Disagreement Is Not Always Information: Task-Relevant Counterfactual Consensus Repair Under Multi-View Corruption

Disagreement between multiple observations is often treated as evidence of corruption or complementary information, but disagreement need not be useful for task-relevant repair. We ask when a sample's own disagreement pattern adds value beyond the pair consensus and scalar disagreement severity. Given clean representation $z_0$ and corrupted representations $z_1,z_2$, we define $M=(z_1+z_2)/2$ and repair target $T=M-z_0$. Under squared loss, disagreement has incremental value only when conditioning on it changes the task-visible conditional mean of $T$. We introduce Counterfactual Consensus Repair (CCR), which compares the target's own disagreement detail with a severity-matched shuffled control while holding fixed the target consensus, scalar severity, global parent, pair architecture, and optimization. Our primary contrast is $$\begin{gathered} Delt{a}_{ \\ mathrmdetail}=R_{ \\ mathrmcorrect}-R_{ \\ mathrmseverity \\ text-matched} \end{gathered}$$, so negative values indicate value from the target's own disagreement detail. Across Speech Commands V2, ESC-50, and UrbanSound8K, this contrast is consistently favorable in asymmetric, heteroscedastic, and structured cross-family regimes; in the cross-family regime, $$\begin{gathered} Delt{a}_{ \\ mathrmdetail}=-0.755,-0.903, \end{gathered}$$ and $-0.562$, respectively. Exact representation-space probes recover the predicted IID Gaussian null and heteroscedastic boundary, and the matched effect replicates on ESC-50 with a frozen pretrained BEATs representation. Zero-shot DHAuDS-Pair experiments, with no repair retraining, show that the learned disagreement-to-repair mapping transfers selectively: external corruption families can be beneficial, near null, or adverse. These results support a conditional rather than universal view of disagreement.