Research article · Results

Results: is there a reason for a learning rule to develop?

Kāula research note, November 2, 2026. Results for the pre-registered gate in Is there a reason for a learning rule to develop?. Every number below comes from the pre-registered analysis script.

Verdict

All three pre-registered checks pass, but only on the contextual task.

Following the pre-registered plan, the research task moves from the plain bandit to the contextual one.

1. Headroom

The cost of a wrong τ is the reward lost by using the other delay's best τ. 2,000 lifetimes per condition; 95% bootstrap CIs.

TaskBest τ, delay 0Best τ, delay 3Cost at delay 0Cost at delay 3Pass
Plain two-armed bandit0.2540.006 [0.006, 0.006]0.005 [0.002, 0.009]No
Contextual bandit, 4 cues1120.045 [0.044, 0.047]0.151 [0.146, 0.156]Yes

The rule needs both costs to exclude 0 and the larger to be at least 0.05. In the plain bandit both costs are real but tiny, at about half a percentage point of reward: choices are sticky, so a short trace works at any delay. In the contextual bandit, the wrong τ costs 15 points of reward at delay 3. One τ cannot serve both environments.

2. Positive control

A hand-built developmental rule estimates the delay during childhood and sets τ from the headroom lookup. On the contextual task it estimated the delay correctly in every lifetime, choosing τ = 1 for twins raised at delay 0 and τ = 12 for twins raised at delay 3.

MeasureDevelopmental ruleFixed τ = 6
Adult crossover interaction0.192 [0.186, 0.198]0.000 [−0.006, 0.006]
τ-transplant effect0.196 [0.192, 0.199]0
Weight-transplant effect−0.004 [−0.009, 0.001]0.000

All three pass conditions hold: the crossover CI is above 0, the τ-transplant CI is above 0, and the fixed rule's crossover is far below a third of the developmental rule's. Transplanting τ carries the whole effect; transplanting the learned weights carries none. So the analysis can tell a developmental change in how an agent learns from a difference in what it learned.

3. Optimizer check

A mirrored evolution strategy (population 16, 40 generations) searched over τ = exp(w₀ + w₁·d̂), where d̂ is the delay estimated in childhood.

RuleAdult fitness
Best fixed τ0.727
Evolved rule0.753
Built-in rule0.755

The evolved rule recovers 93% of the gap (pass: ≥ 50%). It settled on τ ≈ 0.28 at delay 0 and τ ≈ 14.4 at delay 3, close in spirit to the built-in rule's 1 and 12.

4. Deviations from the plan

5. Limitations

What's next

The developmental learning-to-learn project uses the contextual bandit as its task, and moves on to evolving rules that must discover for themselves what to measure in childhood.

All notes · The four pre-registrations