Causal Models and Interventions › Three Questions a World Model Must Answer

Michael BrenndoerferJune 21, 20263 min read
World Models HandbookAll chapters
Part 4 of 12Space, Physics, and Causality
Chapter 21Causal Models and Interventions › Three Questions a World Model Must Answer
Reading section 2 of 14 · Three Questions a World Model Must Answer · 3 minSection 2 of 14 · 3 min

Book contents

World Models Handbook12 parts · 72 chaptersView the whole book →
Active chapter 21 of 72

Section 2 of 14 · 3 min read

Causal Models and Interventions › Three Questions a World Model Must AnswerOne section · Sections in this chapter2 / 14 ↓
Sections in this chapter14

Pearl's hierarchy of causal queries gives a compact way to see what is at stake. Causal queries in this framework fall into three rungs. Moving upward generally requires additional assumptions about how the world generates data; more passive data alone cannot establish assumptions that are observationally indistinguishable.

The Ladder of Causation
  • Rung 1: Association. "What does YY typically look like when I observe X=xX = x?" Answerable from a joint distribution P(Y∣X)P(Y \mid X). This is what purely predictive models learn.
  • Rung 2: Intervention. "What will YY look like if I make X=xX = x?" Requires knowledge of P(Y∣do(X=x))P(Y \mid \mathrm{do}(X = x)), which generally differs from P(Y∣X=x)P(Y \mid X = x).
  • Rung 3: Counterfactuals. "What would YY have been, had I acted differently, in this particular situation?" Requires a structural model that supports reasoning about a specific individual across alternative histories.

The distinction matters because different parts of an agent live on different rungs. An observational video predictor answering "what does the scene look like next?" is a Rung 1 machine. A model-predictive controller asks "what happens to the trajectory if I apply action sequence at:t+Ha_{t:t+H}?", a Rung 2 question when its action-conditioned model has valid intervention semantics. Diagnosis such as "would the collision have happened had the planner chosen the other lane?" is a Rung 3 question. Supporting all three requires more structure than association alone and generally more than intervention alone; a counterfactual-capable SCM normally also supports the lower-rung queries.

Here is the uncomfortable fact that motivates the chapter: planning poses Rung 2 queries, while passive prediction is a Rung 1 task. Model-based reinforcement learning can exploit model error, but confounding is only one source alongside approximation error, finite data, recursive rollout, and limited support. Optimization may seek favorable model errors; whether it does so depends on the planner and error landscape. Causal structure can help distinguish stable mechanisms from collection artifacts when that structure is justified and identified.

Many baseline MBRL systems minimize average prediction loss and use ensembles or penalties to discourage uncertain rollouts. Error is not uniform: finite-sample variation, approximation error, support gaps, and confounding behave differently under optimization. Noise need not average away, and a planner need not exploit every bias. Randomized tests, sensitivity analysis, robust bounds, and structural knowledge can all help separate or constrain these errors; a complete causal decomposition is not always available.