I posted earlier on emergent causality. One aspect that needs to be elaborated on is the concurrent, self-interdependent nature of emergence, or in other words the chicken and egg problem.
When agents emerge, the dynamic processes involved in their emergence sharpen simultaneously via a feedback of information from the higher level to the lower.* There is no use in asking which happened first – dynamics at level 1 or emergence of level 2 – for they are dualisms, each reinforcing the other until at some point we recognize that something happened. This runs counter to the Western understanding of causality, which requires that we fix one or the other as the cause so that we might see how the effect came into being. But this fixture destroys comprehensibility just as a holding onto a ball in the middle of a juggling act causes the whole procession to come crashing down.
Let’s look for a moment at the emergence of a particular social agent, a corporation. In the beginning there is no corporate agent. There may be an idea (in the mind of a person) to create it. Or there may be several people who in going about their lives – be they working for an existing corporation, exploring a new business concept, etc – come to the conclusion that a new entity should be formed. But this still does not qualify as the actual formation or creation of a new entity. Some of the first acts founders typically do include the creation of a written business plan, naming the company, creating an organizational chart, filing incorporation documents and much more. But none of these individually or collectively are essential for the emergence of the new agent, nor is the order of their happening. At these early stages, agency is weak; the outside world barely registers, let alone validates the company’s existence. Even the founders struggle to take seriously the validity of their enterprise. What if it never gets off the ground, never brings in revenue, never turns a profit? But at some point between the lightbulb being turned on and the largest company in the world, we all would agree that General Electric somehow “came into being” as a real thing. It emerged. It is unequivocally (now) an agent that acts to perpetuate itself, improve its wellbeing, protect itself against threats, set and achieve goals, communicate with other agents, sometimes cooperating with them, sometimes competing, selling to them, hiring them, merging, spinning off, etc.
The corporate agent is a cooperatively emerging one, dependent on its constituents somehow aligning their own goals and actions with one another. At first, when it’s just a few close friends or colleagues, this is easy. But as time goes on and new members are added, new functions, sub-goals and sub-plans are created, corporate agency develops and subtle conflicts naturally arise. Conflict between individuals who have different self-interests and whose interests are not entirely aligned with that of the company. Functional components (departments, processes) and corporate goals, once singular are now multiple and not always harmonious. So structure is created (hierarchy, policy, rules, procedures and decision-making processes) to mediate this conflict in service of one main thing: corporate agency. This is an example of information at the higher level (that of the corporate agent) being fed back into the lower level dynamics, constraining constituent agents’ actions to those that promote corporate wellbeing. This is self-reinforcing too. Structures and dynamics which do lead to corporate wellbeing get solidified and amplified, pushing out those elements which do not, further reifying the higher-level agent, and so on. The family health coverage and equal opportunity policy each work to the benefit of certain members and at the expense of others, but both work to the benefit of the company as a whole. These are examples of dynamics which simultaneously promote corporate agency and which are a result of it. To look at the dynamics as prior to agency or vice versa misses a crucial understanding.
The chicken and egg are dual aspects of another type of emergence: autocatalysis. At first glance, autocatalysis and cooperation don’t seem to have much in common, but they share an important characteristic. Both are coherent dynamics that align the existence of two or more agents, just as waves in phase with one another are said to be coherent. Cooperation is about the coherence of agents that exist simultaneously (but at different points in space) whereas autocatalysis is about coherence over time. A begets B which begets C which… begets A. Note that an autocatalytic system might also translate (i.e. move) its constituent agents in space, but it doesn’t need to. In the Game of Life, gliders are autocatalytic systems that move their constituent agents (and thus themselves!) in space, while blinkers are ones that do not.
To see coherence more clearly, it helps to look at systems with decoherent and incoherent dynamics. In the former case, one agent’s existence is negatively correlated with another’s, for example matter and antimatter particles. As we all know, when electrons and positrons meet they annihilate one another. A system with incoherent dynamics is one in which agents coexist and possibly interact, but do not destroy one another or help one another survive. They are coherently neutral. A closed system of gas particles bouncing off one another gets the picture across. If coherence is represented by an upward pointing arrow, and decoherence a downward pointing arrow, then incoherence is one that points orthogonally sideways. Of course, real systems come in many flavors and the arrows more than likely are at non-right angles. Even “purely competitive” systems like marketplaces and predator-prey ecologies are far from decoherent. Market competitors agree to cooperate on the institution of trade itself, and predators survive the meeting, and eating, of their prey (quite nicely, thank you very much).
Emergence itself is tightly coupled with the concept of coherence. And these are also both coupled with the concepts of agency, system, complexity and existence. Alex Ryan introduces a formalism for emergence which while not immediately intuitive, has great explanatory power, and therefore should be taken seriously. One fallout of his formalism is the ability to separate the “interesting stuff”, what he calls novel emergence (life and such), from the more mundane weak emergent results of averaging (e.g. temperature) and scaling (e.g. growth). In the arrows analogy above, perhaps there is a nascent formalism with additional explanatory power. One can imagine a vector analysis and field theory to talk about emergence, including agency. Perhaps we will find that emergent structures like ice and other ordinary “matter” don’t intrigue us as much as “complex systems” because the arrows tend to go sideways in the simpler systems. The interesting stuff happens where agents can interact and affect each other’s very existence.
Natural selection is interesting because the arrows are non-neutral, they point mildly downward. Novel emergence (which includes both the cooperative and autocatalytic types) points north of the horizon, and is also very interesting. Without it there would be nothing interesting for natural selection to shape. But this is stating the obvious. We already knew that chickens and eggs are both very interesting.
* See the arrow labeled “Self Referential” in this diagram.