 In this section to the course, we'll be starting the major theme of non-linearity and self-organization within social systems. We'll use this video to give an outline to the overarching process of self-organization and then in future modules dig further into the different topics covered here. Self-organization is really a type of pattern formation, a means through which some form of order or coordination is developed. There are essentially just two basic methods through which social coordination and order can occur. Within linear systems it may be imposed in a top-down fashion from some centralized global authority, or within non-linear systems it may emerge from the interaction of the agents on the local level in a bottom-up fashion and this is what we call self-organization. As such, self-organization is a non-linear process of pattern formation. We will be defining more precisely what we mean by this term non-linearity in a coming module, but here we'll be using it to refer to distributed interactions. Self-organization is a product of these distributed local interactions. Within a linear system where there is a low level of connectivity and relatively few components interacting in a well-defined linear fashion, it is possible to control and coordinate that system through some centralized regulatory mechanism. We can use this centralized governance mechanism to impose or maintain order within the system. That is to say, by influencing or controlling the agent's choices towards a coordinated outcome, we can get some state of order within the system. We can only have this form of centralized top-down coordination when a relatively large percentage of the interactions are being routed through this centralized coordinated mechanism. But this top-down form of regulation and control is only really possible within the linear systems. As we turn up the distributed connectivity, the number of components, and their capacity for autonomous decision-making, the system will become more and more difficult to coordinate from some centralized location. And it will become easier and easier for patterns to form on the local level through this high level of distributed interactions. Above some theoretical point where we have more non-linear distributed interactions than centrally routed connections, we're starting to get a significantly large enough space that is unregulated. It is in this unregulated space that has sufficient density of non-linear interactions that self-organization can take hold as a significant mechanism for coordination. Self-organization, in contrast to this linear top-down model, is a product of these local non-linear interactions. When I bump into my neighbor on my way out in the morning and say hi, this is an example of a local interaction. These local interactions are often spontaneous as in this example. I didn't plan to meet my neighbor, it just happened. And these local interactions are non-linear in the sense that they typically happen in a distributed fashion. I, myself, have chosen to say hi to my neighbor. I didn't have to go and ask someone for permission. This is a distributed peer-to-peer interaction. And these distributed local peer-to-peer interactions are very difficult to manage through a centralized model. A centralized regulatory model will always have to use some form of abstraction in order to manage the system. Because a centralized model means that very few people are trying to regulate very many. We can only do that by using abstraction. A president of a country with 1.3 billion people can't go around telling each one what to do. There has to be many layers of bureaucracy between them. And information has to flow in a linear fashion out from the center to the periphery. The further we go out, the more people we have and the more possible cross-links we can have between them. Each one of these peer-to-peer links is a possibility for a local pattern to form. So self-organization often happens out at the fringes where the chain of command is weak and there are many local interactions. As a side note, today we see self-organization becoming more of a mainstream form of social coordination because we're increasing these distributed non-linear interactions through information technology. Thus making it more difficult to manage these social systems through centralized methods and easier to get local self-organizing patterns of organization. Either way, we can call this state an unregulated environment and it is the conditions or grounds on which self-organization can take hold. Self-organization is then a form of distributed non-linear pattern formation. All patterns, forms of order or organization are going to involve some correlation between the states of the system's constituent elements. This is essentially what organization is. When there is no correlation, we have randomness, the absence of order. So we have randomness and order which is some form of correlation between the states. Like two people dancing together, a change in the state of one's motion will be correlated with that of a change in the other. Now the dancers on a fundamental level can really only do two different things. They can move together in the same direction which is a positive correlation or they can move in the opposite direction which is a negative correlation. A positive correlation means the two elements states are synchronized. They move together in the same direction. A negative correlation means they are desynchronized. They move in the opposite directions. This dance has what we call a symmetry to it. This idea of a symmetry is at the heart of modern mathematics. During the mid 1800s, we came to understand algebra on a deeper level in terms of symmetric transformations and invariance. We've since gone on to use this within fundamental physics to understand the basic workings of our universe in terms of these transformations because these symmetries and transformations apply to all forms of organization. As such, mathematics came to be understood by some as the study of patterns and that's what we're talking about with self-organization. Patterns of correlation between states that can be understood in terms of symmetries. In applying this to social systems then, we're talking about agents and thus we're talking about correlations between the choices of agents. We're asking, do they choose to do the same thing, opposing thing or is there no correlation between their choices. This is the very basics of what we're dealing with when we talk about self-organization within social systems. Self-organization is then a process that is going to change the correlation between agent states within the system. It is going to coordinate them and this is done through what are called feedback loops. Positive feedback loops have been identified as playing a central role within the process of self-organization. We'll be talking about feedback loops in a future video but a positive feedback loop is one that is self-reinforcing. More begets more. The more products a business sells, the more it can invest in its business. The more it can produce better, cheaper products meaning it will sell more which means it can reinvest more and so on. This is an example of a positive feedback loop. It is a non-linear process of change. Through it, the business can grow in an exponential fashion. Feedback loops are the mechanisms through which some small event which is often random in nature can get amplified into a new macro-level pattern of organization and this is at the heart of the whole process of self-organization. To give a quick illustration of this let's think about a beach of people sunbathing on their holidays. Now let's add some initial random event. We have someone with headphones on listening to their favorite piece of music and they get so excited that they jump up and start dancing around. What happens now depends on the state of the other agents around them. Typically these random events will get dampened down and die out. Everyone will look at the guy like he's weird. But by this person occupying this differentiated state we've already created a feedback loop. It is now much easier for anyone else with a propensity for dancing to jump up and join him and if they do we would now have some distinct pattern. Two people occupying the same state and although they are still a significant minority the positive feedback has got stronger. It is now even easier for the next person to join and with every new person that joins it becomes more attractive for anyone else to do likewise. As this positive feedback process of change continues we'll get to some point where there are more people dancing than not dancing. This has become what we call an attractor. You will now be considered more normal if you are dancing rather than not dancing. If we added a new agent to this system who just wanted to be normal and follow the course of least resistance then he or she would end up dancing. And thus through this process of change driven by positive feedback we now have an attractor. A default set of states within the system. This attractor is the pattern of organization. All these agents dancing has correlated their states in some way. As another example of an attractor we might think about the languages and cultures within different regions. Within any different region there will be a stronger attractor towards speaking the same language and adopting the same culture and behind the creation of these cultural attractors there was a positive feedback loop the same as with our people on the beach. But as we know there are many different languages and cultures in our world representing many different attractors because if the system is large enough this process of self-organization through positive feedback may take hold around a number of different components within the system at different locations and grow outwards from there until it reaches another pattern at which point we get a boundary condition like the national borders in Europe marking the limits to the different cultural attractors that have formed over a prolonged period of time. At this point where all the elements in the system are aligned within local level attractors positive feedback will die out and negative feedback will take hold as the different attractors balance each other out to create a semi-stable configuration. These different attractors then have to compete or cooperate in order to enable some form of global coordination. This type of interaction will largely be a product of how the attractors were created in the first place that is to say were these different local level attractors created out of exclusive or inclusive conditions. We can create social organizations by the individuals overcoming their differences to find common ground, common purpose and identity or we can create this organization by defining our differences and degrading others what is called outgroup derogation which is a form of negative externality. We are creating the pattern of organization by simply exporting the entropy outside of the system's boundary to some other system. For example, German Nazism created their sense of identity around the Aryan race through a systematic derogation of other outgroups including the Jews amongst others. This exporting of social entropy creates division and conflict between the different attractors. By the individual attractors exporting entropy this will create a state of disorder within the overall system. Any group that is dependent upon the exporting of entropy in order to define and maintain its internal state of order will not be able to integrate within the overall state space and conflict will be the primary type of interaction between the different patterns of organization. But of course this will not always be the case we can also get positive externalities the net result being emergence as the different local attractors overcome their differences and we get the emergence of some global pattern of organization. Emergence is a process whereby larger entities patterns and regularities arise through interactions among smaller or simpler entities that themselves do not exhibit such properties as such it is very much analogous to self-organization but subtly different. Emergence is also a distinctly nonlinear phenomena in that it cannot be derived from any one component within the system it is the product of many distributed interactions across the system. An example of this might be a wave at a football match this is an emergent self-organizing phenomena no one is coordinating it some small initial event takes hold and gets amplified into a large macro phenomena. People crossing a street is another example of pattern emergence through self-organization. We have dense distributed interactions as people going in either direction try to pass each other. Those who meet first have to coordinate their activities but once they do this will create an attractor for others to follow as we get the emergence of some macro level pattern distinct streams of people going in different directions. Nonlinear dynamics like the process of self-organization are unpredictable. In these nonlinear systems there is a breakdown of linear cause and effect. If I hit a ball with a bat and you ask me why the ball moved off in the direction it did I can say the cause of that phenomena was me hitting the ball with the bat. This is linear causality and through it I can predict the next time that I hit the ball with the bat it will again move off in the opposite direction. But this is not the case with nonlinear dynamical systems the effect of self-organization that is to say the output to the system is not a product of the input it is an emergent phenomena of the overall state to the system and the feedback loops over time. Almost all real world complex systems are going to have randomness, fluctuation and noise in them. Nothing is perfectly ordered but these small events typically do not affect the overall state of the organization. If we take a large enough society there will always be some people who are discontent with the current state of that sociopolitical system and thus trying to change it but they will typically not gain traction. Thousands or even millions of small events will take place without any effect because of the overall state of the system. But when all of the components come to be aligned within a similar configuration all the agents come to adopt a different perspective than some small event can gain traction and propagate through the whole system. For example we might think about some oppressive political regime. For every act of oppression from the ruler this may not have a direct consequence but it does create resentment among the people. This resentment builds up until we have a critical state all the people are synchronized in their discontent at this point some small random phenomena that may have happened many times before can now propagate through the system rapidly. The system has organized into a critical state and this critical state is systemic. It is distributed out across all the elements in the system and thus we can say that no one cause created the effect. It emerged out of the overall state of the system and the feedback loops that drove it over time. This type of nonlinear pattern formation is then unpredictable. There are many small fluctuations and it cannot be determined in advance which one will gain hold. Thus we cannot know where they will come from because the actual event emerges out of the state of the system and through the feedback loops but those feedback loops play out over time after the event has happened that is to say the outcome does not exist at inception. It is not determined by the initial cause but instead it is created along the way. The academic journal of democracy describes this phenomena as such regime transitions belong to that paradoxical class of events which are inevitable but not predictable. Other examples are bank runs currency inflations, strikes migrations, riots and revolutions. In retrospect such events are explainable even over determined. In prospect however their timing and characteristic are impossible to anticipate. Such events seem to come closer and closer but do not occur even when all the conditions are ripe until suddenly they do. In summary then we've been talking about self-organization as a nonlinear process of pattern formation that requires dense distributed peer-to-peer interactions within an unregulated environment in order to take hold. We discussed social organization as a form of correlation between agents choices. We talked about social organization as a form of correlation between agents choices. How randomness can be equated to lack of correlation whereas order may be understood in terms of symmetries. We talked about how positive feedback loops are the key engines behind self-organization as they can work to amplify some small event into a large systemic phenomena creating local attractors that then have to cooperate or compete to get global coordination. Finally we saw how the idea of linear cause and effect breaks down within these nonlinear dynamical systems. Phenomena emerge out of the distributed state of the system and the feedback loops that play out over time making events fundamentally unpredictable in nature.