In this chapter we will be talking about complexity theory, or Complex Adaptive Systems. The reason I am so interested in complexity theory is because I believe it represents something fundamental about the nature of the universe. Once you understand how complex systems work, there are suddenly many areas of life that will make more sense and be easier analyze. So, what is complexity theory? Well, we are going to go back down the rabbit hole into Wonderland to explore by analogy. If you remember, last episode ended with you escaping the maze, equipped with philosophy as the tool which will help you navigate the world. This means using your own senses and mind to perceive and integrate your surroundings. But what is the fundamental nature of the world we are trying to approximate and understand? What are the underlying mechanisms which determine what can and does exist? How do they work? Well, let’s explore a little...
Part #1. Mathematical Efficiency
You walk out of the maze and are greeted by a vast forest. Not knowing where to go, you decide to go straight ahead until you encounter something of interest. However there isn’t really anything interesting at all, just trees. So, after a while you decide to study one of those. You plop down on the ground with your back up against a trunk, looking up at the canopy of branches above you. As you examine the web of interlocking lines you start to notice the rules which give the trees their form. How the same patterns repeat over and over again: each branch grows, sometimes with a few offshoots, until eventually splitting up into two smaller branches, and then then the same process repeats, all the way up to the top of the tree. Going from one to two to four to eight. And each time a split occurs, the overall proportions remain the same. You can imagine how cutting off a branch at the top of the tree would create a much smaller version of the whole, maintaining the same trunk-to-branch ratio. Except, of course, for the leaves.
But the leaves too, you notice, picking a fallen one up off the ground, follow the same patterns of self-similarity. The stem turns into a vein that runs down the centre of the leaf, and from that main vein smaller ones shoot off on either side, and then even smaller ones grow off of those. Then the pattern seems to be lost, as the capillarities lose their more formal arrangement and turn into seemingly random networks. But are they? As you hold the leaf up to your eyes you notice that the dry mud cracked on the ground beneath you shares the same pattern. With the cracks, like the capillaries, always meeting at 3 or 4 point junctions, breaking up into smaller and smaller subsystems. Now why would this be?
As you consider this, you hear a low rumble as the skies of Wonderland quickly turn overcast and a thunderstorm begins to roll in. You watch as the first bolt of lightning shoots down. However instead of a quick flash, the skies of Wonderland allow you to see the entire process in super-slow motion. You watch as the lightning branches down from the sky in a pattern that looks almost identical to the trees growing in the forest, with a network of leaders splitting up and trailing down in different directions. The leader that makes contact with the earth first then sends a second, much stronger bolt back up into the sky. What is going on here?
The electricity is stretching out, seeking the path of least resistance. And once a connection is made, lightning strikes. The same process is happening to determine how the drying mud cracks in a way to relieve the most tension, and the capillaries on the leaves form to allow for the most efficient transportation of nutrients. So too are the trees, governed by this process which is driven by ease. Other examples would include the hives of honey bees, or soap bubbles drifting in the breeze.
What you are witnessing are the manifestations of mathematics in nature. For the same reason that plants grow in fibonacci spirals, all of the natural world operates in a way that seeks to minimize energy expenditure; getting the most bang for your buck, if you like. This means that many seemingly disparate phenomena are regulated by the same underlying drives and mechanisms. This is why the trees, and the leaves, and the cracked mud, and the lightning all share some similar features. The patterns that you see repeating everywhere are not accidental, but innate. This is due to a few very simple decision making rules which lead to intelligent, complex behaviour. And that’s all math is really, simple rules with complex consequences. Although not comprehensive complex adaptive systems in and of themselves, manifestations of mathematics in nature provide key insights to some aspects of how complex systems work.
Part #2. What is a Complex Adaptive System?
Complex adaptive systems are systems composed of individual agents which interact with one another as well as their environment, giving rise to emergent behaviour which cannot be reduced to the sum its parts, and they can be found almost anywhere. From the activity of ant colonies to the formation of sand dunes, the evolution of plants and animals, the motion of flocks of birds and schools of fish, the dynamics of ecosystems, even the weather. And that’s only in the natural realm! Complexity is even more abundant in the social sphere. Some examples would include financial markets, traffic jams, the rise and fall of culture, the evolution of language, the internet, social movements, academic citations, machine learning algorithms… it would probably be easier to list the realms of human activity that aren't governed by complex systems, rather than the ones that are. But what’s really interesting about complex systems isn’t where they are (because you can find them almost everywhere), but how they work.
I’ll try to explain through the classic example of an ant colony. In any complex system there are three main components: agents, drives, and signals. In an ant colony, the ants are the agents, the food they search for is the drive, and the pheromones they leave behind are the signals. A lonely forager ant will wander along until it discovers a source of food, which it will then collect and bring back to the colony. But it leaves behind a pheromone trail which can be picked up by other ants to guide them to the food supply. As more ants discover this source the signal will become stronger and recruit even more ants, until all the food is harvested. After this point, new ants that follow the trail only to discover no food will return home without leaving a signal, therefore dampening that trail over time and allowing for new food sources to be discovered. Now, if all of the ants in the colony directed all of their attention to only harvesting this one resource, better, more fruitful alternatives may be overlooked. Thankfully, ants are not very intelligent and some will wander off of the beaten path, allowing for the possibility of a better alternative to be discovered. In this way, complex systems use random error to their advantage. As Nassim Taleb would say, they are antifragile. The imperfection of their agents makes them stronger, rather than weaker.
Another example of a complex system would be how memes spread, both in real life and on the internet. Here, the users would be the agents, the drive is entertaining or interesting content, and the signal is how many likes or shares a piece of content has. Some funny image that is posted to 4chan will be picked up by community members and then shared to other online spaces, where the same process reoccurs on larger and larger scales until it eventually hits the front page of reddit and is being talked about by late night hosts. But of course, a given meme doesn’t stay relevant for long, so the same process must keep recurring as internet culture evolves. This is why they are called complex adaptive systems. They are not static but constantly changing and evolving in relation to a given context or environment.
I should also mention that complex systems are not that same thing as complicated ones. Complicated systems have linear, 1-1 cause and effect relationships, like the inner workings of your car. There may be lots of different parts and components working together to produce a given outcome, but the parts are not interrelated. You slamming on the breaks of your car will have no impact on your windshield wipers. But in a complex system, it would. Complex systems are nonlinear, meaning that the agents of the system are interrelated in a way that makes causality much more difficult to track. A small change at one point in the system can create a ripple effect that produces much larger consequences in seemingly unrelated areas.
This idea of nonlinearity has another important consequence—complex systems are entangled. Their components cannot be broken apart and studied in isolation as traditional scientific methods would attempt to do, information is lost in the process. It is not just the agents you wish to study, but the relationships between them. Imagine trying to understand how an ant colony operates by taking each ant and observing it in isolation, I doubt you would get very far. The interactions between the agents have meaningful consequences. While made up of individual components, the system is not simply the sum of its parts. It has certain unique, emergent characteristics.
So how can we study complex systems if the whole is too specific, and the parts are too general? Well, they are all mediated by and subject to certain governing features. Despite taking on such different forms as social networks and sand dunes, there are certain consistent dynamics that all complex systems share. If we can understand one system and the mechanisms which govern its behaviour, it becomes much easier to take those insights and apply them elsewhere.
Part #3. Governing Features
Different complexity theorists will all have slightly different opinions about what characteristics define a complex system, their governing features are often overlapping and interrelated. I have tried to select a few key concepts to outline and define, and hopefully by understanding these ideas you can begin to develop an intuition as to what complex adaptive systems are and how they work.
To start, complex systems are self-similar. This is concept I already alluded to at the beginning of the chapter back in Wonderland; the fractal-like way the same patterns and features can be observed at different levels of analysis, zooming in or out. Trees composed of “mini trees” and lightning bolts composed of “mini lightning bolts”. Similarly, financial markets or weather patterns can be analyzed and understood on either a global scale or by country or city. Each operating by its own internal dynamics while simultaneously being subject to influences from the broader systems it is nested inside of. You can take this same idea and apply it to the internet, where communities are subdivided into smaller and smaller niches. For instance, you can get online and go on social media. Specifically, tumblr. Specifically, fandom tumblr. Specifically, the Doctor Who fandom. Specifically, Doctor Who fan-fiction. Specifically, Doctor Who fan-fiction set in a setting where… you get the idea. In each of these progressively smaller communities you will find that there are certain key players who are well known and have a disproportionate amount of influence.
In fact, influence on any level of analysis will follow power law, or Pareto distributions, where 20% of the agents will account for 80% of the overall impact. It’s important to point out here that when I say agents, I do not necessarily mean people. If you are looking to analyze book sales then books are the agents, not the people purchasing them. We tend to think of normal, or bell curve distributions as being the universal default, but this is untrue. Random distributions only occur when the agents are independent from one another. So people’s heights or shoe sizes would follow a bell curve distribution. However when the agents are interdependent, like in a complex system, this is not the case. In fact, complex systems are so ubiquitous that if you are ever unsure of a statistic, you can likely make it up with a relatively high degree of accuracy. For instance, 20% of books published make up 80% of sales, 20% of world cities hold 80% of the global population, 20% of roads get driven on 80% of the time… you get the picture. But why is this the case?
I like to call it the “law of gravity,” or, that which has, gains! The rich get richer. If two actors of equal talent go in for an audition and one lands the part, then they are more likely to get more acting jobs in the future. Having one major role under your belt increases the odds that you land another. A book that ends up on the New York Times bestseller list is bound to sell more copies than one that doesn’t. A stronger signal, be it book sales or pheromone trails, means that more agents are likely to go down that path in the future. This is what we call a feedback loop. Understanding how these dynamics work can help us become aware of their potential shortcomings. It is completely possible for a feedback loop to start perpetuating inefficient or outdated systems. For example, the QWERTY keyboard. QWERTY is an intentionally inefficient typing system, designed back in the day of mechanical typewriters where typing too quickly could jam the keys. Now, this is no longer an issue, but since everyone has learned to type on QWERTY keyboards their legacy has remained, and we are all much slower typers than we could be as a result of it. Efforts to introduce better alternative typing systems exist, but established norms can be hard to get away from when we’ve all grown accustomed to a certain way of doing things. Being mindful of how feedback loops work can help us to disrupt them when necessary. But we have to know why we are doing it. What changed?
This goes back to the idea that complex systems are adaptive, meaning that the behaviours of a system will adjust and change over time. This can either be in response to a shift in the system’s internal state of affairs, or the external surrounding environment. The most obvious example of this would be Darwin's theory of evolution by natural selection. A common misconception about Darwinism is that Darwin believed in “survival of the fittest,” while really what he championed is, “survival of the best suited to the environment.” This more nuanced position takes into account the fact that the environment is constantly changing and evolving. Therefore “fitness” cannot be conceptualized in a vacuum. It must exist in relation to a surrounding context. This process works through variation, selection, and retention. For biologists, this refers to genetic mutation, but it can also apply to ant trails, book sales, or scammy emails.
A bot that has been programmed to send emails soliciting money will probably vary the types of messages it sends, and then generate more messages depending on which style is the most successful. As people learn not to trust mysterious messages from Nigerian princes, the program will then have to adjust its behaviour and employ new tactics. That being said, there are certain attractor states in all complex systems that adaptations will gravitate towards. In the realm of evolutionary biology, you could consider camouflage to be a good example of this. Certain patterns and colourings will consistently crop up in insects, amphibians, and other animals. Despite having a dramatically different evolutionary heritage, some outcomes are simply more effective than others. Although again, potential degrees of freedom always exist in relation to a given environment. In some circumstances, a multiplicity of equally viable attractors is possible, while other contexts demand a singular solution. An example of this might be the path you take to cut through a forest. Depending upon the obstacles at hand, there might be one ideal path which is clearly the most efficient, or there may be a few different ones that are all equally viable.
Complex adaptive systems are also nonlinear, meaning that the size of an effect can be much greater than the initial cause. Think about how a rock jostled from the right nook on a steep cliff will cause an avalanche, or the notion of the straw that broke the camel’s back. There can be certain tipping points in a complex system where a small, seemingly irrelevant push will have very large consequences. You may have heard of this phenomenon referred to as the butterfly effect, where it is said that the flapping of a butterfly’s wings can cause hurricanes; how some small difference in initial conditions can produce dramatically different outcomes. This is because, due to the feedback loops mentioned earlier, a small cause can compound into a much greater effect. Maybe one ant wandering north rather than south discovers a food supply which ensures the colony having enough food to survive the winter. Or maybe one upvote on a meme makes the difference between it going viral or never being seen again. We like to think that the consequences of our actions are limited to ourselves, but complexity theory suggests that the stakes may be far greater than you could possibly imagine.
Complex adaptive systems are also bottom-up, meaning that they are driven by the simple decisions of individual agents. If you have ever wondered how flocks of birds or schools of fish move in such beautiful patterns, here is your answer. There is no head bird or fish that is calling the shots, each one is simply reacting in accordance to the actions of their immediate neighbours while maintaining a certain speed and distance from the others. Simple decision making rules can compound and lead to complex consequences. A great example of this is how researchers have started using slime mold to map roadways. This unique type of mold will, like ant colonies, spread out and create networks in relation to available food sources. A team of researchers took oat flakes and arranged them in a pattern that mimicked cities surrounding Tokyo. Within a few hours, the slime mold had spread throughout the flakes and taken on a shape that looked nearly identical to Tokyo's subway system. This subway system had been meticulously designed by a team of expert engineers, and yet some unintelligent slime mold was able to replicate their results with an amazing degree of accuracy. Without any centralized control or planning, an organized, efficient outcome was achieved. But how is this possible?
This leads us to the final and arguably most interesting component of complexity theory, which is that complex adaptive systems have emergent characteristics which cannot be reduced to the sum of their parts. Often, this emergence can look like some form of intelligence, self-organization, or coordinated control. But you will not find any of these features located in the individual agents. They only come to exist through the relationships and interactions between them. However once these emergent characteristics come about, they possess their own unique causal power. We understand this pretty intuitively in terms of culture: there is no government body or top-down control determining what music, food, or fashion is going to be popular. Our preferences are determined by the people around us. But once a certain fashion trend or fad food takes hold, it can in turn impact the behaviours of individuals. Causality begins to work in two directions. People make culture, but culture also makes people. I think this phenomena of emergence has interesting implications in terms of understanding consciousness, but more on that in the next chapter. Now that I’ve outlined some governing features of complex systems, hopefully you are beginning to develop an intuition for what they are and how they work.
Part #4. A Metaphysics of Complexity
So, to tie all of this back to metaphysics: I do not wish to argue that all aspects of reality are innately complex adaptive systems, however they do consistently give rise to increasingly complex behaviour. They become more organized and more structured as time goes on, rather than less—and they can be found almost everywhere, both in the natural and the social realm. Understanding how these systems work increases our capacity to think critically about the environments we find ourselves in. The mechanisms that brought about their creation as well as the potential shifts that could lead to their undoing. There are consistent rules at play, but they lead to contingent and ever changing outcomes. For this reason, an emphasis on complex thinking demonstrates how any truth claim always exists in relation to a given context. But those truths emerge out of, and are regulated by, certain higher order systems and mechanisms which are governed by consistent principles. This is the case I want to argue, particularly because I think it has extremely interesting consequences when it comes to the problem of what I like to call, “nesting objectivism and relativism.”
If you recall, last episode I described metaphysics as asking, “are there firm and stable laws which regulate the universe? Or are things entirely relative and relational? Is there a consistent, coherent structure to reality? Or is everything random and chaotic?” These questions essentially boil down to the timeless debate between objectivism and relativism. Now, the beautiful answer that an understanding of complexity provides is: both. There is fix, and there is flux. If you like, you can reframe these two seemingly contradictory stances in terms of the Taoist vision of Yin and Yang, chaos and order. “The Way” is being able to traverse that fine line down the middle, straddling both extremes. There is no rigid, consistent, perfect, Platonic form, but there are certain rules which govern how and why any given form manifests. Although, these motives might not necessarily be identifiable through so many entangled threads. Therefore thinking complexly demands much more nuance than traditional interpretations of reality, precisely because things can never be boiled down the simple matter of either/or. There are both objective truths and relative contexts at play.
My ultimate metaphysical argument is that we must focus on the broader processes that determine how reality manifests, rather than any particular manifestation. Specific structures matter too of course, since they represent the very real constraints that you must learn to navigate, but they are much more context and path dependent than the meta-structures which determine their form. To go back to the metaphor of the maze, the particular form of your maze isn't as important as the rules which govern all mazes. And the correct principles of solution will allow you to navigate any given maze successfully. Understanding the underlying generative rules is always going to be more valuable than mastering a specific manifestation. We have a natural intuition for this idea, concept formation is all about abstracting away from specifics towards more general principles. What makes a video go viral? Why does a tree take on a certain form? We are not interested in the reasons underlying any specific instance, but the mechanisms which are consistently at work upon repeated iterations.
I think employing complex thinking is an extremely useful analytical tool, especially in social studies where natural scientific methodologies are no longer viable. For instance, try evaluating a truth claim like “pink is a girl’s colour.” Well, we all know this be true and untrue simultaneously. Obviously no colour itself has innate, objectively gendered characteristics. 100 years ago blue was the predominate colour used to communicate femininity. However we can trace an objective (if not complex) path through history, fashion, and advertising to determine how “pink is for girls” came to be a “true” statement.
Another example would be how everyone learns in history class that the assassination of Franz Ferdinand kicked off the start of the first world war. You could conceptualize this event as a “tipping point” that set off a domino chain of reactions producing an effect much greater than the initial cause. After the first world war, the consequential Treaty of Versailles put Germany in a weakened economic position, causing a growing support for the Nationalist Socialist Party in later years, eventually resulting in the second world war. But linking Franz Ferdinand’s assassination directly to the Holocaust is much tricker. There were many underlying causes for the Nazi’s rise to power, and history is rife with competing interpretations as to which events held more causal power in determining a given outcome. Maybe the Holocaust was an “attractor state” (although I shudder at the thought) which would have happened with or without his assassination. It is impossible to know, there are too many entangled threads.
The character Martha from Doctor Who understood this notion of causal complexity on her first time travel adventure, expressing concern that stepping on a butterfly may lead to a compounding series of unintended consequences. Later in the show, the Doctor reveals that this is not an issue since there are fix and flux points in time. I think this statement may get at the heart of the idea I am trying to express here. In any system there are going to be certain attractor states which are inevitable products of the systems themselves, and other areas full of wiggle room. The notion that pink is for girls obviously falls into the latter category, this is a path dependent outcome and probably wouldn’t arise upon repeated iterations. It is important to understand how these processes unfold so we can try to evaluate which is which in a given scenario. Another gendered truth claim is that, on average, women are more interested in people, whereas men are more interested in things. This idea can help explain why men and women tend to have different career interests. And in societies that have worked the most to promote gender equality, these differences are heightened rather than diminished. This would suggest that these differences are not the result of path dependent feedback loops, but are instead the product of hundreds of thousands of years worth of evolutionary psychology. In other words, it is an attractor state representing a much deeper truth.
In the broadest of strokes, I wish to posit that the notion of “truth” always exists in relation to a given environment, which itself is constantly shifting and evolving. But there are consistent rules which dictate a contingent structure. Relativists, or postmodernists, emphasize how truth claims are contextual, and ever changing. While Objectivists, or realists, care about the consistent rules which cause a given truth claim to emerge. They are both attending to only one half of a much broader picture. We must be able to synthesize and acknowledge how both of these forces work in tandem simultaneously. Neither can exist in isolation, they are two parts of a greater whole.
So where do we go from here? If my metaphysics is grounded in complexity, then what implications does that have for my epistemology? How can and should we go about acquiring knowledge? Well, as I teased earlier, a special feature of complex systems is that they have emergent characteristics. This notion of emergence has dramatic implications on what we are capable of knowing. The philosophy of science therefore becomes of utmost importance, and critical realism is a philosophical approach that I believe contends with these issues most persuasively. So, in the next chapter I will begin to unpack critical realism and discuss what implications it has on scientific inquiry.
Continue to Part 3: Emergence & Epistemology
Wonderland is a free publication that outlines a philosophy and political vision for the 22nd century. To support my work and receive updates on new posts, sign up for a free or paid subscription.
1.2 The Metaphysics of Complexity