Monday, April 20, 2009

The Emergent System

The concept of emergence seems to be cropping up everywhere one turns these days. The term is used to describe not only new events in organized religion, as described in the book I reviewed last week, but many other examples of otherwise unexplained behavior in the social sciences, such as dynamics in financial markets.

Emergent properties are collective properties, which means they are properties of collections, not properties of individual parts. The cause can often be traced back to the
interactions between the parts of which a system is made.

Often the nature of those interactions is more important than the identity of the parts. Thus, flocking behaviors are observed in systems composed of insects, birds or even people, as seen in the accompanying
photo. This concept has found a use in the computer science world where swarm intelligence, a type of artificial intelligence, is used to control flocks, or swarms, of robots.

The term "emergence" has been around for awhile and can be traced back to systems theory. Systems theory has also been around for awhile, going back to the 1950s, but it was not until 1972 when Philip Anderson wrote an article in Science entitled "More is Different," that scientists really began to grapple with the fact that systems are different than the isolated parts which collectively make up the system.

In this seminal article, Anderson wrote:

"The ability to reduce everything to simple fundamental laws does not imply the ability to start from those laws and reconstruct the universe...At each level of complexity entirely new properties appear. Psychology is not applied biology, nor is biology applied chemistry. We can now see that the whole becomes not merely more, but very different from the sum of its parts."

This idea that the whole is more from the sum of its parts is one of the main reasons that the emergence concept has been so compelling. When a new quality, or new behaviors, emerge from a simpler system we scientists long for a theory or explanation.

So far, our understanding of emergent behavior is only partial. We understand that the interactions between the parts of the system is a key feature of the mechanism that leads to emergence. We also understand that the behavior of the system as a whole can
feed back on the parts which make it up, changing the behaviors of those parts.

It is this last aspect of emergence that makes it seem more than a little scary to those who first hear about it. If we are the parts and the system is society, the idea of the "system" imposing its will on us can be quite frightening.
Just what is this "system" that is imposing its behavior on us? And is it something that I, as an individual, have any choice about?

This topic is a deep and intricate one and I will have much more to say about the science of emergence in future posts - stay tuned!

16 comments:

  1. Thanks for your twitter update that led me here. I'm fascinated by this topic, now called emergence, and have been recently tantalized by seeing bits of it everywhere. I wonder how this all relates to the system we call "The Cloud." I'm off now to rifle through your blog archive. Thank you!

    ReplyDelete
  2. It seems that some congregations have realized they have to change somehow and they are trying to exchange ideas which is great

    ReplyDelete
  3. can someone explain the difference between emergent systems and chaotic systems?

    ReplyDelete
    Replies
    1. That's simply a characterization of behaviour. A chaotic system is one which exhibits 'chaotic' behaviour. See also: 'chaos theory'. It suggests that there are 'states' in a given system in which behaviour is linear (like laminar flow in air flow analyses) and 'chaotic'. The term 'emergent system' is actually a misnomer. We speak of 'emergent properties' of a system. Like the way that complex systems have a personality or purport to function differently than the purpose for which they were designed.

      Delete
  4. Hi - thanks for your comment/question. A chaotic system is an emergent system in which chaotic behavior emerges from the interactions of the system's parts. Chaos isn't the only thing that can emerge in these systems -- sometimes very orderly behavior emerges. So, the short answer is that chaotic systems are a sub-category of emergent systems.

    ReplyDelete
  5. This comment has been removed by the author.

    ReplyDelete
  6. Fun! & FYI, found my way here via
    http://bigthink.com/ideas/38557 "emergent system" link (May 24, 2011). Started at http://bigthink.com/ideas/38554 (Marriage is an Emergent System, David Brooks) Interesting, also bumped into you a while back via fractal cookie discussion. (grin) Here's to feedback loops between nodes!

    ReplyDelete
  7. I like how it all resembles fractals, nature, life and the universe is all about fractals, once that we understand this concept, we can understand the laws of all.

    ReplyDelete
  8. Nice :) I have been wondering about this issue, I wanted to thank you for this special read.

    ReplyDelete
  9. I'm less worried about control than lack of it. Emergent systems are adaptive - they can modify their behaviour in response to environmental stimuli as a whole for the benefit of the whole, regardless of the impact on an individual. This makes emergence a survival trait which is beneficial overall. However, circumstances outside the normal range of environmental stimuli can cause the entire population to self-destruct. This is of course just natural selection at work on the system as an individual, where systems that can adapt that little bit more are preferentially selected. The problem for us comes when there is just one copy of the emergent system - the potential for self-destruction due to circumstances never predicted (a 'black swan' event) is real and for a society dependent on that emergent system, potentially fatal.

    We now have emergent behaviours appearing from interconnected programmes (not computers themselves). A particularly troubling one is automated trading, where computers trade assets faster than a human could possibly monitor. There have been numerous 'flash crashes', some unexplained, in which computer trades cause wild fluctuations in a share price or, worse, an entire market. The rules that govern the market can't take into account every possible external 'black swan' event, so they have to be designed to protect the system from self-destruction whatever happens. We are a long way from this at the moment and until then, our financial systems are extremely vulnerable.

    ReplyDelete
    Replies
    1. Thanks for your comment - many excellent points in here! It seems to me you are making an excellent argument for continuing our pursuit of a detailed understanding of emergent systems. I don't think, though, that understanding them better will ever allow us to control them, except at the level of systemic design. I very much agree with your last point about the vulnerability of the financial system, which is clearly subject to these types of self-correcting events, so the only way to prevent that is in the initial design.

      On the other hand, emergent systems in the natural world (those we have not "designed" ourselves) may also be subject to these self-destructive corrections. Our response to that can be a desire to control or change the system, but who are we to say that these seemingly catastrophic changes are bad? I think it requires some trust on our part and a willingness to admit that maybe we don't understand the bigger picture here. As the saying goes, what seems like the end of the world to the caterpillar is, to the Master, a butterfly.

      Delete
  10. the root of the concept comes from cybernetics. One of it's prime proselytizers was Douglas Hofstaedter, whose essay in Godel Escher and Bach: An Eternal Golden Braid discussed how the emergent properties of an ant colony exhibited the intelligence of a single mind. And his idea was that this was how we could get computers to have consciousness eventually.

    ReplyDelete
    Replies
    1. Yes, Hofstaedter's work is very relevant to all this. Thanks for your comment, including the one that seems to have now been deleted about "emergent system" being a misnomer. I agree that the proper way to word this is to say that the system exhibits emergent properties, not that there is anything that could be called "an emergent system." However, the title of the post was in response to the usage of that term in the press, so I wanted to explain what people seemed to be trying to write about, but we're missing some of the subtleties.

      Delete
  11. This comment has been removed by the author.

    ReplyDelete
  12. Raima, You write with simplicity and clarity, which makes it so easy to understand this not-so-easy-to-understand topic. Am anxious to read more of your posts about the science of emergence.

    ReplyDelete
  13. wow its the great content ...i got the god information,...i like that thank you for sharing the blog please give me many more information...

    Financial industry regulation

    ReplyDelete