are and how we think, when we take what we see for what there is. The way forward, I think, is to consider both model and modeler. After all, every telescope and microscope contains a mirror.