Thesis #19: Complexity ensures collapse.
by Jason GodeskyPredicting the proximate cause of collapse is impossible, though, as we have seen, both environmental problems and peak oil present serious threats–precisely the kind of threat that has toppled civilizations in the past. On their own, however, such proximate causes are probabilistic. Peak oil may mean the end of civilization; or, perhaps we will be able to transition to some alternative. Environmental problems may destroy the most basic necessities of civilized life, or perhaps we will solve them, instead. What makes collapse a certainty, rather than a probability, is, ironically, the very thing that defines civilization in the first place: complexity.
Graph theory is ultimately the mathematics of relationships. Here, a graph means a set of nodes and the edges (lines) that connect those nodes to one another. Such a graph can represent nearly anything. A graph of air travel has nodes of airports, and edges of routes. A graph of the internet has nodes of webpages, and edges of hyperlinks. A graph of the electrical grid has nodes of power stations, and edges of power lines. A graph of social power has nodes of people, and edges of power relationships. Graphs can be directed, where edges are all one-way, or bidirectional.
Take, as an example, our power grid. It is, as mentioned above, a graph. We can define the nodes as the power stations, and the recipients who need power. The edges, then, are power lines. There are obviously a great many nodes here, and a great many edges. But buildng a new edge is expensive, and redundancy is only useful when something goes wrong. By taking the resources that might be used to create more redundancy and instead creating new edges, any power company can increase the number of nodes connected by its edges for the same cost. The problem, though, is that the resulting graph is complex, and fragile. Removing a single edge can disconnect a huge sub-graph from the rest of the graph–and in the case of a power grid, that can mean an enormous blackout.
That’s what happened on 14 August 2003, when insufficent tree trimming in a Columbus, OH caused a single power line’s capacity to wear. That caused a power surge throughout the power grid, and the largest electrical blackout in North American history. One-seventh of the United States’ population, and one-third of Canada’s, went without power. The economic toll was estimated at $6 billion. All for an untrimmed tree in Ohio.
Increasing complexity without increasing redundancy means an escalating probability of disaster for the whole network. We can look at the power grid as such a graph, or we can broaden our scope and see all of civilization as such a graph. We now see a global economy, with currencies pegged to the American dollar or the Euro, and interdependent stock markets. Hospitals and security rely on power grids that themselves rely on a complex network of commodities and components some of which, while crucial, yet have no redundancy. Were any natural or political disaster to befall Taiwan, for example, the “Information Age” would come grinding to a halt, with 80%of the world’s mainboards and graphics chips, 70% of the notebooks, and 65% of the microchips suddenly disappearing from the table.
The solution to such vulnerabilities, of course, is simple: create redundancy. That is the only solution to such a conundrum, but it is a solution civilization is incapable of implementing.
As we have already seen, civilization must always grow (thesis #12 and thesis #13). That kind of competition creates an environment where building redundancy is impossible. An entity that spends its resources building in redundancy to guard against possible future vulnerabilities is not using those resources to grow. A competitor that chooses to grow is more vulnerable, but has significant short-term advantages that will allow it to out-compete its more forward-thinking competitor, and makes all her planning for the future a moot point. Running two power stations, or twice as many power lines, makes a power grid more robust, but it also makes it more expensive to maintain. Another grid with less redundancy costs much less to maintain, and so will out-compete the other–at least, until something goes wrong. On a long enough timeline, something always goes wrong.
While some part of the globe remains unincorporated into that graph, there is room to grow. However, once that room is consumed, room for growth can only be bought at another entity’s expense–meaning that the overall graph is incapable of any further growth. In the case of civilization, that means that the process of collapse begins. As Jeff Vail writes in “Rhizome, Communication, and Our ‘One-Time Shot’“:
In the past, such peer-polity resource races led to periodic regional collapse. Today such a collapse is not possible—with the ‘Closing of the Map’ it is no longer possible for one region of the world to collapse while progress, technology, and “civilization�? are maintained in another location, much like epidemic diseases. Instead, our global civilization simply swallows up non-performers or attempts at regional collapse and immediately reintegrates them into the global system. … In today’s world, without the ability for regional collapse and reconstitution, the entire world functions as an integrated system. We have had a remarkable run of development, fueled by the twin processes of improving energy subsidy (coal, nuclear, oil, petroleum based fertilizer, etc.) and globalization (always newer and cheaper labor pools, newer and cheaper resource sources). But this will soon come to an end. The fundamental reality of the finite nature of resources upon which we depend (fossil fuels, uranium, metals), combined with the accelerating depletion of renewable resources which without regional collapse can no longer recover (forests, topsoil, clean water) is leading down the road to an inevitable global collapse.
Thus, we find ourselves hemmed in by the very complexity that has so often solved our problems in the past. The diminishing returns of complexity make it increasingly difficult to use complexity to solve our future problems, even as our complex society finishes its 10,000 year march to complete domination of the earth, and we find that the result is more fragile than anything we could have foreseen, and disastrous because it has finally succeeded in eliminating all those alternatives it had once relied on when it had previously failed. The result is a fine, gossamer web of a culture that is doomed to fall apart in the slightest breeze–wherever that breeze may come from.
Watts, Duncan J. “A simple model of global cascades on random networks,” PNAS, April 30, 2002, vol. 99, no. 9, pp. 5766-5771.
I should’ve fit that in somewhere….
Comment by Jason Godesky — 13 December 2005 @ 1:19 PM
You can’t edit your post after publishing?
Comment by Somebody — 13 December 2005 @ 3:51 PM
I could, but it isn’t worth it here. These theses are the rough draft of a book. As I wrote here:
So you’ll also see me making notes to myself in the thesis comments from time to time, as above.
Comment by Jason Godesky — 13 December 2005 @ 3:58 PM
The last part of the article is saying that collapse will only come because some finite resource will be consumed (room for growth, minerals, biosphere whatever you want to call it). Doesn’t this contradict your thesis that complexity ensures collapse? After all it seems that as long as there’s room to grow then growth can continue.
Wouldn’t it be a coincidence that the complexity required for collapse is reached at the same time as no room is left? Or is it because complexity does not matter?
This thesis seems redundant and confusing.
Comment by DigitalDjigit — 13 December 2005 @ 4:57 PM
Well, you can certainly collapse due to lack of resources, but even in a world of infinite resources, increasing complexity will cause collapse. First, by its increasing cost, as discussed previously. This thesis caps the previous two with a reminder of that, and adds the point that increasing complexity without also increasing redundancy makes a network fragile, such that the probability of its breakdown approaches 1.
Comment by Jason Godesky — 13 December 2005 @ 5:16 PM
Ok, fine. But then why include the last 3 paragraphs?
Comment by DigitalDjigit — 13 December 2005 @ 5:26 PM
That’s why we’re incapable of redundancy. When the map is closed, the only room left to grow is by cannibalizing other sub-graphs in the system.
Comment by Jason Godesky — 13 December 2005 @ 5:37 PM
I don’t see how it follows that “we are incapable of redundancy” from “the only room left to grow is by cannibalizing other sub-graphs”.
Comment by DigitalDjigit — 13 December 2005 @ 5:58 PM
Because redundancy is room to grow. Resources being used to make a second edge between two nodes are resources that can be consumed for growth.
Comment by Jason Godesky — 13 December 2005 @ 6:31 PM
1. It seems that you are not really talking about complexity but something else. It seems more about redundancy and competition and global concentration of specializations.
2. I have thought about this when I hear how many farms are foreclosing and being taken over by agribusiness and such. 1 farm is lost every 30 minutes in the US I heard. And I think about this with the dynamics of how farmers can’t compete with cheap food imports from industrialized countries, so they have to go to the sweatshops. I also read an article about how China is going to have to lose more farmers. And so, they will be dependent on the complex network too.
3. That’s crazy about the untrimmed tree. It makes you think how easily things could domino as a result of things collapsing.
4. I don’t think that globalization was necessarily inevitable. But because it is in the interest of the elite, it is. But in the long run, it isn’t. I don’t think they are totally stupid and don’t see what would happen, but yet they did it anyway when it was necessary to increase their profits. The Fortune 500 have this interest who have a vast majority of the money in the economy, but I don’t see how it helped the economy as a whole. It wouldn’t be necessary if they didn’t lean on the government to do it.
4. I know people who are against economic growth, yet they think they can have a decent standard of living by America’s standards. I read some of another thesis about how the Great Depression was because we were growing 75% of our normal rate (I think, I forget). I don’t know.
Comment by planetwarming — 13 December 2005 @ 7:11 PM
Were any natural or political disaster to befall Taiwan, for example, the “Information Age” would come grinding to a halt, with 80%of the world’s mainboards and graphics chips, 70% of the notebooks, and 65% of the microchips suddenly disappearing from the table.
Uh, no. All this stuff is made in China now, except for certain very specialized stuff.
Michael
Comment by Michael Turton — 13 December 2005 @ 8:26 PM
Exactly. Specialized stuff–the critical component that makes everything else work. I was looking for a specific news story from a long time back, but couldn’t find it. A specific kind of plastic coating without which a microchip is impossible to build–and only two factories in the world make it. One of them had a fire and was shut down for six months. If anything had happened to the other in that time, the entire IT sector would have shut down.
If anybody remembers the incident I’m talking about and had a link, it’d be appreciated.
Planet…
Redundancy, specialization and competition is why complexity leads to fragility. Globalization is inevitable in any system where you have a collection of entities that must grow as much as possible or die. One of those entities will out-compete the others, and eventually take over the whole world. It has no other choice: to do otherwise would be to stop growing, and to die.
The “steady state” economy people are ignoring a lot of important details, in my opinion.
Comment by Jason Godesky — 13 December 2005 @ 8:53 PM
It seems like you’re getting growth and complexity mixed up.
Growing too big is the same as running out of resources. If there’s still lots of resources, you still have room to grow. If you’ve grown too big, that just means you’ve outgrown your resources and now must cannibalize.
But what does this have to do with complexity?
Comment by Anonymous — 13 December 2005 @ 9:00 PM
Also, complexity is redundancy. A web has many redundancies. A bunch of nodes in series has no redundancy. Is a web not more complex than a series of nodes?
Redundant and complex:
o—o
| X |
o—o
Nonredundant and simple:
o–o–o–o
Comment by Anonymous — 13 December 2005 @ 9:07 PM
No, I’m not getting them mixed up at all. You grow a graph by making it more complex. The first graph you offer is not complex, but strongly connected. Your second graph is weakly connected. The problem with complexity is that V/E approaches 1 (where V is the number of vertices, and E is the number of edges), because complexity grows by making a weakly connected graph. A simple series is easier to follow. A complex, but weakly connected, graph, on the other hand, can be broken easily, but the vulnerabilities may not be immediately obvious–such as a single power line in Ohio.
See if you can find the vulnerabilities in this graph.
Comment by Jason Godesky — 13 December 2005 @ 9:26 PM
So I take it your definition of complexity in respect to graphs is “the number of nodes.” Is that right? Do you consider “growing” and “increasing in complexity” to be synonymous?
Comment by Anonymous — 13 December 2005 @ 9:36 PM
It seems to me that “complexity” is more intuitively synonymous with “connectivity” in graphs. I’ve never heard the number of nodes referred to as “complexity.” That might be why some readers can’t follow this thesis.
Also, I don’t see how complexity leads to collapse. I think what you’re really arguing is “lack of redundancy leads to eventual failure.” Who’s to say you can’t have complexity and redundancy? Although you have explained that complex empires don’t build redundancies due to competition, complex entities outside of empire do build redundancy — we do have two kidneys, after all. Look at the biosphere of Earth - I’d say it’s incredibly complex, and it bounced back after at least 5 mass extinctions.
Comment by Anonymous — 13 December 2005 @ 10:04 PM
It seems a lot of the need for growth is that the corporations gobble up more and more of the wealth like what happens in the game Monopoly, so there is more need to grow because the rest of the people left have to do that just to make a living, so they have to exploit more.
The top 20 percent owns over 80 percent of all wealth.
The top 5 percent own more than half of all wealth.
In 1998, they owned 59 percent of all wealth. Or to put it another way, the top 5 percent had more wealth than the remaining 95 percent of the population, collectively.
http://multinationalmonitor.org/mm2003/03may/may03interviewswolff.html
Comment by planetwarming — 14 December 2005 @ 12:08 AM
I heartily agree with this article, Jason, but there are a couple of things that could use improving for the book.
Obviously, people are getting bogged down in the complexity/growth issue. It didn’t occur to me when I first read it, but the comments made it apparent.
Second, the article might benefit from pointing out the shallowness of the metrics of all the mentioned types of “growth.”
Maybe it’s because I read “The Tao of Pooh,” and then went on a Taoist reading kick while I was also becoming enamored of Systems Theory in Planning School, but I tend to conflate the two philosophies. So, apologies if necessary.
Anyway, the complexity/growth issue could be solved by a discussion of how systems that have too many pieces and variables to be fully understood and manipulated accurately in all their individual pieces (one would think “complex”), can nonetheless be coherent enough to be wielded and used, but only by their own rules.
Computer software, as I’ve pointed out elsewhere, is an excellent example of this. Even seemingly simple commercial software is made with hundreds of thousands of lines of code. There has yet to be born a human brain capable of containing all this in such a way that it can reliably predict absolutely any reaction of the code to absolutely any action of the manipulator. When you try to extend the functionality of the program to do things it was never intended to do, you cause unintended and often seemingly illogical consequences. Now, you know computers are just logic machines, so how could this be?
A concrete and (probably for many of us) familiar example of this would be if the address on an envelope is just not quite in the right spot when using the envelope function in Microsoft Word. The return address is hitting the curve towards the edge of the envelope and thus not printing clearly, say.
You’ll spend hours and hours fighting the program, trying to bend it to your will. As you make one change to nudge the address over, something else will change, so you’ll have to fix that. That change will create another which you’ll have to fix, etc.
It’s this struggle of fixes of fixes of fixes of fixes that is analogous to the complexity you’re talking about in the world system.
A more salient example might be beef cows. Cows are very complex systems, with all those stomachs and stuff. As a coherent thing, they’re fairly easy to understand: they’re born, turned loose in a field with their mothers, nurse from their mothers, increasingly eat the grass at their feet until they fully wean from their mothers, then grow to their full size on just grass. Once they’ve achieved maturity, you can decide whether to breed them, breed them and use them for dairy (thus using another part of the system), or kill them for meat.
At some point, somebody realized you could get broader expanses of more tender meat by feeding them just grain. Now you have “corn-fed” beef, and it’s all anybody will eat.
Thing is, cows aren’t built to eat grain. They have several stomachs (and chew cud) precisely to process very fibrous plants which have just a little bit of grain or germ (grass gone to seed) into enough fuel to keep them moving at a fairly slow pace. Giving them pre-processed food, cut down to just the germ part causes those several tummies to be upset and even diseased. I don’t know this, but it would logically make them want to act more energetically, thus toughening the meat.
So, the cows start to get sick, and have to be restrained to keep them from building strong muscles. Here comes the complexity.
First, you start giving them antibiotics and other medicines to combat the diseases caused by feeding them corn.
These drugs also kill off bacteria that are supposed to be there to help process grass, so you have to go back and develop drugs to fight the maladies this causes.
When you start to confine them to keep them from developing strong muscle, since this isn’t tender enough, they start to develop get too fatty, so you give them steroids to develop more muscle mass.
Now, you have so many pounds of beef coming onto the market that you have to start look at ways to grow pounds more cheaply. This leads you to start packing more and more cows onto the land you have. The tighter quarters mean a rise in communicable disease….
The cow-system came with all the redundancy it ever needed, but you’ve decided you can enhance the cow-system and make it do things beyond it’s original design capabilities. Each change you add forces more changes, thus increasing complexity, such as more and more drugs, more and more enriched feed, until you get to the point where the cow is using none of it’s multi-layered natural abilites and is getting sick with Mad Cow Disease.
The original nodes would be “cow” and “meat,” with the edge being slaughtering to selling. The greater complexity comes in with the new nodes of “richer food to make more desireable meat,” then “giving the cow medicine to fight the effects of richer food,” and “too much beef on the market.”
The lack of redundancy comes from changing one thing at a time (richer feed) to achieve one narrowly-defined result at a time (produce more expansive and tender meat).
This example does stretch the edges and nodes model a bit, so maybe you should reconsider that.
Now to the shallowness of the metrics….
The post doesn’t inherently depend on this addition, I just think any discussion of the drive for efficiency should include mention of how that efficiency is measured. Now that I think about it, this certainly speaks to redundancy. Thinking even more, it speaks to the complexity/growth issue as well.
I guess my main addition to the article would be that the complexity comes in when you try to go against what a system is really for. A system itself can be highly complex, but operate and be understood in simple ways. Violating the inherent rules of that system necessitates a mushrooming sequence of fixes that creates complexity that must eventually fall of its own weight.
The shallowness comes in when it’s one metric that defines the efficiency of a complex system.
Unemployment has been hanging around five percent. Unemployment is measured by a survey of people of working age in a certain week of the month, which I believe is used to adjust the numbers of those drawing unemployment compensation. If you’re compensated for two hours or more of labor during that week, you’re not counted as unemployed. Now how far is two hours of paid work going to get you?
As long as the economy can keep employing people for two hours or more per week, unemployment’s not a problem.
If an increased number of people are paid for more than two hours of work in a week, then there’s growth.
So, these very narrow metrics can give evidence of growth, when there’s really a major systemic problem.
I’d planned on developing this point and bringing it back around to the complexity, systems, and redundancy points, but I’ve suddenly run out of energy. If somebody else can see where I’m going and develop it, feel free.
Comment by Sam — 14 December 2005 @ 3:48 AM
As I just wrote over at IshCon:
Complexity with redundancy in its links can be a successful adaptation, but complexity without redundancy is fragile. Civilization is a system that seeks to maximize complexity to the exclusion of redundancy.
The critique of complexity in civilization stems from the fact that as it grows and maximizes efficiency — in other words, reducing redundancy — it becomes increasingly fragile. Also, as civilization grows, more resources must be allocated to its systems to support it, resulting in a positive feedback loop that will result in those systems’ failure. This is due to the diminishing returns of complexity as an adaptive strategy. Civilization is an unsuccessful adaptation because it seeks to continually grow and maximize efficiency at the same time, making it prone to collapse.
Past a certain point, a theoretical point of diminishing returns, increasing complexity becomes an increasingly maladaptive strategy. Yet because of path dependence and other positive feedback loops, civilization has locked itself in to run a course of perpetually increasing complexity despite its increasingly maladaptive results. The only way out of this loop is collapse, as termed in previous theses.
—-
Hypotheticals such as “in a system of unlimited resources” are absolutely useless when examining such limited systems as civilization, which exists in the real world. Putting a system of “civilization” in a “system of unlimited resources” (name one please) is an absolutely useless idea and completely contradictory, as civilization is a very specific definition based on human interaction with a limited set of resources. You might as well say that human societies have no problems and use the hypothetical “if humans didn’t exist” as your justification. That’s dumb. Really dumb.
- Devin
Comment by Devin — 14 December 2005 @ 7:44 AM
Complexity does not preclude redundancy, obviously, but if you do not have redundancy, then complexity creates vulnerability, and worse, vulnerabilities that are very difficult to recognize or address. So, I would argue that it is still complexity that’s the problem, even though that problem could be solved with some redundancy–if we were capable of creating redundancy.
Planet: That disparity of wealth is essential to the system. Without elites, agriculture would simply not be worthwhile. But that’s a separate phenomenon from the need to continually grow, which we covered in theses #12 and #13. It’s a game of prisoner’s dilemna; if you don’t grow your complexity (which requires more energy), then you’ll fall prey to those who do. While the option exists, you must take it–or be destroyed.
Sam: Yes, this is definitely the weakest of the theses yet. Fortunately, we’ve already established the marginal returns of complexity, which is the mechanistic, inescapable reason that collapse is certain, rather than just a probability. Thanks for the suggestions, I may wind up incorporating those ideas in the book version.
Comment by Jason Godesky — 14 December 2005 @ 11:00 AM
I’m unconvinced. Perhaps complexity itself leads to collapse, in a single system, like a power grid. I don’t think any of us know for sure — even a simple, non-redundant power grid would crash once in a while, I bet. But it’s a big leap to say that society as a whole will collapse. The ohio tree is a perfect example. Yes, there was the largest blackout in history, or whatever. But there were no riots, or starvation, or anything else. After a few days it was fixed. Airplanes, I’m told, are incredibly complex, and no one person understands how all the systems work (like software, i suppose). But they generally do work, at least well enough. Sometimes they crash, but most don’t, and life goes on. Theoretically, perhaps, in a closed system like the earth, we will approach the limit of growth, and everything will crash. But there’s no way anyone here can say with certainty that we’re at that point, or will be within the next 100 years. Anyone who says they absolutely know for a fact this will happen should have their own psychic friends network. Yes, global warming, bird flu, oil peak, the derivatives market, are all looming disasters (that also have a degree of uncertainty — no one knows their scope or schedule with certainty — global warming is happening now, but we don’t know when lower manhattan will be underwater, etc.) but there are simply too many variables in each issue, to predict with certainty what will happen at any definite point within the next, i don’t know, 10 or 20 years. And all of them together, or any one of them, causing collapse and die-off, is just too far of a reach.
I think a lot of people here are hoping for collapse. In a way I am. I hate a lot about our materialistic and ridiculous society. Mostly, though, I hate my job. And I think that affects how much I want some big collapse to happen, so that I can go hiking and camping all I want. But I make a good salary, and I have a family to support, and the kids have to go to school, and we need health insurance, etc., etc. I’m interested in this site because I want to get a sense of probabilities: what is the likelihood of any specific thing happening within a specific time frame. I just find a lot o the absolute certainty that is expressed here about what will happen unreasonable. One thing we should all have, given what we’re discussing, is a little humility. None of us can know everything about everything.
Comment by coffeenow — 14 December 2005 @ 11:10 AM
Statistics is the mathematics of uncertainty. It’s how to come up with things you can be sure of, when you’re not really sure of anything. That’s why complexity–its diminishing returns, and its increasing fragility–makes collapse certain, rather than merely probable. At the same time that we’re facing our greatest crises–global warming, mass extinction, peak oil, etc.–we are at our most vulnerable, and our ability to develop new solutions is most diminished.
There’s a probability that global warming will end civilization, and a probability that mass extinction will, and a probability that peak oil will, etc. The probability of all of them happening is the product of all those probabilities–which would be very, very small. But the probability of any of them happening is the sum, minus the probability of the overlap. And that probability increases with each additional factor that you include.
We don’t need to know everything about everything; we don’t even need to know all that much to understand that a world of increasing reliance on a network that is increasingly fragile is not a trend that can continue very long before it suffers catastrophic breakdown.
Obviously, we’ve never seen the collapse of a global, industrial society before, but we’ve seen plenty of collapses, and our discussion so far has been heavily informed by those past collapses. This is how they all collapsed before us, and this is exactly what we’re doing now. Maybe we’re the ones who are so special that we can succeed where they all failed, but at that point, who’s the one in need of humility? Us, for suggesting we’re subject to the same pressures that have shaped all of history before us, or those who think that the basic laws of physics no longer apply to them, and Jesus will poof new oil reserves into existence because he’s so gosh-darned impressed by our Protestant work ethic?
I would agree with you that it’s a threat, but not a certanty, if it were simply a matter of peak oil, or global warming, or mass extinction. But we have passed the point of diminishing returns for complexity, and that means that our collapse is assured. The only question left is when and how the killing blow will be made.
Comment by Jason Godesky — 14 December 2005 @ 11:38 AM
But there are some redundancies built into the system.
For example, most people work 8 hours a day 5 days a week. If there is an emergency, they can work more hours and contain an emergency.
The Internet is a robust system with multiple redundancies built in by design. Many power plants are designed to operate on several fuels.
There are giant stores of food and fuel set aside for emergencies by governments. If food suddenly becomes harder to grow, the governments have an option of rationing the reduced supply, which they don’t have to do right now. Governments contain many redundant functions and are not at all maximally efficient. Neither are military forces. Increasing the military by draft will certainly reduce the overall complexity, and this option is available in an emergency. Since there are ways of reducing complexity gradually, and reduction of complexity is an economizing process, why wouldn’t these options be taken as needed? Your argument that a more efficient competitor will emerge to destroy the economizers no longer applies because in a global system, there are no more competitors. We have become one global system, and there is not another system competing with the global system
The civilization is not so fragile that any destabilizing input will immediately collapse it. If it were, it would have collapsed by now because there are plenty of destabilizing inputs into the system
Comment by _Gi — 14 December 2005 @ 4:29 PM
It’s a balance, heavily tilted against redundancy. Some redundancy is necessary, but there’s a pressure against it.
Though one system, civilization is made up of competing sub-systems: competing countries, competing corporations, etc. That’s where the pressure is exerted.
Collapse–not just a reduction of complexity–is economizing. It cannot be taken as needed, because any net reduction in complexity will trigger a catabolic collapse, and set off a new positive feedback loop that won’t stop until civilization is gone.
Comment by Jason Godesky — 14 December 2005 @ 4:38 PM
But if this is a balance, and some redundancy is necessary and present, won’t there be a minimum level of redundancy which will resist any pressure to further reduce it?
Your notion of competing subsystems goes against your observations about globalization. As globalization occurs, corporations tend to merge and become giant, overall competition tends to decrease. Same process is occuring in relations between states, with borders and individual differences between states becoming less relevant as time goes by.
On the Tainter’s complexity graph, there is a complexity level that generates maximum ROI. Let us call it optimal complexity.
Overshooting this level leads to increased inputs for decreased returns. The solution is ofcourse collapse that reduces complexity so that it becomes less than optimal complexity. However, if there was a process that would reduce complexity only to the optimal complexity and not any further, wouldn’t this process be preferrable?
Comment by _Gi — 14 December 2005 @ 5:07 PM
right. didn’t civilization survive at a reduced level of complexity/growth during the great depression? If there is peak oil, couldn’t we be at a steady state of depression for a generation or more, until some genius, using a computer powered by a wind farm built from scrap metal invents cold fusion, or whatever? Yes, we’d all be living like russians–freezing cold, with tremendous shortage of consumer goods, and hording toilet paper if we could find it, but we wouldn’t be out in the woods eating grubs. america is a fantastically wasteful place. If there were global financial collapse, because of peak oil, or whatever else, we’d be forced to economize to the level of most 3rd world countries today. no disposable diapers, a lot less meat in our diets, no star wars figurines. but most people would still live in houses, go to school, vote for mayor, etc. why wouldn’t there be a steady state of depression? hasn’t it happened before?
Comment by coffeenow — 14 December 2005 @ 10:14 PM
Hmm. There are still major conceptual problems here. I thought I was onto a solution, but frankly got too tired to wrap it all up.
To me, a visual concept of the issue makes sense. Coherent systems that work and do specific things (cows, software, airplanes) I see as self-coherent packages like eggs or, for some reason, large tropical nuts (just what pops into my head). The complexity resulting from trying to get those systems to do what they weren’t designed to do I visualize as a Rube Goldberg crazily over-mechanistic superstructure on them.
When you find yourself having to invent more and more complex solutions, that’s when it’s time to go back and see if maybe you’re getting too far from the original system’s intentions.
On the other hand, I do have an unusually strong tendency to work with complex issues in big chunks. Mostly, my brain sorts out the internal complexities of those chunks at the sub-verbal level. What I’m “aware of” is working with the chunks in a mostly visual, 3-D way.
There are also people who have strong tendencies to work with things on a very sequential, strictly verbal level. I think it’s those people who tend to make up the largest number of those in charge, possibly because their perspective is the most easily communicated. The result of that kind of sequential thinking is that if the latest fix improves on the last one, everything must be groovy. This is especially true in those fields that most people can’t fathom, or just can’t be bothered about.
I wanted to comment on some of the later posts, regarding collapse.
I think there’s collaps and there’s collapse. What I think Jason sees is the whole world going farther and farther into a completely fallacy-based reality. This reality must logically collapse of its own weight, thus destroying all it’s component systems. We’ve already gone so far on this path, largely through globalization, that the component systems pretty much cover the entire earth. The collateral damage one would expect from such an implosion should take care of the little left over.
There’s another line of thought, that I think I’m more a part of, that sees a more near-term collapse of the economic system, primarily as it exists in the US, and has taken over most of the rest of the world. We see this system as having grown so large and fragile that any of several things could bring it down. That’s where we end up living in a Great Depression or Third World circumstances.
On the other hand, a decent-sized collapse of this sort could possibly avert the more complete catastrophe. If we have to roll back the clock to 1750 America, maybe we won’t make the planet completely uninhabitable.
Comment by Sam — 15 December 2005 @ 2:53 PM
Interesting… but the blanket statement “complexity ensures collapse” obscures the fact that you’re talking about artificial complexity, and without redundancy.
For instance, complexity is the ecosystem’s redundancy.
It’s also worth pointing out, I think, that civilization is complex in some ways but simplified in others. No matter how complex our food distribution systems, for instance, they are still rooted in monoculture(s).
Comment by Mews — 15 December 2005 @ 9:33 PM
Good luck! It looks very promising.
Comment by Somebody — 16 December 2005 @ 3:43 AM
Complexity ensures collapse.
I agree. My mental picture on the topic is the following: you can fold a sheet of paper 6 times over, then it’s over. Try it!
Our economies, laws and social systems have been folded 6 times now. They need to be unraveled by a catastrophy before progress can resume (but from a collapsed starting point).
Comment by willy — 16 December 2005 @ 9:20 AM
Hey –
Right on Mews… if you keep reading you’ll see some discussions on the difference between ‘Complex Solutions to simple problems’ (civilization) and ‘Simple solutions leading to complex results’ (ecology etc)
Janene
Comment by Janene — 16 December 2005 @ 10:06 AM
Jason,
I think the issue, or at least part of it, is that you are using the word complexity to indicate some measurement but you have not really shown how you are measuring, and given values to that measurement. For example, it seems to me that an egalitarian society is MORE complex than a heirarchical society IF you are measuring power relationships. This is because in a strict heirarchy there is basically one power relationship per person (Think of a B-tree, each node has one parent) while in an egalitarian society there are many power relationships per person. However, if you are counting the number of specialization areas then a heirarchical society is probably more complex.
You have previously defined civilization as a society that responds to all pressure by increasing complexity. This definition has the same problem, among others.
So: What are you measuring to determine how complex a society is?
Jim
Comment by JimFive — 16 December 2005 @ 11:05 AM
Hey —
Jason has said that he is using the Anthropological Criteria for complexity — ie simply add up all of the ‘artifacts’ that a given society creates and that gives you thier complexity. So a typical hunter gatherer band may have 100-200 artifacts, whereas modern america has tens of thousands… (or more?)
Its cargo
Janene
Comment by Janene — 16 December 2005 @ 11:47 AM
Unfortunately, I spent most of this thesis on a tangential point, rather than addressing the thesis itself. Complexity ensures collapse; the arguments about fragility are another way in which that statement is true, but the strongest argument is simply the reiteration of Tainter: complexity is subject to diminishing returns, and once that point is passed, collapse is ensured.
Poorly written thesis, I agree. The inevitability of collapse was actually covered much more thoroughly in thesis #14 and #15.
Comment by Jason Godesky — 19 December 2005 @ 10:59 AM
This thesis reminds me of the Henry Hub.
Comment by sevenmmm — 26 December 2005 @ 8:33 PM
WITH ALL THE DASASTERS LAST YEAR SOMETHING IS GOING TO HAPPIN
Comment by RON — 6 January 2006 @ 3:05 AM
http://globalguerrillas.typepad.com/globalguerrillas/2006/01/the_example_of_.html
I need to follow up with more stuff like this in the re-write.
Comment by Jason Godesky — 27 January 2006 @ 11:48 AM
For a new approach to the evolution of complexity visit my website.
Comment by Jacques Richard — 29 January 2006 @ 3:53 PM
Resulting complexity:
http://www.cbc.ca/cp/business/060224/b0224122.html
Comment by Rick Larson — 25 February 2006 @ 3:22 PM
If complexity ensures collapse — and the entire thrust of life’s evolution on Earth has been towards greater and greater complexity — how did the last four to six billion years happen? How are we here?
I haven’t seen you draw a distinction between types of complexity. Certainly just adding more nodes to a network won’t improve computation speeds beyond a 10-13% increase, but that’s just complexity for it’s own sake. Evolution clearly provides us with a robust complexity, because it’s been working for one hell of a long time.
Also, like Diamond, you make a great deal out of previous civilizations collapsing, but Jason, THEY KEEP COMING BACK. And they’re more complex every single time.
I just don’t see how any of this stands up to mathematical or biologial facts.
Comment by Wombaticus Rex — 28 February 2006 @ 5:53 PM
It hasn’t. See thesis #2, or Gould’s Full House. Evolution drives towards diversity, not necessarily complexity.
The collapse of greater complexity is here referring to social complexity, not necessarily biological complexity, but there’s a certain anologue here, as well: there does appear to be an asymptote of biological complexity.
Not really. There have been about six autochthonous civilizations. Each one must keep on growing. All but one of them collapsed; the last one now fills up the whole world.
Very rarely has a civilization “come back.” More often, other civilizations around it carry on, continue expanding, and bring everything around to a new, higher level of complexity, until it collapses again. (In the case of Rome, see Byzantium and Ireland).
In this particular case, it’s endgame–because there are no more civilizations to carry on after this one falls, and the resources to rebuild are beyond any primitive society.
Comment by Jason Godesky — 28 February 2006 @ 6:02 PM
John Robb, “Big Bangs“:
Comment by Jason Godesky — 1 March 2006 @ 11:17 AM
Thanks for the brainfood, I definitely don’t agree with you (or anything Gould says), but I will keep digging and perhaps realize I’m wrong. Peace.
Comment by Wombaticus Rex — 1 March 2006 @ 5:06 PM
A similar idea was reached by Joseph Tainter in his Collapse of Complex Societies.
Comment by Anonymous — 10 September 2006 @ 9:11 PM
Note to self: escalating probability of disaster.
Anonymous: if you flip through the earlier theses, you’ll see I rely heavily on Tainter’s work, and cite him directly earlier on.
Comment by Jason Godesky — 20 September 2006 @ 10:11 AM