Holism is the idea that everything is connected.
"You can't kill time without injuring eternity"
"Do unto others as you would have them do unto you."
Holism's insights are simple enough for a 10-year old to grasp, but they are profound enough to generate novel experiences for a lifetime. With holism's insights, the viewer's focus shifts from the parts of the systems to the connections between the parts. Our shared reality is perhaps not constituted so much of masses of objects in space as with the relations between them. Things (shoes, cars, dolphins, frying pans) come and go, but the patterns created by relational connections among these various things point to enduring themes. This is true especially in adaptive systems: relations among the consituent parts come to the fore, because managing changing relations is how the system reconfigures itself to adapt to its changing environment. Contrast this relational view with that of the reductionist program, where "things" are the focus. "If we can just take this apart and find out what things are inside, we'll gain understanding."
The holistic search is not to identify the disparate components of the universe, but rather to understand how the universe has assembled itself together out of disparate components. Fortunately for us the universe likes variety, and yet all of this variety has assembled itself into a whole. It is this process of assembly, of connecting and relating things, of building the whole(s), that fascinates the holistic thinker. This "building process" incorporates disparate media, but it is not limited to any one medium. This is why the adaptive systems school takes cues from physics, biology, culture, behavior, physiology, psychology, anthropology, et al, and in turn provides novel perspectives on these fields.
A System is a set of related parts, along with the relational patterns of connection and interaction. What constitutes adaptive systems, as opposed to nonadaptive sets, is that adaptive systems do not respond uniformly to external stimulus. A pool ball will always respond the same way when struck with a pool cue; every action is always followed by an equal and opposite reaction. But with an adaptive system the response to external stimuli will vary over time. An adaptive system will reconfigure itself over time, and will respond differently according to its latest configuration. If I tell a dog "fetch", it may or may not fetch the stick, depending on if it has been trained to obey my voice commands, among other things (it may be tired, it may be ill, there may be multiple sticks on the lawn and it is confused, etc).
An adaptive system will change the distribution of relations among its constituent members in response to changing environmental conditions. Consider a ceramic bowl full of marbles: there may be agates, cat's eyes, chinas, and glass marbles. Each small hard sphere rests against its neighbors. But the neighborly relationships in this set of marbles are static, and are irrespective of outside stimulus: cold, heat, moisture, force. In an adaptive system, the parts joined together in relations will transform the nature of their relationships due to changes in the environment. Consider the relations in the family. If one of the siblings is elevated due to success in society, and becomes esteemed greatly by the parents, then the other children might naturally resent this, become jealous, and even conspire against the "Goody-Two Shoes". Relationships change in adaptive systems, while relationships don't change in the static systems. The relations in a bowl of marbles (a set) is different than the relations in a family (a system): one doesn't reconfigure its constituent relationships in response to stimuli and the other does.
Agents are sub-systems which have a degree of limited autonomy to carry out work cycles deemed useful by their "uber-system", the system directly above them. A dog, for example, is an agent: it can be told to carry out a work cycle ("Fetch!") and left to figure out the means (self-propulsion, visual/olfactory coordination, purposive behavior, lifting and carrying actions, lawn navigation, etc). A hammer, however, is not an agent. It is a tool; it is useful to carry out work (driving nails, etc), but it has no autonomous capacity to carry out work cycles on its own.
So we may say that systems are sets of similar members, grouped together, whose dynamics (relations) change over time. Sets are static groupings, and adaptive systems are dynamic groupings; they are, to some degree, autocatalytic sets. I say "to some degree" because the systems draw energy and resources from the meta-systems which enclose them; they are "open systems" and thus are not completely autonomous.
Agency is important because agents are a way to convert low power (signals) to high power (work). If I want to cut down a tree, I can get my chainsaw and bulldozer and go out into the woods in search of a tall, straight tree. Or I can give a low-power signal to an agent. I can say, "Fred, please cut 2,000 feet of oak." Fred is a logger, my "timber-cutting agent", and he is waiting for such a signal. When he hears my command, he collects his nonadaptive tools - chainsaw, bulldozer, axe, etc - and sallies forth into the forest.
Agents are the lifeblood of systems. They are subsystems which have a degree of autonomy and can carry out work cycles with low power inputs (i.e. commands). A man with several agents is much more powerful, and productive, than a man with several tools (in principle, that is -- it could be argued that an atomic bomb is more 'powerful' than several gardeners or carpenters).
An agent is a goal-oriented sub-system which prioritizes its purpose around a value system. "To my wife Suzie, without whom I would have nothing"; an author (John Horgan, here) gives a dedication, or acknowledgment, at the preface of his book which is a subjective assessment of relative worth. The agent promulgates values, which are deliberate distortions of objective reality. "X" is worth(valued) more than "Y" (to me, that is). The agent imprints bias upon the landscape, in order to carry out his purpose as he understands it at the time. Purpose, or teleology, is (presumably) filtered down from the Ur-system above.
Communication is low-power energy transfer among agents; it is also known as signals, or information. A man walking around in the forest with a bucket of red paint, who occasionally stops at a tree and makes a red "X", is nonsensical, when seen out of context. Is he crazy? Is he a lunatic escaped from the asylum? So you wonder...later you notice that another man comes by with a chainsaw and cuts down every tree marked with the red "X", and later another man comes by with a skidder and pulls the log out of the forest, where you see a pile of logs, each one marked with the red "X", waiting roadside. Eventually a large truck arrives, loads the logs and takes them to the sawmill. By now you've realized the man with the red paint is connected to, and communicating with, a "tree harvesting system". And you don't need to look further, because you know that the sawmill is connected to several stores where the cut boards will be sold.
The man walking around in the forest with red paint was conveying information via his markings. His activity is logical within the agreed-upon communication of his system. A red "X" means "cut this tree" to the next person in the system. And this system, called "logging" or "forestry", is part of a larger system, let's say "the housing market", which delivers value to its own parent system "society", and is thus rewarded. So the man with the red paint is paid money to walk around in the forest and put "X"s on some of the trees. If seen out of context, his activity is nonsensical. But if seen as communication within a system, he is being a rational, purposive agent.
Lest the reader thinks that this last example is too prosaic, consider the process called "opsonization". Within the cellular matrix, an antibody (IgG, IgA) wanders around in search of pathogens. When it finds one it marks it with an antigen. Along comes a phagocyte, sees the antigen marker, says to itself, "Hmm, food", and eats the pathogen. Consider the antibody as if it were wandering around with a jar of ketchup. Whenever it sees something it doesn't like, it tags it by placing a splotch of ketchup on it and the phagocyte ("big eater") comes along and ingests the invader, ketchup and all. An antibody walking around and painting invaders with ketchup doesn't make any sense unless you see it as part of a larger system. The antibody is communicating with the larger system: what is a friendly, and what is an invader. Low-power communication is what makes high-power systems work efficiently. Without communication, there is no control, and systems are frustrated in their ability to carry out work cycles, and therefore efficiently dissipate thermodynamic energy.
Low-power energy transfers (signals) close the energy dissipation 'work cycle' loops more rapidly than high-power energy expenditures (work) can do alone. Here is another example. A fox slinks about, looking for food. When he spies a rabbit munching in the grass, he begins to creep closer. If the rabbit sees the fox coming, it will stand on its hind legs, observing the fox. The fox now realizes that its been discovered, and it will turn away from the hunt. The rabbit could run, but that would entail wasteful energy expenditure. So it simply signals the fox. The fox gets the "I see you" signal, and turns away, because it also doesn't want to expend energy on a futile chase. So both animals come out ahead, by the use of a signal. The rabbit's work loop (stay alive) has been completed with minimum energy expended, and the fox's work loop (find food) has been terminated unsuccessfully, but with less energy used than if it had included a fruitless chase. Signalling in the biotic world is ubiquitous; it is not merely confined to what are viewed as "symbiotic", or partner, relationships.
Game Theory is the term for individual agents deciding how to maximize their responses to the environment. Each agent wants to exploit the environment to the fullest. But the trick is, you can't overexploit the environment, or you fail. You have a goose laying golden eggs, and you want to get every egg possible, but you don't want to kill the goose in so doing. We are playing a card game, and we will be dealt cards, and we have a pot of money. Our pot is the capacity to play. Our cards are the changing environment. If we "play" the cards right, we increase the pot. If we overbid our hand, we reduce our capacity to play until we are removed from the table, and will no longer receive any cards.
Each agent wants to maximize return. Suppose the game is "21". You want to get as many points as possibe, up to the arbitrarily agreed-upon number 21. If you get a 22 or more, you lose the hand, your pot is diminished, and your capacity to "play" is thereby reduced. If you lose enough hands, your pot will be gone and you'll no longer be able to play. But, if you take too conservative a strategy, fearful of overbidding, you may stop at the number 14, or maybe 15. But then the house gets a score of 18, and you lose. So you realize you need to maximize return on your hand, without "overmax", without going over.
The reason each agent has to maximize its hand is because the environment includes other maximizing agents who will do a better job of exploiting the environment than you, and take your pot and push you off the table. So a score of 12 is probably going to be unsatisfactory. Each agent becomes a "probability machine", examining its environmental cues and creating heuristic platforms to decide whether to accept another card or not. So my definition of "game theory" is "agent heuristics in multivariate settings".
Consider a young lady, who is eyeing 2 young gentlemen as possible suitors. One is pleasant and bland, and fabulously wealthy. Another is pleasant and extremely handsome, but impoverished. She would ideally prefer a young man to arrive who has both looks and money (the "handsome prince"), but at the moment she seems to have 2 men who partly fulfill her expectations. So she considers her environment. If she is very young, she may decide to wait, and hope for a better one to come along. If she is older, or available suitors are scarce, she may just make do with what is available. If she has money, or if she lives in an environment where job opportunities are plentiful, she may take the handsome fellow because they both can earn money later. But if the bulk of the available resources belong to the bland young gentleman, she may decide that he is more suitable. And if she is not deemed ravishingly beautiful by social consensus, she may simply grab the first one whose eyes linger too long in her direction, figuring "that's the one".
Now, all the young ladies are going through this same heuristic process. Do I choose the man with the sports car or the man with the big biceps? Ideally I will find a man with both, but time is short and other ladies are out there grabbing available men so I may have to satisfice; I may have to make do with the best that the environment gives me. If I hold out for the perfect match, I may end up old and alone.
Lest any reader be insulted by my generalizations, remember that this is a stochastic process. These rules apply to most of the agents, most of the time. Every group has its exceptions, and its outliers. But enough of the bulk fall under this process to afford us some generalization.
Heuristics: Each agent has a decisionmaking matrix to help it navigate its environment. Most heuristics, even with people, are quite automatic. But some are consciously developed, like with a sports team, or a politician running for office, or a company. They realize they need a set of simple rules to navigate a complicated, hazardous, and shifting landscape.
What is needed is a simple method for turning data into information, information into knowledge, and knowledge into wisdom. The right set of rules, and suddenly you are surfing the big waves. Life is good. Seeing the wave, lining up, moving, getting up on the board, those are the hard parts. Once your heuristic has gotten you aloft, you simply surf the second law of thermodynamics.
Last modified 11/18/11