Chapter 23 eternal novelty

eternal novelty Holland begins by pointing out that economics is a prime example of the "complex adaptive systems" that the Santa Fe Institute is committed to studying.In nature, such systems include the human brain, the immune system, ecosystems, cells, developing embryos, and ant colonies, among others.In human society, such systems include cultural and social institutions, such as political parties and scientific societies.In fact, these systems become ubiquitous once you learn how to recognize them.But no matter where you find these systems, they seem to have something crucial in common.

First, each such system is a network of many "actors" acting in parallel.In the human brain, the agent is a nerve cell; in an ecosystem, the agent is a species; in a cell, organelles such as the nucleus and mitochondria; in an embryo, the agent is a cell, and so on.In an economy, actors may be individuals or households.Or, if you look at the business world, the actors would be companies.If you look at international trade, the actors are entire countries.But no matter how you define it, each actor finds itself in a system environment formed by the interaction of itself and other actors.Each actor is constantly taking and changing actions based on the movements of the other actors.Because of this, basically nothing is fixed in this system environment.

Taking it a step further, the control of a complex adaptive system is fairly diffuse, Holland said.For example, there is not a single main neuron in the human brain, nor is there a main cell in a developing embryo.The continuous and consistent behavior results produced by this system are generated from the mutual competition and cooperation among the actors.This is the case even in the economic sphere.Ask any president who has struggled with a protracted recession: no matter how Washington adjusts bank rates, tax policy and the supply of money, the overall effect of the economy is still the result of countless day-to-day economic decisions by millions of individuals .

Second, every complex adaptive system has a multi-level organization, and the actors at each level act as building blocks for higher-level actors: for example, a group of proteins, fluids, and amino acids will form A cell, a group of cells will form a physiological tissue, a group of physiological tissues will form an organ, the combination of organs will form a complete organism, and a group of different organisms will form an ecological environment.In the human brain, one group of neurons forms the language control center, another group of neurons forms the motor cortex, and a third group of neurons forms the visual cortex.In exactly the same way a group of workers forms a sector, and many sectors form higher-level sectors, which in turn form corporations, branches of the economy, national economies, and finally the global economy.

And Holland thinks it's important that complex adaptive systems learn from experience, often improving and rearranging their building blocks.The next generation of organisms will improve and rearrange their physiological organization in the process of evolution; people will continue to learn in contact with the world, and the human brain will continue to strengthen or weaken the countless interconnections between neurons; a company will promote Effective individuals rearrange organizational plans for greater efficiency; nations enter into new trade contracts or realign into entirely new alliances.

On some deep and fundamental level, all these processes of learning, evolving, and adapting are the same.One of the most fundamental adaptation mechanisms in any system is the improvement and reorganization of one's own building blocks. Third, all complex adaptive systems anticipate the future.Clearly, this should come as no surprise to economists.For example, anticipation of a protracted recession can cause individuals to abandon plans to buy a new car or take a lavish vacation, which in turn deepens and prolongs the recession.Likewise, anticipation of oil shortages can lead to waves of panic buying and overselling in the oil market—whether or not shortages come and go.

But in fact, this ability and awareness of anticipation and prediction is not unique to humans.Predictive codes are embedded in the genes of everything from tiny bacteria to living things. "In one environment or another, organisms with such a genetic blueprint are well adapted." Likewise, all living organisms with brains have hidden codes of innumerable predictions in their empirical inventory:" In case of ABC, action XYZ may be taken." Holland said that, more generally, every complex adaptive system is constantly making various expectations, which are based on its own internal assumptions about the external world. clear and vague perceptions.Moreover, these inner hypothetical models are far from being passive genetic blueprints.They are active, like subroutines in a computer program, that can be activated under specific circumstances, enter a running state, and produce behavioral effects in the system.In fact, you can think of internal hypothetical models as the building blocks of behavior.They, like all other building blocks, can be tested, refined, and rearranged as the system learns from experience.

Finally, complex adaptive systems will always have many niches, each of which can be exploited by an agent able to adapt itself to develop within them.That is why the economy can accommodate computer programmers, plumbers, steel mills, and pet stores as the rain forest can accommodate sloths and butterflies.Moreover, each agent that populates a niche opens up yet more niches, which opens up even more niches for new parasites, new predators, new prey, and new symbionts. living space.And this in turn means that it is pointless to discuss the equilibrium of a complex adaptive system: such a system can never reach a state of equilibrium, it is always unfolding and changing.In fact, if the system does reach an equilibrium state, a steady state, it becomes a dead system.In the same sense, says Holland, it is simply impossible to imagine that the actors in such a system will always "maximize" their own fitness, or utility.Because the space of possibility is really too big, the agent cannot find the channel of reality close to the maximization.The most they can do is change and improve themselves based on the behavior of other actors.In short, complex adaptive systems are characterized by perpetual novelty.

Various actors, building blocks, built-in hypothetical models, and perpetual novelty—all taken together, not surprisingly, make complex adaptive systems very difficult to analyze using conventional theoretical mechanisms.Most conventional techniques like computation or linear analysis are well suited for describing invariant particles in invariant environments, but really deep understanding of economies, or complex adaptive systems in general, requires mathematical And computer simulation techniques that can be used to emphasize the intricate web of interrelated interrelated underlying hypothetical models, new building blocks, and multiple actors.

Arthur was making quick notes while Holland was talking.Arthur's notes picked up speed as Holland began to describe the various computer technologies he had developed over the past thirty years to make these ideas more accurate and useful. "It's incredible," he says, "I sat there all afternoon with my mouth open." And not just because the perpetual novelty Holland points to is exactly what his Economics of Increasing Returns has been thinking about for the past eight years. The meaning of elaboration is not only because the niche that Holland pointed out happened to be the problem that he and Kaufman discussed when talking about the autocatalytic group two weeks ago, but the integrity and clarity of Holland's entire view of things And impartiality makes you slap yourself on the forehead and say, "Of course! Why didn't I think of that?" Holland's thoughts shook and identified with him, which in turn sparked more ideas in his mind.

Arthur said: "Every sentence of Holland is an answer to all the questions I have been asking myself over the years: What is adaptation? What is emergence? And a lot of questions that I didn't recognize myself. ’ Arthur didn’t know how to apply all this to economics.In fact, as he patrols the conference room, he can see that many economists are either skeptical or confused. (At least one is taking an afternoon break.) "But I believe that Holland's research is much, much more sophisticated than our work." He even felt that Holland's point of view was extremely important. The Santa Fe Institute certainly thinks so.However novel Holland's ideas might have been to Arthur and the other economists at the Economic Symposium, Holland himself was already a familiar and very influential figure among the regular members of the Santa Fe Institute. His first contact with the Institute was in 1985 at a seminar entitled "Evolution, Play and Learning".The workshop in Los Alamos was organized by Farmer and Packard. (It was at this symposium that Farmer, Packard, and Kaufman gave their first presentation on the computer-simulated autocatalysis group.) Holland's talk, on emerging research, seemed to be a great success.But Holland remembered a man in the audience who kept asking him very pointed questions.The white-haired man, with a focused, cynical expression on his face, stares at him through black-rimmed glasses. "My answer was pretty blunt," Holland said. "I don't know who he is. If I knew who he was, I'd probably be scared to death." Regardless of whether Holland's answer was polite or not, Mari Gell-Mann obviously liked Holland's answer very much.Shortly after this, Gell-Mann called Holland and invited him to join the Santa Fe Institute Advisory Board, which had just been formed at the time. Holland agreed. "I really fell in love with the place as soon as I got there," he said. "My immediate response to the issues people are talking about and researching here is 'Of course I hope these guys like me too, because I belong here!'" It's a shared feeling.When Gell-Mann refers to Hollander, he uses the word "brilliant"—not a word he uses casually to compliment those around him, and Gell-Mann doesn't often stare in amazement at anything. eyes.In the early days, Gell-Mann and the other founders of the Cowan Institute had almost always been thinking about the new redundant science in terms of concepts from physics they were familiar with, such as problems like emergence, collective behavior, and self-organization.Moreover, it seems that simply applying these metaphors to the study of the same ideas, such as emergence, collective behavior, and spontaneous organization, to the study of fields such as economics and biology seems to have created a rich research program. up.But Holland showed up, with his analysis of fitness, not to mention his computer simulations.Gell-Mann and others suddenly realized that there was a big omission in their research program: What exactly were these emergent structures doing?How do they respond and adapt to their environment? In the following months, they have been discussing that the research topic of this institute should not be just complex systems, but complex adaptive systems.Holland's personal research project—understanding the intertwined processes of emergence and adaptation—essentially became that of the entire Institute. In August 1986, Holland played the leading role at the Complex Adaptive Systems Symposium, a large meeting at the Institute chaired by Jack Cowin and Stanford University biologist Mark Feldman Introduced into the Santa Fe seminar).David Paines also arranged to take Holland to talk with John Reid and other members of Citibank on the same day as the complex adaptive systems workshop.Under the arrangement of Anderson, Holland participated in this large-scale economic seminar in September 1987. Holland participated in this series of academic activities very happily.He has been working in obscurity on the concept of adaptation for twenty-five years, and it is only now being discovered at the age of fifty-seven. "To be able to talk one-on-one with people like Gell-Mann and Anderson and be on an equal footing with them is so good, it's unbelievable!" head of a science library), he will spend much longer in New Mexico than he does now. But Holland was always an optimist.He had spent his life doing what he really liked, and was always amazed by his luck, so he had the frankness and good-nature of a really happy man.It's almost impossible not to like Holland. Arthur, for example, didn't even think about resisting Holland's attraction to him.On the afternoon of the first day, after Holland finished his report, Arthur couldn't wait to go up and introduce himself.In later sessions, the two quickly became good friends.Holland found Arthur a pleasant man. "It's rare for someone to embrace the concept of fitness so quickly and then integrate it into their own thinking so quickly," Holland said. Just go deep in." At the same time it seemed to Arthur that Holland was clearly the most complex and fascinating intellectual he had met in Santa Fe.Indeed, Holland was one of the main reasons why he remained in a sleepless state for the remainder of the economics seminar.He and Holland spent many evenings sitting at the kitchen table in the house they shared, drinking beer and discussing various issues well into the night. He remembered one conversation in particular.Holland came to this economics seminar because he was eager to know what are the key issues in economics. (Holland says: "If you want to do interdisciplinary research, into other people's disciplines, the very least you should do is take their problems very seriously. They've spent a lot of time working on them." problem.") That night, as the two of them sat at the kitchen table, Holland asked Arthur, bluntly, "What is the real problem in economics, Brian?" Arthur replied without thinking: "It's like playing chess!" International chess?Holland was puzzled. Well, Arthur took a sip of his beer, trying to find the right words.He himself didn't quite know what he was trying to explain.Economists talk all the time about systems that are simple and closed, where they can quickly figure out one, two, or three ways of behaving, and then nothing else happens.They always tacitly assume that economic agents are always extremely intelligent, and can always make the right and correct best choice immediately in any situation.But think about what that means when playing chess.In the mathematical laws of game games, there is a theorem that tells you that any finite, two-player, zero-ending game, such as chess, has an optimal solution, that is, there is a way to choose a move Being able to allow both black and white players to make better moves than they would have otherwise made. Of course, in reality, no one knows this solution, and no one knows how to find it.But these idealized economic agents of which economists speak find the solution immediately.When chess starts and two armies play against each other, the two players can imagine all the possibilities in their minds, and can reverse all the possible moves that can defeat the opponent.They are able to reverse moves over and over again, counting all the possibilities, and then find the best move to start the layout with.This way, there is no need to actually play chess.No matter which chess player has the theoretical advantage, let's say the chess player with white pieces, he knows that he will always win anyway, and he can declare victory immediately.And the other player knows that he will always lose anyway, so he can declare defeat immediately. "Who plays chess like that?" Arthur asked Holland. Holland laughed, fully understanding how absurd this was.In the 1940s, when computers were just being invented and computer researchers were just beginning to design "intelligent" programs capable of playing chess, Claude Shannon at Bell Labs, the father of modern information theory, estimated that The total number of chess moves.The answer he came up with was 10 to the 120th power, which is an incomparably large number.The time since the big bang to the present is measured in microseconds, and there are not so many microseconds.There are not so many elementary particles in the universe visible to our naked eyes.No computer of any kind can count all these moves, and of course it's even less likely that a human brain can.Human chess players can only judge which strategy is the best under what circumstances based on actual experience. Even the greatest chess masters have to constantly explore their chess moves, as if falling into a bottomless black hole. Find your way by a faint lantern.Of course, their chess moves will continue to improve.Holland, a chess player himself, knew that a master chess player of the twenties would never be able to beat a contemporary chess master like Gary Kasparov.But even so, they seem to have only advanced a few yards in this unknown world.This is why Holland called chess fundamentally an "open" system: its possibilities are virtually endless. That's right, said Arthur. "The types of actions that people can actually predict and take are very limited compared to so-called 'optimization', where you have to assume that economic agents are much smarter than economists." However, "the assumption of optimization That's how we deal with economic problems. Trade with Japan is at least as complicated as chess, and yet economists are still there saying: 'Assume this is a rational game.'" So, he told Holland, this is where the real problem with economics lies.How should we build this science in the face of agents who are not perfect, but are very intelligent and constantly exploring endless possibilities? "Aha!" said Holland, as he always did when he figured something out.International chess!Now he understood the metaphor.