Pattern/Systems Thinking/No. 0318
Emergence
Emergence is the formation of patterns or abilities in a whole through interactions among its parts. Also called emergent behavior, it is studied in physics, biology and systems theory. George Henry Lewes distinguished emergents from simple sums of parts in 1875.
Also called Emergent Behavior
- Evidence
- Well established
- Read
- 6 min
- Links
- 8 connections
01You've seen this when…
- in life
At a family dinner, everyone speaks a little louder to be heard. Ten minutes later, the whole table is shouting, though nobody intends to make the evening noisy.
- at work
Each department adjusts its orders to avoid running out of stock. As those adjustments pass upstream, suppliers see huge swings in demand that customers never created.
- out in the world
Traffic slows to a standstill on an open highway. When you finally reach normal speed again, there’s nothing to pass: no crash, no construction, no obvious cause.
02The idea
Nobody at the dinner table chooses the final noise level. Each person reacts to the people nearby. Together, those reactions create something none of them controls.
That is emergence: interactions among parts generate a pattern or capacity at the level of the whole. A driver can accelerate and brake. A line of interacting drivers can produce a traffic wave. Explaining the wave requires looking at how drivers interact, beyond the details of one car’s engine or one driver’s personality.
The crucial word is interactions. A heavier bag of groceries illustrates simple addition: you can explain its weight by adding the items’ weights. A crowd that funnels itself into a dangerous crush requires a different explanation: where people stand, what they can see, how they move, and how pressure travels between them.
Emergence shifts the question from what are these parts like? to what happens when these parts affect one another? The parts still matter. But their arrangement and relationships matter too. The same people, connected by different rules, can produce a very different result.
03Why it happens
- Each part responds to a limited view. Drivers watch nearby vehicles, not the entire highway. Teams respond to their own workload, not every dependency in the organization. A locally sensible response can contribute to a collectively bad result.
- Responses become inputs for other responses. Your braking makes the next driver brake. Their braking affects the driver behind them. These feedback loops can amplify a small disturbance or dampen it, depending on how people respond.
- Connections determine what can spread. Who interacts with whom, how quickly, and through which channels changes the outcome. An identical set of components can behave differently when its connections change. A tightly connected network can spread both useful coordination and failure.
- The combined effect need not grow smoothly. A few more vehicles may be enough to turn flowing traffic into unstable traffic. Such nonlinearity means doubling an input needn’t double the outcome. Crossing a threshold can produce a different pattern altogether.
Emergent patterns vary in which of these mechanisms they use. Explaining the whole requires an account of the relationships among its parts.
04A worked example
In an experiment reported in 2008, Yuki Sugiyama and colleagues put 22 vehicles on a circular track about 230 meters long. Drivers were asked to maintain roughly 30 kilometers per hour while keeping a safe distance. There were no intersections, lane changes, or obstacles.
Small variations in spacing and speed grew. Eventually, a traffic jam formed, with some vehicles stopping temporarily. The congested region moved backward around the circle while the vehicles traveled forward.
What it looks like There must be an obstruction somewhere, or a particularly troublesome driver holding everybody up.
What’s actually going on The moving jam is a collective pattern. Drivers respond to changes ahead of them, with delays and differences in how they brake and accelerate. Under these conditions, those responses amplify fluctuations rather than smooth them out. Vehicles enter and leave the jam, but the jam continues.
What would have helped A response aimed at reducing the propagation of speed fluctuations in traffic with no obstruction to remove. Candidate measures include coordinated speed control and following behavior that keeps small changes from growing by absorbing them. Those measures need testing at the group level: helping one vehicle move faster and stabilizing traffic are different outcomes. This experiment demonstrated the failure mode. Evidence for a particular cure would require further testing.
05How to spot it
These are clues, not proof. Sometimes one broken component really is responsible.
06What to do about it
- Draw the interactions before replacing the parts. Map who reacts to whom, what information they receive, and how long each response takes. Include connections across your chosen system boundary, such as suppliers or customers.
- Measure the collective outcome. Track total completion time, traffic flow, or customer experience alongside individual performance. A better result for each department can still produce a worse result for the organization.
- Change one interaction rule. Try a different handoff, shared queue, information channel, or response timing. A pattern continually recreated by the rules connecting capable people will persist when those people are replaced.
- Use models to expose possibilities. A simple simulation can show how local actions might generate a collective pattern. Establishing that your proposed mechanism explains the actual situation requires comparison with observations.
- Test small and watch beyond the target. Use a safe-to-fail experiment where possible. Check effects on neighboring teams and downstream outcomes while tracking the number you meant to improve. Stop if the new interaction creates a harmful pattern.
07Where it doesn’t replace an explanation
Calling an outcome emergent leaves its mechanism to be explained. You still need to identify the interactions that produce it and show why your account fits the evidence. On its own, emergence gives an unexplained outcome a sophisticated label.
Emergent outcomes can still be predictable. Many emergent patterns can be modeled and reproduced. The everyday scientific sense, often called weak emergence, concerns collective behavior generated by interactions. Strong emergence makes the much larger claim that higher-level properties cannot, even in principle, be derived from lower-level facts. That claim remains philosophically contested and would require evidence beyond traffic waves.
Emergence and self-organization overlap and address different features of a system. Self-organization concerns order developing without central direction. Emergence concerns properties of a whole arising through interactions. Emergent behavior can arise in deliberately designed systems, including behavior their designers didn’t anticipate.
08Roots
In 1875, George Henry Lewes was trying to connect the study of mind with the rest of nature in Problems of Life and Mind. Chemistry offered a revealing contrast. Adding two weights produces a total you can calculate by addition. Combining chemical substances can produce something with properties unlike those of its ingredients.
John Stuart Mill had already explored this contrast between mechanical and chemical effects. Lewes gave it memorable vocabulary: resultants for effects that could be treated as combinations of their contributing causes, and emergents for qualitatively different outcomes. The distinction offered a way to discuss novelty in nature without treating every new property as a supernatural interruption.
Later thinkers extended emergence into accounts of life and mind. In 1972, physicist Philip Anderson brought a related argument into modern physics with More Is Different. Knowing fundamental laws, he argued, leaves further explanatory work at larger scales. Collective behavior requires concepts and theories appropriate to its scale. Complexity research and computer simulations subsequently made that lesson especially visible: simple interaction rules can generate elaborate patterns, without any component containing a plan for the whole.
09How solid is this?
Collective patterns generated by interactions are well documented in physics, biology, and social systems, including controlled traffic experiments. Emergence is a broad explanatory concept; specific predictions require a separate account of the system. Claims of strong emergence remain contested.
10Connections
- Can follow from Feedback Loops
- Includes Network Effects, Self-Organization
- See alsoNonlinearity, Local vs. Global Optima, Safe-to-Fail Experiment, Complex Adaptive System, System Boundary
11Origin and sources
George Henry Lewes introduced the distinction between emergents and resultants in 1875, building on John Stuart Mill’s account of combined causes. Later developed in philosophy, physics, and complexity and systems traditions.
- [1]Lewes, G. H. (1875). Problems of Life and Mind. First Series. The Foundations of a Creed. Vol. II. Trübner & Co.
- [2]Anderson, P. W. (1972). More Is Different. Science, 177(4047), 393–396.
- [3]Sugiyama, Y., Fukui, M., Kikuchi, M., Hasebe, K., Nakayama, A., Nishinari, K., Tadaki, S., & Yukawa, S. (2008). Traffic jams without bottlenecks—experimental evidence for the physical mechanism of the formation of a jam. New Journal of Physics, 10, 033001.
Suggest an edit· Updated 2026-10-02