Visit Your Local PBS Station PBS Home PBS Home Programs A-Z TV Schedules Watch Video Donate Shop PBS Search PBS
NOVA ScienceNOW

Emergence: Expert Q&A

  • Posted 07.16.07
  • NOVA scienceNOW

On July 16, 2007, John Holland, one of the world’s leading authorities on emergence, answered questions about this mysterious phenomenon.

John Holland

John Holland

One of the world's leading experts on emergence, John Holland is a professor of Psychology as well as of Computer Science and Engineering at the University of Michigan. Full Bio

Photo credit: Courtesy John Holland

John Holland

One of the world's leading experts on emergence, John Holland is a professor of Psychology as well as of Computer Science and Engineering at the University of Michigan. He is also an external professor and member of the Board of Trustees at the Santa Fe Institute, a MacArthur Fellow, and a Fellow of the World Economic Forum. His two most recent books are Emergence: From Chaos to Order (Oxford University Press, 2000) and Hidden Order: How Adaptation Builds Complexity (Helix Books, 1996). Forty-nine students have received Ph.D.'s under his guidance.

Q: Has anyone developed a widely accepted definition of emergence? It seems like a phenomenon that is "obvious when we see it," but is there a way to define it when we don't always know what will "emerge" from simple individual instructions? Anonymous

Q: The concept of emergence is a self-evident property of matter/energy derived empirically from naïve observation. It is the fundamental principle of Gestalt psychology. However, it falls short of science, as strictly defined, since from it no hypotheses can be formulated that can be tested that could support theories that would yield predictions; in other words, there is no way to understand how it works. In this sense, it is more bizarre than quantum mechanics or string theory. As an explanatory device it is more akin to intelligent design than to evolution.

Isn't emergence merely a statement of the obvious? A solution in search of a problem? In any real way, how does the study of emergence advance knowledge? Michael, Norwich, Connecticut

John Holland: Let's consider emergence in a fully defined arena, like checkers or chess. The rules of the game, though few in number, give rise to a huge number of possibilities, most of them irrelevant or outright bad if the objective is to win the game. However, there are certain recurring patterns (regularities) among these possibilities that greatly influence the possibility of winning. The regularities even have names, like "sacrifice," "pin," and "gambit." These regularities depend upon particular interactions between the pieces. The regularities obey "rules" at a higher level. These new rules (strategies) for exploiting the regularities cannot contradict the rules of the game, but they are emergent in the sense that they are not obtainable by simply "summing up" the rules applying to individual pieces.

Similarly, the persistence of a vortex in a stream, and its interaction with other vortices, is not simply a summing up of the properties of individual water molecules. It is noteworthy that Maxwell used vortices as a rough analogy to guide him to his theory of electromagnetism.

Emergence is a regular phenomenon in complex adaptive systems (cas). A cas consists of a large number of interacting individuals, called agents, that adapt or learn as they interact. Examples, with the corresponding agents in parentheses, are: markets (traders), ecosystems (organisms), the immune system (antibodies), biological cells (proteins), and so on. Some of our most complex current problems center on cas:

  • strengthening the immune system (antibodies)
  • preserving ecosystems (organisms)
  • understanding the evolution and acquisition of language (humans)
  • controlling spam, viruses, and other intrusions on the Internet (sites)
  • encouraging innovation in dynamic economies (firms)
  • predicting the effects of trade laws on global trade (governments)

All cas that we have examined closely exhibit a phenomenon that I'll call lever points, "inexpensive" actions that yield desired outcomes. In a game, a lever point is, say, a pin in chess; for the immune system, a vaccine is a good example. We mostly discover lever points in cas by trial and error, but a good theory would give us a principled way to find them. Pieces of that theory exist, but a general theory of cas is a deep and difficult problem. Still, it is clear that this young discipline has much to contribute to our understanding of the world.

Q: Is a hurricane part of emergence? Germaine, Noelani Elementary School, 6th grade, Honolulu, Hawaii

Holland: I would say a hurricane is emergent because (1) it depends on the interaction of many components, (2) it arises because of interactions that are not simple sums of the behaviors of the component water molecules, and (3) it is a pattern imposed on a flow. Again, a vortex in a stream of water offers a simple analogy.

Q: In a philosophy workshop we tried to link emergence to the unpredictability of nonlinear systems like weather or walking pedestrians. We utterly failed, because we couldn't determine the exact "spot" where the new quality/behavior should arise. Does it just lie in the eye of the beholder (we humans naming something we didn't expect), or is there a real new quality emerging? Andreas, Henke, Germany

Q: How do you separate the observer from an emergent pattern? In other words, if birds flock in the woods and no person sees them, do they make a pattern? Thank you. Carey Sherrill, Columbus, Ohio

Holland: Starting with a clean tabletop, add one grain of sand at a time to the center of the table. When do we reach a point when there is a pile of sand which can exhibit small avalanches of sand when new grains are added (an example of the phenomenon known as "self-organized criticality")? There is not a sharp transition from a "no pile of sand" to "a pile of sand," but the transition does exist.

Many complex phenomena, up to and including "life," have no sharply defined "boundaries," but they have been studied scientifically. In the case of life, for example, these studies have produced astounding, unexpected insights that go far beyond "eye-of-the-beholder" interpretations. For example, the program-like properties of DNA, with genes being "turned on and off," are emergent properties not simply determined by summing up the properties of the component nucleotides.

The idea that patterns do not exist without an observer is a bit like the old philosophical conundrum: Does a tree falling in the forest make a sound if no one is there to hear it? If you define sound as a pattern of vibrations in the air, the answer is "yes." If you define sound as vibrations in the cochlea, the answer is "no." We can define patterns rigorously without reference to eyes.

Q: Einstein's mass-energy equivalence. Space and time intertwined. The number pi seemingly endless. What does emergence say about the geometry of the universe, as we know it? Thank you. Randy Caba, Long Beach, Washington

Holland: The concept of emergence has played an important role in the physical sciences, though not under that name until recently. For instance, the properties of water (transparency, refraction, etc.) are not properties of individual molecules on H2O. They are properties of the aggregate, and they depend upon the interactions of the molecules.

Much of the story of science is one of using selected combinations of the elements of one level (e.g., the atoms of atomic physics) as elements of the next level (the molecules of chemistry). The laws of each level are constrained by, but not fully determined by, the laws of the previous level.

Q: Are there, or do you predict, any emergent properties in the Internet and/or other computer networks? Jim, Sweden

Q: Might a large enough collection of the right kinds of computers, programmed in an appropriate way, create an emergent behavior? Hal Lane, Durham, North Carolina

Holland: It seems there are already emergent properties of networks, both positive (e.g., new kinds of collaborative markets such as eBay) and negative (e.g., self-reproducing subroutines such as viruses, worms, etc.). There are also simulations that exhibit emergence, one of them being the BOIDS simulation shown in the Nova segment.

Q: What insights does emergence present about the stock market? Is it possible to trace back a heavy volume day to a single stock and find the "leader" that caused the trade-off? Peter Andersen, Jersey City, New Jersey

Holland: Q: As an organizational performance consultant and theorist, I have for several years since reading your books Hidden Order and Emergence been captured by the notion that emergence is the best description of how corporate culture occurs. Further, the idea that organizational culture can be best appreciated as a complex adaptive system seems appropriate to me—in my experience organizational culture always seems to occur from the inside out, or from the bottom up, if you will (nonlinear interactions of agents in a system, etc.) If this is true, then the method by which it is created or amplified seems the antithesis of what traditional management theory espouses (top-down control, directive, extrinsically motivated alignment). Does this make sense to you?

Many of my esteemed colleagues argue that people are not ants or fish or antibodies because they are "choiceful," and this renders complexity science useless as a lens for understanding organizations. My view is that organizations must first be viewed as organisms, albeit primitive ones, before a full appreciation of them as living, complex adaptive systems can be had.

Am I off base here? Dave Guerra, Holland, Texas

Holland: Q: Do you think your studies have an application in marketing? That was all I could think of while watching your wonderful piece. Nick Moreno, San Diego, California

Holland: Thinking individuals in the aggregate often act in predictable ways (e.g., crowd behavior or traffic jams, as illustrated in the NOVA scienceNOW program). The study of these complex adaptive systems usually involves agent-based models, wherein individual agents (e.g. firms, traders, or even antibodies and organisms) modify their behavior by learning or adapting their strategies on the basis of experience.

In building these models, there can be several levels, where selected aggregates (combinations) of agents at one level become agents one level up (think of traders forming trading groups which then act as agents in the larger market).

The successful Prediction Company (founded by members of the Santa Fe Institute) provides an example of market trading based on complex adaptive systems.

Q: To my layman's mind, there seem to be similarities/overlaps between 1) emergence, 2) chaos theory, and 3) Stephen Wolfram's New Kind of Science/cellular automata. Are the three ever considered variations on the same theme? Anonymous

Q: What general field of study does emergence fall under? What course of study (from a B.S. through a Ph.D.) would you recommend for someone who is interested in studying emergence? Anonymous

Q: What is the title/author of the easiest book to read on this subject? Allan Gehring, Plant City, Florida

Holland: The general field of study goes under the name "complexity," which Stephen Hawking calls "the science of the 21st century." For 20 years, the Santa Fe Institute has been the center of such studies. Its story is well-told in Mitchell Waldrop's book Complexity. In Europe, the recently founded Institute Para Limes has similar objectives. Some universities now offer graduate study in this area, e.g., the Center for the Study of Complex Systems at the University of Michigan. A visit to the Web sites of these organizations will produce much more information.

Q: Is it right to say that life is the evolution of simplicity into the emergence of complexity? Roger Landriault, Chelsea, Quebec, Canada

Q: How has your research been received by those on both sides of the issues concerning evolutionist and creationist perspectives? How have they responded to the fascinating segment on "emergence" that aired this evening on NOVA scienceNOW? Anonymous

Q: Hi. Something I have pondered for some time is emergence in plant community ecology, i.e., 1) we have assembly rules 2) we have patterns of vegetation that 3) respond predictably to certain disturbance regimes. It has been a long time since grad school, but I do not remember this being modeled.

My ultimate question is: I have a feeling that emergence is the other end of the spectrum from the Second Law of Thermodynamics—things eventually fall apart. In many cases, coherent plant and animal communities begin to disorganize (de-emerge) when disturbance regimes are removed. Has anyone looked at this from the standpoint of energy input into ecological systems?

Thanks. This has been bugging me for a long time. May be a goofy question, but what the hell. Jim Eidson, Rockwall, Texas

Holland: If we equate simplicity to a limited number of "building blocks" (atoms, nucleotides, linguistic phonemes, computer instructions) and complexity to the vast number of ways of combining those building blocks (molecules, DNA, speech, programs), then we open the possibility of deriving complexity from simplicity.

Darwinian selection is, from one point of view, the selection of persistent combinations. One way to become persistent is to reproduce. There is no violation of the Second Law, because, in a sense, the great flux of energy from the sun is "degraded" to produce organization (think of the action of chlorophyll). Networks of interaction between persistent entities, such as trading and recycling of resources, can encourage the emergence of ever-increasing diversity. The great diversity of tropical rain forests provides a good example (over 1,000 species of insects can inhabit a single tree).

My early research used computer programs called genetic algorithms to solve difficult problems by mimicking the role of recombination and selection in evolution. The execution of such a program, where nothing is hidden, can create quite complex solutions. As with Samuel's checker-playing program, which I describe in my book Emergence, the result can outdistance the designer. Such concrete illustrations of emergence give little comfort to those advocating intelligent design.

Q: I feel a similar kind of complex order that stems from simple instructions when I improvise with members from my dance company or when we perform a unison piece of choreography. Has any part of this theory been applied to artists, specifically dancers or chamber music groups? When I watched the NOVA scienceNOW segment tonight on PBS, I definitely had a resonating feeling. There seems to be an organizing chaos that we all can willingly tap into. I was just wondering if there was evidence to support that. Troy Ogilvie, New York, New York

Q: Can emergence address a pattern of memory within one's own body, as in muscle groups when performing Tai Chi movements that have been done a thousand times in the same order? Daniel Piatchek, St. Louis, Missouri

Holland: A: David Cope's book Computer Models of Musical Creativity is suggestive. He discusses the recombination of motifs. This, I think, is similar to first acquiring certain patterned, well-practiced "building blocks" that can then be combined. Much as is the case with children's building blocks, combinations of such motifs can yield fresh variations and improvisations. It is easy to establish that most inventions (e.g., the internal combustion engine) are derived from innovative combinations of well-known "building blocks."

Related Links

  • Emergence

    How does the "intelligence" of an ant colony or the stock market arise out of the simple actions of its members?

  • Everyday Examples of Emergence

    From bird flock to slime mold, chess to consciousness, see emergence in action all around us.

  • Bird Brains

    Clues to the origins of human language are turning up in the brains of birds.

  • Games With a Purpose

    Luis von Ahn explains the higher purpose of his computer games, the downside of being a MacArthur "genius," and more.