← Dynamic Systems Engineering Foundations · Semester 1, Week 3

Causality: Why Systems Change

Theme

A system's behavior cannot be understood by observation alone. Engineers seek to explain why changes occur. This week introduces causal reasoning as the foundation of prediction, diagnosis, and system design.

Learning Objectives

By the end of this week, students should be able to:

Distinguish correlation from causation.
Build simple causal models.
Identify feedback loops.
Recognize hidden variables and confounding factors.
Generate multiple competing explanations for the same observation.
Evaluate explanations based on evidence rather than intuition.

Core Vocabulary

Cause · Effect · Correlation · Causation · Necessary Condition · Sufficient Condition · Confounding Variable · Feedback Loop · Positive Feedback · Negative Feedback · Intervention · Prediction · Counterfactual

Fundamental Principles

Principle 1 — Events have causes, but systems often have multiple interacting causes.

Principle 2 — Correlation alone does not establish causation.

Principle 3 — The best causal model explains both successful and failed predictions.

Principle 4 — Always generate multiple hypotheses before selecting one.

Principle 5 — Every causal claim should suggest an observable prediction.

Anchor Example — Thermostat

Observation: The room temperature increased.

Possible causes: Heater activated. Sunlight warmed the room. A window was closed. More people entered the room.

Lesson: The observation alone does not identify the cause.

Anchor Example — Human Learning

Observation: A student's exam score improved.

Possible explanations: More practice. Better sleep. Easier questions. Improved teaching. Random variation.

Objective: Rank explanations by evidence rather than selecting a favorite.

Anchor Example — AI Assistant

Observation: The AI produced an excellent answer.

Possible explanations: Better prompt. Better retrieved information. Familiar subject. Luck in token generation. Strong reasoning process.

Task: Evaluate which explanation has the strongest supporting evidence.

Anchor Example — Traffic

Observation: Traffic suddenly slows.

Possible causes: Accident. Lane closure. High traffic volume. Weather. Phantom traffic wave.

Discussion: How local driver behavior can create system-wide congestion.

Anchor Example — Small Business

Observation: Revenue declined.

Possible causes: Reduced customer demand. Increased competition. Supply shortages. Poor marketing. Seasonal effects.

Task: Identify what additional evidence would help distinguish among these possibilities.

AI Laboratory

Choose one observation from an anchor example. Ask AI to generate five plausible explanations. Then ask: Which explanation best fits the available evidence? What additional observations would distinguish between the remaining hypotheses? What prediction follows from each explanation? Compare the AI's reasoning with your own.

Reflection Questions

What evidence supports my preferred explanation?
Which explanation have I not considered?
What evidence could disprove my current model?

Assignment

Select a real-world event. Construct: at least five competing hypotheses, evidence supporting each, evidence against each, predictions implied by each hypothesis, and your current confidence ranking.

End-of-Week Competency

A successful student understands that engineering begins with causal explanation rather than description and develops the habit of evaluating multiple competing models before drawing conclusions.

Next: Week 4 — State Estimation: Reasoning Under Uncertainty