Suppose a Hilbert space has a direct-sum decomposition H=n0Hn,\mathcal H=\bigoplus_{n\ge 0}\mathcal H_n, where the sectors are eigenspaces of a self-adjoint operator S^c\hat S_c , S^cψn=snψn.\hat S_c|\psi_n\rangle=s_n|\psi_n\rangle. Let J^\hat J be a self-adjoint generator of unitary evolution, U(τ)=eiJ^τ,U(\tau)=e^{-i\hat J\tau}, and suppose [J^,S^c]=0.[\hat J,\hat S_c]=0. A separate clock/phase operator S^t\hat S_t is introduced with the intended canonical relation [J^,S^t]=iκ.[\hat J,\hat S_t]=i\kappa. My question is: Can J^\hat J , while commuting with S^c\hat S_c , generate genuine transitions HnHm,mn,\mathcal H_n\longrightarrow\mathcal H_m,\qquad m\neq n, or must U(τ)U(\tau) preserve every eigenspace of S^c\hat S_c ? If the latter, what is the standard operator structure for introducing genuine sector-changing evolution while retaining a separate clock/phase operator? For example, would one require ladder/shift operators that do not commute with S^c\hat S_c ? I am particularly interested in the spectral/operator-algebraic issue rather than in a specific physical model.