Deborah Gordon on Ants, Humans, the Division of Labor and Emergent Order
Aug 20 2007

Deborah M. Gordon, Professor of Biological Sciences at Stanford University, is an authority on ants and order that emerges without control or centralized authority. The conversation begins with what might be called the economics of ant colonies, how they manage to be organized without an organizer, the division of labor and the role of tradeoffs. The discussion then turns to the implications for human societies and the similarities and differences between human and natural orders.

Wally Thurman on Bees, Beekeeping, and Coase
Wally Thurman of North Carolina State University and PERC talks with EconTalk host Russ Roberts about the world of bees, beekeepers, and the market for pollination. Thurman describes how farmers hire beekeepers to pollinate their crops and how that market...
Yuval Harari on Sapiens
Yuval Harari of Hebrew University and author of Sapiens talks with EconTalk host Russ Roberts about the history of humanity. Topics discussed include the move from hunting and gathering to agriculture, the role of fiction in sustaining imagination, the nature...
Explore audio transcript, further reading that will help you delve deeper into this week’s episode, and vigorous conversations in the form of our comments section below.


Aug 21 2007 at 6:32pm

My bussiness partner likes to say we are like ants. Meaning bussinesses and people.

Jose Jimenez
Aug 22 2007 at 7:19am

This podcast is one of three that I listen to every week without fail, the other two being NPR’s “This American Life” hosted by Ira Glass and WNYC’s “Radiolab” hosted by Jad Abumrad and Robert Krulwich-both can be found on Itunes or through their respective websites.

By coincidence the topic of Radiolab this week was also Emergence. If you were interested in this week’s Econtalk you will probably find the Emergence podcast at Radiolab also interesting. Weblink:
Itunes Link:

Clayton Roche
Aug 22 2007 at 9:40am

I have to comment on this despite that it is now irrelevant.

I signed on for the sole purpose of linking the above show myself, which I listened to a few days ago.

Ken Willis
Aug 22 2007 at 11:53pm

Unfortunately, everyone has a “we have to do something” mentality and it is a near impossible task to convince anyone that sometimes the best solution to a perceived problem is to do nothing.

Ken Willis
Aug 23 2007 at 12:01am

An example is the claim by some that evolution could not have produced complex life forms without the help of an “intelligent designer.”

Aug 23 2007 at 4:38am

Well, here are some lessons economists might take from studying ants, etc.

1) the role of the ant follows a progression; hunter/gatherers probably progressed from working around the camp to being the forager (hunter) to being the patrol (wise sage). So, when economists argue that laying off old workers causes no problems, you expect the worker to go against nature, and start over again, after going from trainee to copilot to pilot, the laid-off pilot just starts over as trainee.

2) While ants might be born the same, once they have reached a certain point in life, the are now part of a single colony. So, you are part of the ivory tower colony and I’m from the college drop-out progressed to computer systems engineer and scientist by experience colony, so my messing in econ is objectionable to you, because I should be starting as a student of econ to grad student and then I might become part of the econ colony. But that is going against nature. Likewise, saying the laid-off airline captain pilot switches to a new career goes against nature.

3) The ants live in a harmony with the environment that has evolved by trial and error. While colonies/ants must incorporate random variation in their innate rules which provides behavioral variation, the benefit to the environment must balance the cost. Ants wouldn’t have survived hundreds of millions of years by destroying the environment by their simplistic rules. Colonies that do as man is doing will destroy their local environment and then perish as a colony, weeding out the unsustainable rules.

For example, you believe the ants walking over the pile of seeds is irrational, failing to consider how unsustainable it is. As a neoclassical economist, you seek the lowest cost solution for the moment. Thus you would have the rule require scouring the area closest to the nest first, then expanding out as that is depleted. But that isn’t sustainable. Early on, the ants finish early with little effort. When they find the pile of seeds they are really efficient. But as time goes on, they spend more and more time hunting for food, which dries them out and presumably they expire. If they start over nearby because more seeds should have blown in, some will be found, but not in the quantity initially exploited, so they still need to forage further out, which may be so costly the colony suffers a recession of depression with loss of life while conditions recover. By foraging near and far always, they maintain the harmony required to survive at a small cost in efficiency and extra labor.

4) ants differ by their environment; there is diversity in the tactics in the rules the ants use based on their environment. Economists often see that as inefficient so food becomes mono-culture, often designed to meet some efficiency objective, rather than promoting diversity and higher labor costs. And while the ants might operate by what seem inefficient rules, they still spend most of their time not laboring, but just milling around.

5) What happens to the excess ants when the recession or depression comes wasn’t discussed, but I assume the laid off workers are recycled some how. If economists were brave and honest, they would say “well, too many workers, too many people, to be supported, so its time to eliminate the too many mouths, time for Soylent Green.” Now people would really avoid economists at parties if they said that.

6) One might ask, “why not kill all the annoying ants?” In a way, they are like the people who swarm over office building after dark, emptying trash, sweeping, etc. In my experience, one thing that annoyed a lot of people was the unknown rules they followed. Stuff in a trash can was thrown out, even if not trash. Garbage marked “trash” stayed. The ants play an important role in the environment, otherwise they wouldn’t exist. Not because there is a planned role for them, but because the role they randomly came to play was useful to all the other millions of species in the environment, so they collectively support the ants existing.

The question many have is whether humans will survive with the new rules of the fossil fuel burning of the cheapest first ensuring ever increasing costs in an unsustainable quest for growth will converge on new emergent behavior that is in harmony with the environment, or will the environment snuff out the industrial ant colonies, leaving only the inefficient non-developed colonies with their older rules of life.

Donald Browne
Aug 23 2007 at 6:12am

I find the best interviews and most relevant somehow, seem to be the offbeat ones….ants with professor Gordon and the stadium scalper interview. Keep up the great podcasts, and I look forward to more of the unusual which for some reason seems more theoretical, yet very practical.

I am wondering if we humans are really different than ants…we widen highway 101 and when finished find it’s not wide enough…the ants march right over, and ignore the new hill of seeds, to travel to a smaller seed cache found the day before. Hmmm.

[Author URL corrected for typo–Econlib Ed.]

Ken Willis
Aug 23 2007 at 10:30am

I am wondering if we humans are really different than ants…we widen highway 101 and when finished find it’s not wide enough…the ants march right over, and ignore the new hill of seeds, to travel to a smaller seed cache found the day before. Hmmm.

In the case of ants, the foragers practice of always following the path alerted by the patrolers has apparently proven to be a good strategy for survival. But ants live in pretty much the same environment in which they evolved. We humans have been able to substantially change the environment from that which we evolved in a few hundred thousand years ago. Some of our behavior is anachronistic and makes less sense today than it might have then. Fortunately we have intelligence and can learn new behaviors without waiting for evolution to work it out. Ah, can’t we?

Aug 24 2007 at 9:21pm

big pile of seed = black swan?

Aug 26 2007 at 9:02pm

Fantastic! I’ve always been fascinated by ants and their societies. Of course my wife was less impressed by the emergent properties of their socieites then by the fact they are “run by” women. I know….I know…”run by” but no the discussion ended there.

Thanks again.

Robert Bennett
Sep 4 2007 at 2:46pm

Great service, econtalk! I find it very helpful.

I’ve been thinking a lot about how patterns arise out of form and rules arise out of patterns. Mutations in form (e.g. a species of ants with better eyes), may enable their beneficiaries to assume new behavior patterns which, when repeated in concert, result in new rules enabling new capacities and benefits for the entire colony depending on the given state of the environment and the mutation’s utility given the prior rule set of the colony. No matter what happens, though, in ants form, pattern and rule all determine purpose and are all both perceived and applied information.

Humans are different in that with the exception of our absolute needs, our form, patterns and rules are components affecting our purposes but are not solely determinative of our purposes. Hence our freedom to determine the value of things, which calculation assesses the utility of any given thing to affect some purpose less cost. This is why the price mechanism usually works so well while political organizations do not.

Notice how we humans have developed legal rules attempting to organize, limit, enhance or otherwise optimize our behavior around the patterns, rules and values (although I distinguish these from moral principles) prevailing in society. These legal rules are often adopted in the attempt to optimize behavior patterns without first assessing whether the legal rules conflict with common purposes (utility to value less cost) to the extent rending them costly rather than useful. Worse, economists often fail to recognize that here lies the line between an incentive and a burden, causing them to often characterize a burden that people will avoid as an incentive that will motivate the desired behavior (I should distinguish between positive and negative incentives here, but I won’t because I want to keep this short). Still worse, I’ve encountered few economists who are willing to admit that due to our constantly changing environment, incentives ofen become burdens and vice versa, rending the predictive capacity of their pronouncements unreliable at best.

Has any economist treated the problems described above?

Brad Hutchings
Sep 23 2007 at 10:27pm

I just caught up on this one while working late last night… Russ, this was a brave effort. I could see where you were going with this and the parallels you trying to tease out of her toward the end. I fear that there is a tremendous amount of marketing to do on the concept of “emergent order” so that it doesn’t end up being a concept like “social Darwinism”. Her pushback on emergent order in social systems was “yes, this occurs in nature, but we are better than that in social systems”. That seems very similar to me to how Stephen Jay Gould dealt with Darwinism.

The difference between social Darwinism and emergent order of course, is that we’re not advocating laissez faire so the best and brightest will rule us, but so that unexpected contributions can come from unexpected corners and make things a lot more interesting. The total knowledge landscape can’t really be sorted out except by letting it sort itself out. It sounds to me like the analogy probably breaks down as a result of Professor Gordon’s ability to dissect and understand all the messaging mechanisms used by a particular species she studies. Whereas, in an economy made up of autonomous people, sorting it all out to pick the winners and losers is an intractable problem. Given that we all want to be better off and that the wealth is not a zero-sum game, we trust that good ideas will come from unexpected places and know that our self-annointed best and brightest won’t have a monopoly on those ideas.

Comments are closed.


About this week's guest:

About ideas and people mentioned in this podcast:Books:


Podcasts and Blogs:



Podcast Episode Highlights
0:36Intro. Spontaneous order, complexity, emergence, self-organizing systems: the way that order emerges without control. Central to economics, parallelism in world of ants. From Ants at Work:
The basic mystery about ant colonies is that there is no management. A functioning organization with no one in charge is so unlike the way humans operate as to be virtually inconceivable. No insect issues commands to another or instructs it to do things in a certain way. No individual is aware what must be done to complete any colony task. Each ant scratches and prods its way through the tiny world of its immediate surroundings. Ants meet each other, separate, go about their business. Somehow these small events create a pattern that drives the coordinated behavior of colonies.
How did scientists come to recognize these patterns, history, what was intriguing? A bit of the Old Testament, Proverbs, first part familiar: "Look to the ant, thou sluggard. Consider her ways and be wise. Without chief or ruler or overseer she gathers her harvest in the fall to save for the winter." Understood 6000 years ago that there is nobody in charge. Mostly sterile female workers. Ant that lays the eggs is named the Queen since the 18th century, suggests authority, newer idea than the reality that nobody is in charge. The Queen is bigger, merely the ovaries of the colony, not the one who decides what needs to be done. But movies like "Ants" and "A Bug's Life" sets up whole organizations with hierarchy. Stories are reflections of how we think about society, but not about ants. Last 20 years: questions about organization about hierarchy has become more interesting with the advent of computers. Ants don't have cell phones like CEOs, no hierarchical organization.
6:0110,000 species of ants, 50 of which well-studied. Most ants can't see. Some can distinguish shapes or direction of light. They smell well with their antennae. Many different glands secreting different chemicals used for communication. Within a colony, species all have same glands. Another chemical communication, not just an alarm signal that is a chemical secreted immediately, pheremones are put outside the body, as opposed to hormones. Volatile, dissipates in the air quickly. Ants, bees, wasps also secrete and spread by grooming a layer of grease, fatty acids, secreted in gland in mouth parts, thought to help body from drying out, hydrocarbons spread over cuticle or outside body of ant. These carry a specific odor of the colony, a signature. Can take very young ants and put them in another colony in even a different species and they will be accepted because they come to take on the odor spread on them by grooming. Harvester ants: within a colony everyone has the same signature, but also ants of different tasks smell different. Conditions of the job make them smell different. Lives in desert, hot dry conditions changes chemistry of the hydrocarbons on their bodies, so outdoor ants come to smell different from the inside ants, allowing ants to recognize the tasks it does. Coal miner, calluses on carpenter's hand. Ants use the rate at which they smell each other, bump into each other, to decide what to do next. Ants respond to glass beads with the odors and changed tasks if rate was high. Are odors receivable by other species? Can humans smell them? Humans can smell some of the ants volatile pheromones like the lemon smell of one tropical species, but not the hydrocarbons conveying tasks.
14:26Experiments. Foragers: early in morning a group of patrollers goes out, and foragers won't go out unless the patrollers come back. Avoid neighboring ants. Fighting season sometimes, but mostly just avoid others for reasons of competition. If you take away the patrollers the foragers never go out. What if patrollers encounter other species, what do they do? Predator, horned lizard, might eat them. Direction matters. Each day the patrollers choose a few directions, so not only whether they come back but where they choose to mark. If you map the trails, if the older ones meet one day they are not likely to use that trail the next day. Adolescents return and keep fighting. What makes foragers go out and what direction do they choose? Mike Green, U. Colorado at Denver. If you extract the hydrocarbons from patrollers and put it onto glass beads, drop beads into the entrance, the foragers come out. Learned that it had to be at a particular rate, not too fast or too slow. Has to be about 10 seconds. Too long and it's as if the ants forget it ever happened. At first only a few patrollers go out; if successful then more. Finally some 30-50 patrollers, returning every ten seconds, which is enough to end up signaling to 30-40,000 foragers. Next question: How do foragers know what to do with response to the beads, which direction to go? Just milled around. In the absence of patrollers foragers just go where they went the day before. Kept patrollers off nest mound. Extracted glands and marked a sector where they had been disallowed, then the foragers go there.
23:49Killing ants, feeling guilty. Does the Queen have some kind of control, say, over the mix of types within the colony? Small number of patrollers relative to foragers; after a war is the Queen able to change the mix of fighters and workers? No. In some species the workers come in different sizes, sometimes large ones called soldiers. But very few species have that; and it's not clear that large ones are always fighters; might break seeds with large jaws. In movie "Ants" there is a bureaucrat who stamps the larva at birth. But in most species, like Harvester ants, their tasks change as they get older. Division of labor is self-organized, regulated by wages. Task allocation is a process that has to be driven by simple rules at the level of each ant. Sequence that an ant will go through. Start inside nest with nest-maintenance, simple trips outside, then may become a patroller if more needed, or even more foragers. But never switch back once one has become a forager or patroller, so if more need for nest-maintenance, they have to be recruited from younger ants. "If rate at which I'm meeting foragers really increases, I'm going there too." Is that division of labor? Is it what Adam Smith was talking about? Wasn't he talking about a happy village with people all doing what they do best? Ants don't necessarily get better at doing things; merely transitions imposed from the outside.
31:20It's actually not much different in economics and it is related to the division of labor. Specialization, one piece of the story, think about learning by doing. Some truth to it. But tension, doing same thing over and over can become monotonous. The real power comes from getting the right people in the job because we are not all equally good. In human world, changes in wages and non-monetary satisfaction draw people into particular tasks. Random distribution of jobs is not as beneficial for increased wealth. Russ, at 5'6", would not do as well as Larry Byrd; not just doing well, but who is doing well. When demand for foragers goes up, it would be good for the colony to increase the number of foragers. But probably all ants are equally good at foraging, works pretty much as all equal. Distributed process. Some do seem to work harder than others; many sluggards. Division of Labor podcast distinguishes between Smith and Ricardo, different illustrations of the power of specialization. Ricardo, came from different skills. Smith, division of labor is limited by the extent of the market. Example clarifying that from Jim Buchanan: with a bunch of people who hunt, as number of hunters increases, one finds merely because of sheer numbers that he can supply food, open a deli. Could all be equally skilled. Get specialization just because with enough economic activity you get specialization no matter what. Not just different skills. You presume that if there is a bad hunter he might become a deli operator. Ricardo's insight is that it's the relative productivity not the absolute productivity that matters, so it might be best even if the best hunter becomes the deli operator (if he's relatively even better at deli operation). Different focus, different intuitive parts of the story, not really different theories of specialization. Learn something from each of Ricardo's and Smith's conception of how specialization and the division of labor works.
40:06In Nature article, in this self-organizing system, without a leader, is imperfect. It's remarkable how it works without control, but it's not perfect. Can miss a big pile of seeds because they only go to certain places. Just have to find enough to keep going. Even if you put out a big pile of seeds, the foragers will walk right over it if the patrollers didn't mark it! Trails can be up to 20 meters long. Takes about 25 minutes for an ant. Each ant goes back to same place on successive trips. Patroller tells a direction but not how far to go. Find food randomly in that direction. Foragers go out in that direction; rate at which successful foragers come in determines how many more foragers go out. Sense of smell helps them find it in the sand. Intensity of foraging. What is the cost of foraging? Cost is dessication, drying out. Get their water out of metabolizing the seeds they eat; spend water being out in the sun. Respond to outdoor temperature. Can't move fast when it's cold; hot, run around fast; but at some point too hot for their feet so they stay in. No ability to sweat. Colony acts as if it's wise even though no one member is wise. Like Surowiecki's Wisdom of Crowds, Smith's invisible hand, Hayek's "Use of Knowledge in Society". Limitations of language with talking these self-organizing societies. Do you use "the colony" as the subject of a sentence? Yes, think about it all the time. Easy to talk about ants as if they have agency. Tension. Hayek called the human version of this a "marvel." Coordinatedness without a coordinator is marvelous, interesting. Makes a difference in how we understand what's happening to us. How important is it to us as humans that we think we have agency? Consider Read's "I, Pencil," suppose graphite becomes in higher demand by someone who doesn't make pencils. Makes graphite less available. How does the awareness of the demand for other goods like cars get pencil and tennis-racket makers to cut back (or supply of graphite to increase), or change production methods to use less graphite? Range of choices of how we respond is vast. What knowledge would we want someone to have to make the right decision? Unfathomably much, and may not even exist till the crisis. No way that a central authority could acquire the information and then send it out to users before yet something else changed. But each supplier only looks at one thing--the price of one good. Similar to ants just looking at the hydrocarbons. Self-regulating. Imperfect, things can get missed, but amazing that it works at all and that it works better than any other system. Each of us is like the ant. We have our task, we try to do it well, sometimes we mess up, but overall there is an immense amount of order in our society. But unlike the ants we have sentience, intelligence, can try to respond with whole new ways to do things; dynamic element that ants don't have.
54:46The ants aren't faced with the decision of whether we care more about pencils or tennis rackets. They only keep going. Isn't Russ's story taking the goals for granted? What if we wanted to encourage more of something? Who is "we"? Usually humans don't have a unanimous desire for what we should do next. Sometimes coincide, but sometimes not. Prices help keep down violence when tied to property rights and other institutional arrangements. Value. Suppose I want to stop eating meat, become a vegetarian. In an authoritarian world the people in charge of the whole system (human organizations have control but there is very little control across organizations) could go out and survey people and could possibly make plans based on the answers. But hard to decide. Market solution emerges, we have no language for it and sometimes says "the market does this" as if the market is a sentient actor, harmless in one sense, but can hamper others including economists about what is going on. In the market system, non-organized, when people become vegetarians, a bunch of stuff gets set in motion via signals--vegetarian restaurants, etc. Takes only a few months. The choices don't hamper the choices made by others, except that others may have to pay a higher price. You may care about other people's choices in addition to your own. Original example, conflict in tastes of different people: for the ants if that happens there is no possibility of its having any impact, but for us we have the option of making a decision, even if it's hard to see how to implement it; we can sit here and talk about it. We know about the extended effects. If lots of Americans decide to become vegetarians it can affect tropical forests and we're capable of knowing about that and thinking about what that means. Forces are generated by the type of system that ants operate in, analogies, but in a human system on top of that is that we can understand it and in principle make decisions about changing it. Many self-organizing systems don't work well. For ants, no choice. For us, when we see something with bad outcomes we want to fix them. Pollution, congestion; we have a natural tendency to say we can fix that. Ironically, because we don't have a lot of intuition about self-organizing behavior we often look for solutions that lead to unintended consequences because we don't understand the nature of system-wide change. So we widen 101, the Bay Area thruway; and don't understand why it's still as crowded now as before. We don't look at ants enough.