Chapter 13: Systems Thinking

Summary

This chapter develops the thinking tools that ecologists use to understand complex, interconnected systems. Students learn to build stock-and-flow diagrams, identify reinforcing and balancing feedback loops, find leverage points, and recognize emergence and resilience in ecological and human systems. After completing this chapter, students will be able to model any system using causal loop diagrams and predict how interventions ripple through interconnected networks.

Concepts Covered

This chapter covers the following 27 concepts from the learning graph:

Systems Thinking
System
Stocks and Flows
Feedback Loops
Reinforcing Feedback
Balancing Feedback
Causal Loop Diagrams
Stock and Flow Diagrams
Leverage Points
Emergence
Resilience
Regime Shifts
Scale and Hierarchy
Mental Models
Unintended Consequences
Time Delays
Complex Systems
Dynamic Equilibrium
Positive Feedback
Negative Feedback
System Boundaries
Interconnectedness
Trade-Offs
Cascading Effects
Adaptive Management
Ecosystem Resilience
Threshold

Prerequisites

This chapter assumes only the prerequisites listed in the course description.

Bailey Says: Welcome, Builders!

This might be the most important chapter in the entire course. Why? Because systems thinking is the lens through which ecologists see the world. Once you learn to think in systems, you'll never look at a forest, a lake, or even your school cafeteria the same way again. Everything's connected -- and now you'll finally have the tools to map those connections! Let's build on that!

What Is a System?

A system is a set of interconnected parts that function together as a whole. Your body is a system. A forest is a system. A city's water supply is a system. A smartphone is a system. What makes something a system rather than just a random collection of parts? Three things:

Elements -- the individual components (trees, animals, water, soil)
Interconnections -- the relationships between those elements (nutrient flows, predation, competition)
Purpose or function -- the overall behavior that emerges from the elements working together (a forest ecosystem cycling nutrients and supporting biodiversity)

Remove any one of these and you no longer have a system. A pile of car parts in a junkyard has elements, but without the interconnections of proper assembly, it has no function. And here's the key insight: you can't understand a system by studying its parts in isolation. You have to study the relationships.

Systems thinking is the practice of understanding how things influence one another within a whole. Instead of asking "What is this thing?" systems thinkers ask "What is this thing connected to?" and "What happens when I change one part?"

Seeing the World Through Systems: Mental Models

Before we build formal diagrams, let's talk about what's already in your head. A mental model is an internal representation of how something works. You have mental models for everything -- how your school schedule works, how traffic flows, how a campfire burns.

The problem? Most mental models are incomplete. They capture the parts we notice and miss the parts we don't. A farmer's mental model of crop growth might focus on sunlight and water but miss the complex soil microbiome. A city planner's mental model of traffic might focus on roads but miss the feedback loop where building more roads actually increases traffic (a phenomenon called induced demand).

Systems thinking upgrades your mental models. It gives you a toolkit for making your invisible assumptions visible and testing them against reality.

What happens when our mental models are wrong? We get unintended consequences -- outcomes we didn't predict because we didn't understand the full system. Introducing cane toads to Australia to control beetles seemed logical. The toads ignored the beetles and became an invasive disaster. The mental model was too simple: it assumed a linear cause-and-effect chain instead of a web of interactions.

Stocks and Flows: The Building Blocks

Every system can be described in terms of stocks and flows. A stock is an accumulation -- something you can measure at a point in time. The water in a lake. The population of deer in a forest. The carbon dioxide in the atmosphere. The money in your bank account.

A flow is a rate of change -- something that fills or drains a stock over time. Rain fills the lake (inflow). Evaporation drains the lake (outflow). Births increase the deer population (inflow). Deaths decrease it (outflow).

The relationship is intuitive:

If inflows exceed outflows, the stock grows
If outflows exceed inflows, the stock shrinks
If inflows equal outflows, the stock stays constant -- this is dynamic equilibrium

Dynamic equilibrium doesn't mean nothing is happening. Water is still flowing into and out of the lake. Deer are still being born and dying. But the level stays roughly constant because inputs balance outputs. It's like a bathtub with the faucet running and the drain open at the same rate -- the water level holds steady even though water is constantly moving.

Stock and flow diagrams are visual tools that represent these relationships. Stocks are drawn as rectangles (like a tank or reservoir). Flows are drawn as arrows with a valve or spigot symbol. Here's the basic notation:

▭ Rectangle = Stock (accumulation)
➤ Arrow with valve = Flow (rate of change)
☁ Cloud = Source or sink outside the system boundary

Diagram: Interactive Stock and Flow Sandbox

Interactive Stock and Flow Sandbox

Type: microsim sim-id: stock-flow-sandbox
Library: p5.js
Status: Specified

Bloom Level: Apply Bloom Verb: Construct Learning Objective: Construct a stock and flow diagram for a simple ecological system and predict how the stock changes over time Instructional Rationale: Hands-on manipulation of inflow/outflow rates builds intuition about accumulation and equilibrium

A sandbox simulation with a central stock visualized as a tank/reservoir with a water level. Two sliders control inflow rate (0-10 units/sec) and outflow rate (0-10 units/sec). The tank visually fills or drains in real time. A line chart below tracks the stock level over time. A "dynamic equilibrium" indicator lights up green when inflow equals outflow. Preset buttons load ecological scenarios: "Lake Water Budget" (rain vs. evaporation), "Deer Population" (births vs. deaths), "Carbon in Atmosphere" (emissions vs. absorption). Each preset sets appropriate labels and initial values. Colors: blue water in tank, green for inflow arrow, red for outflow arrow, yellow highlight when equilibrium achieved.

System Boundaries: Where Does the System End?

Every time you analyze a system, you must decide what's inside your system and what's outside. This is called setting the system boundaries. The boundary defines what you're studying and what you're treating as external.

Choose boundaries too narrow, and you miss important connections. Study a single pond without considering the watershed that feeds it, and you'll miss the source of nutrient pollution. Choose boundaries too wide, and the system becomes unmanageably complex. Try to model the entire biosphere at once, and you'll drown in data.

There's no single "correct" boundary. The right boundary depends on the question you're asking. Studying fish populations? Your boundary might include the lake and its immediate watershed. Studying climate change? Your boundary needs to include the entire atmosphere, oceans, and land surface.

Bailey Says: Think About This!

Here's a systems riddle: If you draw your system boundary around a single beaver dam, you see a pond. Expand the boundary to the whole stream, and you see how my family's dams create a chain of wetlands that slows floodwaters and filters pollutants for miles downstream. Same beaver dam -- completely different understanding! The boundary you choose shapes the story you see.

Feedback Loops: The Engine of System Behavior

Feedback loops are the most important concept in systems thinking. A feedback loop occurs when a change in a stock affects the flows into or out of that same stock. The output of a process circles back to become an input. There are two fundamental types.

Reinforcing Feedback (Positive Feedback)

Reinforcing feedback (also called positive feedback) amplifies change. A change in one direction triggers more change in the same direction. It's a snowball effect.

The term "positive" doesn't mean "good." It means the feedback adds to the direction of change. Reinforcing feedback can amplify growth or amplify collapse.

Examples of reinforcing feedback:

Compound interest: More money earns more interest, which creates more money, which earns even more interest
Algal blooms: More nutrients cause more algae, more dead algae cause more decomposition, more decomposition removes more oxygen, fewer fish mean fewer algae-eaters, so algae grow even faster
Melting permafrost: Warming releases methane from permafrost, methane traps more heat, more warming melts more permafrost
Soil erosion: Less vegetation leads to more erosion, which removes topsoil, which supports less vegetation

Left unchecked, reinforcing feedback loops produce exponential growth or exponential collapse. But in nature, nothing grows forever.

Balancing Feedback (Negative Feedback)

Balancing feedback (also called negative feedback) counteracts change. A change in one direction triggers a response in the opposite direction, pushing the system back toward equilibrium.

Again, "negative" doesn't mean "bad." It means the feedback subtracts from the direction of change. Balancing feedback is what keeps systems stable.

Examples of balancing feedback:

Predator-prey cycles: More rabbits support more foxes, more foxes eat more rabbits, fewer rabbits support fewer foxes, fewer foxes allow rabbits to recover
Thermostat: Temperature rises above the set point, the heater turns off, temperature drops, heater turns back on
Body temperature: You get hot, you sweat, evaporation cools you down
Density-dependent population growth: As a population grows, competition for resources increases, slowing growth

Most real systems contain both reinforcing and balancing loops working simultaneously. The behavior you observe depends on which loops dominate at any given time.

Diagram: Feedback Loop Explorer

Feedback Loop Explorer

Type: microsim sim-id: feedback-loop-explorer
Library: p5.js
Status: Specified

Bloom Level: Analyze Bloom Verb: Differentiate Learning Objective: Differentiate between reinforcing and balancing feedback loops and predict their effects on system behavior Instructional Rationale: Side-by-side animated loops with real-time graphs make abstract feedback dynamics visible and comparable

Split screen. Left panel: a reinforcing feedback loop visualized as a circular arrow diagram with animated flow showing amplification. A stock (e.g., "algae population") grows exponentially as the loop runs. Line chart shows exponential curve. Right panel: a balancing feedback loop with animated flow showing counteraction. A stock (e.g., "body temperature") oscillates toward a set point. Line chart shows damped oscillation converging to equilibrium. Dropdown menu selects different ecological scenarios for each type: reinforcing (ice-albedo, erosion cycle, invasive species) and balancing (predator-prey, thermoregulation, nutrient cycling). Speed slider controls animation rate. Labels on arrows show the causal relationships. Colors: reinforcing loop in warm red/orange, balancing loop in cool blue/green. "R" and "B" symbols mark loop types per standard notation.

Causal Loop Diagrams: Mapping the Connections

A causal loop diagram (CLD) is a visual map showing how variables in a system are connected and influence each other. It's the systems thinker's most versatile tool. Here's how to read one:

Variables are written as named nodes
Arrows connect variables that influence each other
A "+" sign on an arrow means "same direction" -- if the cause increases, the effect increases (or if cause decreases, effect decreases)
A "−" sign on an arrow means "opposite direction" -- if the cause increases, the effect decreases
A loop marked "R" is reinforcing
A loop marked "B" is balancing

To determine loop type: trace around the loop and count the minus signs. An even number (including zero) means reinforcing. An odd number means balancing.

Here's a classic ecological CLD -- the predator-prey cycle:

Rabbit Population —(+)→ Fox Population —(−)→ Rabbit Population
        ↑                                           |
        └───────────────(B)─────────────────────────┘

More rabbits lead to more foxes (+). More foxes lead to fewer rabbits (−). One minus sign means it's a balancing loop (B). This is why predator-prey populations oscillate rather than growing forever or crashing to zero.

Bailey Says: See How It All Fits Together?

Causal loop diagrams are like X-ray vision for ecosystems! Once you learn to draw them, you can map the hidden feedback loops in ANY system -- from a forest to a friendship to a school budget. It's the same toolkit everywhere. Dam useful, right?

Time Delays: Why Systems Surprise Us

Time delays are gaps between a cause and its effect. They are one of the main reasons systems produce surprises and why people make bad decisions about systems.

Consider this scenario: You notice a lake's water quality declining. You implement strict new pollution controls. But the lake continues to deteriorate for years. Why? Because pollutants already in the sediment take decades to work through the system. The delay between action and result makes it tempting to give up and conclude "the regulations didn't work" -- when they actually just haven't had time to take effect yet.

Time delays create several problems:

Overshoot: By the time you realize you've gone too far, momentum carries you further (population growth overshooting carrying capacity)
Oscillation: Delays in feedback create boom-bust cycles (predator-prey oscillations)
Policy resistance: People abandon effective policies because results aren't immediate

Complex Systems and Emergence

A complex system is a system with many interacting parts whose collective behavior cannot be predicted simply by understanding the individual parts. Ecosystems are classic complex systems. So are economies, cities, and the internet.

Emergence is the phenomenon where complex system-level behaviors arise from simple interactions between components. No single ant knows the blueprint for an ant colony. No single neuron contains a thought. No single tree decides to create a forest microclimate. Yet ant colonies build elaborate structures, brains produce consciousness, and forests regulate their own temperature and humidity.

Emergent properties are often surprising because they are qualitatively different from the properties of individual parts:

Individual water molecules have no "wetness" -- wetness emerges from billions of molecules interacting
Individual birds follow simple rules -- but a flock produces mesmerizing murmuration patterns
Individual wolves hunt differently than a pack -- pack hunting strategies emerge from social coordination
Individual trees transpire water -- but a forest generates its own rainfall patterns

You cannot predict emergent properties by studying parts in isolation. This is why reductionism (breaking things into parts and studying each separately) has limits. Systems thinking complements reductionism by focusing on the interactions.

Diagram: Emergence Simulator

Emergence Simulator

Type: microsim sim-id: emergence-simulator
Library: p5.js
Status: Specified

Bloom Level: Understand Bloom Verb: Explain Learning Objective: Explain how complex system-level patterns emerge from simple individual rules Instructional Rationale: Agent-based flocking simulation makes the abstract concept of emergence tangible and visually compelling

A flocking simulation (boids algorithm) where each agent follows three simple rules: separation (avoid crowding neighbors), alignment (steer toward average heading of neighbors), cohesion (steer toward average position of neighbors). Students adjust sliders for each rule's strength (0-100%). With all sliders at zero, agents move randomly. As rules are activated, emergent flocking behavior appears. A second mode shows "ecosystem emergence" -- agents representing producers, herbivores, and predators with simple eat/reproduce/die rules producing emergent population cycles. Counter displays: number of agents, average cluster size, number of distinct flocks. Colors: agents colored by their current heading direction (using HSB color wheel), producing beautiful flowing color patterns in flocks. Background: dark blue.

Interconnectedness and Cascading Effects

Interconnectedness is the principle that elements within a system are linked through multiple pathways. In ecology, everything really is connected to everything else -- but some connections matter more than others.

When a disturbance hits one part of an interconnected system, it can trigger cascading effects -- a chain reaction that ripples through the network. The classic example is the reintroduction of wolves to Yellowstone National Park:

Wolves returned → elk populations decreased and changed behavior
Elk stopped overbrowsing riverside willows → willows recovered
More willows → more shade over streams → cooler water temperatures → better habitat for trout
More willows → more food for beavers → beavers built dams → more wetlands
More wetlands → more habitat for songbirds, amphibians, and insects
Stabilized riverbanks → reduced erosion → rivers changed their physical shape

Wolves changed the rivers. That's cascading effects through an interconnected system.

Scale and Hierarchy

Systems operate at multiple scales and hierarchies. A cell is a system. An organism containing trillions of cells is a system at a higher scale. A population of organisms is a system at an even higher scale. An ecosystem. A biome. The biosphere.

Each level has properties that emerge at that scale and don't exist at lower scales. A single tree doesn't have "species diversity." A forest does. A single organism doesn't have a "birth rate." A population does.

The hierarchical structure means:

Higher levels constrain lower levels (climate determines which biomes exist, which determines which species can live there)
Lower levels produce the components of higher levels (individual organisms produce the populations that make up communities)
Interventions at different scales have different effects (protecting one tree vs. protecting an entire watershed)

Leverage Points: Where to Intervene

Leverage points are places within a system where a small change can produce large effects. Not all interventions are equal. Push a system in the right place, and everything shifts. Push in the wrong place, and nothing happens -- or things get worse.

Donella Meadows, one of the pioneers of systems thinking, identified a hierarchy of leverage points from least to most effective:

Leverage Point Type	Example	Effectiveness
Numbers (subsidies, taxes, standards)	Setting a pollution limit at 50 ppm	Low
Buffer sizes (stocks relative to flows)	Increasing wetland area for flood control	Low-Medium
Structure of material flows	Adding a recycling loop to waste stream	Medium
Length of delays	Faster water quality monitoring	Medium
Strength of feedback loops	Making pollution costs visible to polluters	Medium-High
Information flows	Publishing water quality data publicly	High
Rules of the system	Changing who has property rights to clean water	High
Goals of the system	Shifting from "maximize production" to "maximize sustainability"	Very High
Paradigm (mindset)	Changing from seeing nature as a resource to seeing nature as a partner	Highest

Bailey Says: Here's a Helpful Tip!

When trying to solve an environmental problem, start by mapping the system. Find the feedback loops. Then look for leverage points. Changing a rule or a goal is almost always more effective than just tweaking a number. It's the difference between patching a dam one leak at a time versus redesigning the whole structure!

Trade-Offs: No Free Lunches

Every intervention in a system involves trade-offs -- gaining something in one area while giving up something in another. Building a dam generates clean electricity but destroys river ecosystems. Banning a pesticide protects wildlife but may reduce crop yields.

Systems thinking helps you see trade-offs clearly because it forces you to trace the full consequences of a decision, not just the intended ones. A systems thinker asks:

What do we gain?
What do we lose?
Who gains and who loses?
What are the short-term versus long-term trade-offs?
Are there solutions that minimize trade-offs?

Resilience, Thresholds, and Regime Shifts

Resilience is the ability of a system to absorb disturbance and still maintain its basic structure and function. A resilient forest can withstand a drought and recover. A resilient coral reef can survive a bleaching event and bounce back.

Ecosystem resilience specifically refers to how much disturbance an ecosystem can absorb before it fundamentally reorganizes into a different state. Think of it like a ball sitting in a bowl. Small pushes rock the ball, but it returns to the center. That's resilience. But push hard enough, and the ball rolls over the rim into a different bowl entirely.

That rim is the threshold -- the critical point beyond which the system shifts to a fundamentally different state. Cross it, and you get a regime shift -- an abrupt, often irreversible reorganization of the system.

Real-world regime shifts:

Clear lake → algae-dominated lake: Nutrient loading slowly degrades a clear lake until it suddenly flips to a permanently turbid, algae-dominated state
Coral reef → algae reef: Overfishing removes herbivorous fish, and a bleaching event tips the reef from coral-dominated to algae-dominated
Grassland → desert: Overgrazing reduces vegetation cover past a threshold, triggering erosion and desertification that cannot be reversed simply by removing cattle
Arctic sea ice → open ocean: As ice melts, darker ocean absorbs more heat (reinforcing feedback), accelerating further melting past a point of no return

Diagram: Resilience Ball-in-Basin Model

Resilience Ball-in-Basin Model

Type: microsim sim-id: resilience-basin
Library: p5.js
Status: Specified

Bloom Level: Analyze Bloom Verb: Predict Learning Objective: Predict when a system will undergo a regime shift based on disturbance magnitude and resilience Instructional Rationale: The ball-in-basin metaphor makes the abstract concept of regime shifts physically intuitive

A physics-based simulation showing a ball (representing ecosystem state) sitting in a landscape of basins (representing stable states). Two basins are shown: one labeled "Healthy Forest" and one labeled "Degraded Grassland" with a ridge between them (the threshold). Students apply disturbance by clicking and dragging the ball or using a "disturbance" slider. Small disturbances rock the ball but it returns to its basin. A large enough disturbance pushes it over the ridge into the other basin. A second slider controls "resilience" which changes the depth/width of the healthy basin -- deeper basin = more resilient. A "slow degradation" button gradually makes the healthy basin shallower (reducing resilience) until even a small disturbance causes a regime shift. Real-time labels show: current state, resilience level, distance to threshold. Colors: healthy basin in green, degraded basin in brown/orange, ball in blue, threshold ridge in red.

Bailey Says: You've Got This!

Regime shifts can sound scary, and they should be taken seriously. But here's the hopeful part: understanding thresholds means we can monitor systems and act BEFORE they cross the point of no return. Early warning signs include slower recovery from small disturbances and increasing variability. Knowing what to watch for is half the battle!

Adaptive Management: Learning While Doing

Given the complexity of ecological systems, we can never fully predict the consequences of our actions. Adaptive management is a structured approach that embraces this uncertainty. Instead of implementing a rigid plan and hoping for the best, adaptive management treats every management action as an experiment:

Assess the current state of the system
Design an intervention based on the best available understanding
Implement the intervention
Monitor the results carefully
Evaluate whether the results match predictions
Adjust the approach based on what you learned
Repeat

This is systems thinking in action. You acknowledge that your mental model is incomplete, so you build in mechanisms to learn and adapt. It requires humility, good monitoring data, and the willingness to change course when evidence demands it.

Diagram: Adaptive Management Cycle

Adaptive Management Cycle

Type: diagram sim-id: adaptive-management-cycle
Library: p5.js
Status: Specified

Bloom Level: Apply Bloom Verb: Implement Learning Objective: Implement the adaptive management cycle for an ecological management scenario Instructional Rationale: Interactive cycle with scenario branching shows how real-world uncertainty requires iterative learning

A circular diagram showing the six steps of adaptive management as clickable nodes arranged in a circle. Connecting arrows show the cycle direction. When a student clicks a step, the center panel shows a scenario example (managing wolf reintroduction, controlling invasive species, or restoring a wetland -- selectable via dropdown). At each step, students make a choice from 2-3 options. Their choice affects outcomes in later steps, demonstrating how different decisions cascade through the management process. After completing one full cycle, students can see how their monitoring data informs adjustments in cycle 2. A running log shows decisions made and outcomes observed. Colors: each step is a different color in a warm-to-cool spectrum around the circle. Active step is highlighted and enlarged.

Putting It All Together

Systems thinking isn't just an academic exercise. It's a survival skill for navigating a complex, interconnected world. Climate change, biodiversity loss, pollution, resource depletion -- these are all systems problems that resist simple solutions.

The core principles to carry forward:

Think in loops, not lines. Every effect is also a cause.
Identify stocks and flows to understand what's accumulating and what's changing.
Find the feedback -- reinforcing loops amplify, balancing loops stabilize.
Watch for delays -- they cause overshoot and oscillation.
Respect emergence -- system behavior can't be predicted from parts alone.
Look for leverage points -- small, well-placed interventions beat large, poorly-placed ones.
Monitor for thresholds -- regime shifts are hard to reverse.
Practice adaptive management -- learn while doing, adjust while going.

These tools apply to every chapter in this course. The nutrient cycles of Chapter 4, the pollution cascades of Chapter 12, the climate dynamics ahead -- they're all systems. Now you have the language and the diagrams to make sense of them.

Bailey Says: Incredible Work, Builders!

You've just learned the most powerful thinking toolkit in all of ecology! You can now see feedback loops, stocks and flows, and leverage points everywhere. From beaver dams to global climate, it's all systems within systems within systems. Dam, that's powerful! In Chapter 14, we'll sharpen another essential skill -- scientific literacy. Let's build on that!

Self-Test Questions

What is the difference between a reinforcing feedback loop and a balancing feedback loop?

A reinforcing feedback loop (positive feedback) amplifies change -- it pushes a system further in the direction it's already moving, producing exponential growth or collapse. A balancing feedback loop (negative feedback) counteracts change -- it pushes a system back toward equilibrium. In a causal loop diagram, reinforcing loops have an even number of minus signs (including zero), while balancing loops have an odd number of minus signs.

Explain why time delays make systems difficult to manage.

Time delays create a gap between when an action is taken and when its effects become visible. This causes several problems: (1) overshoot -- by the time you see the effect, you've already gone too far; (2) premature abandonment -- people give up on effective policies before results appear; (3) oscillation -- the delay between cause and effect creates boom-bust cycles as the system repeatedly overshoots and undershoots its target.

A lake gradually receives increasing nitrogen runoff. For years, the water stays clear. Then suddenly, it turns green with algae and never recovers. Which systems thinking concepts explain this?

This scenario illustrates a threshold and regime shift. The lake had resilience -- it could absorb increasing nutrient loads while maintaining its clear-water state. But that resilience had a limit (the threshold). Once nutrient loading crossed the threshold, the system underwent a regime shift from a clear-water state to an algae-dominated state. The shift was abrupt because crossing the threshold activated reinforcing feedback loops (more algae → more shading → less competition from submerged plants → even more algae). The regime shift is difficult to reverse because the system is now in a new stable state with its own balancing feedback loops maintaining the degraded condition.

What is emergence, and why does it matter for ecology?

Emergence is the phenomenon where complex behaviors or properties arise at the system level that cannot be predicted from studying individual components alone. It matters for ecology because many critical ecosystem properties are emergent: nutrient cycling, climate regulation, food web stability, and biodiversity are all properties of the whole system, not of any individual organism. This means you cannot understand an ecosystem by studying species in isolation -- you must study their interactions. It also means that removing or adding species can have surprising, unpredictable effects on the whole system.

Why are leverage points near the top of Meadows' hierarchy (goals, paradigms) more effective than those at the bottom (numbers, parameters)?

Leverage points at the top of the hierarchy change the fundamental structure and purpose of a system, which then automatically adjusts everything downstream. Changing the goal of a system from "maximize short-term profit" to "maximize long-term sustainability" transforms every decision made within that system. Changing a paradigm reshapes how people perceive and interact with the entire system. In contrast, tweaking numbers (like a pollution limit from 50 to 45 ppm) only adjusts a single parameter without changing the underlying structure or incentives that produced the problem. The system often compensates for parameter changes, but it cannot easily compensate for a fundamental shift in goals or worldview.

See Annotated References