When Does Reward Teach State? A Hidden-Automaton Instrument and the Group-Language Boundary
Does a reinforcement-learning agent that earns high reward represent its task's latent state, or only a reward-correlated shortcut? Usually this is unanswerable: the "true state" is undefined. We make it exactly answerable with a white-box instrument: express the task as a hidden deterministic finite automaton (DFA), let the agent observe a symbol stream and intermittently choose the next symbol under partial control, with a sparse terminal reward for acceptance. Knowing the automaton gives two things for free: the optimal return (so reward becomes an interpretable normalized score) and the exact latent state at every step (so we can probe the agent's representation without ever showing it). Reward success and latent-state learning become separately measured quantities whose coupling is governed by three axes. Optimizer strength: under weak on-policy RL the agent can lock onto a reward-shortcut, earning reward with the state probe at chance; whether it escapes is set by optimizer strength, budget, and task structure -- a pre-registered control isolates the optimizer (PPO+GAE), which escapes fastest but only partially and with high seed variance. Task structure: permutation (group-language) structure is a warning sign computable from the transition function before any training, and held out on 153 capacity-controlled fresh automata it flags perception gaps at precision 0.86 (89 of 103), in one direction only. Observation informativeness: a label-free auxiliary is vacuous when observations carry no state and recovers it sharply once they carry any (a threshold, saturating near beta=0.1). The payoff is a distinction reward-only evaluation cannot make: a perception gap (state not linearly recoverable, though representable) versus a planning gap (state recoverable but not exploited). Reward alone cannot localize which; high reward alone is not evidence the state was learned.