Imagine you've been studying two dialects of a language for years, compiling separate grammars, cataloguing separate idioms — and then someone hands you a decoder ring showing they're the same language spoken with different accents. That's what this paper does for non-Hermitian physics. The two dominant forms of non-Hermiticity — non-reciprocal couplings (where energy flows asymmetrically between neighbors, producing the skin effect) and imaginary on-site potentials (which give rise to PT-symmetry and related phenomena) — have been used together in models but always treated as fundamentally distinct ingredients. Li Ge proves they aren't. The core tool is the Lanczos transformation, a standard tridiagonalization algorithm from numerical linear algebra, repurposed here as a physics-level isospectral mapping. Given a generalized Hatano-Nelson model H with non-reciprocal, spatially varying couplings, the transformation produces a new Hamiltonian HL that has real symmetric couplings and an imaginary on-site potential — a completely different-looking system with identical eigenvalues. The mapping is exact, not approximate. It works even when the original model already contains an imaginary potential and exceptional points. Why does this matter? The non-Hermitian skin effect and PT-symmetry breaking are among the hottest topics in photonics, acoustics, and condensed matter. Dozens of groups design metamaterials and photonic circuits exploiting one or the other. Knowing these are isospectral means that any intuition, any spectral theorem, any experimental result developed in one framework transfers directly to the other. You don't need to re-derive; you translate. The paper also delivers a practical payoff: exceptional points in the Lanczos-transformed system HL are often easier to visualize and understand than in the original non-reciprocal model H. If you've struggled to build intuition for where exceptional points appear in asymmetric-coupling networks, the imaginary-potential picture may hand you that intuition for free. A third consequence is a construction recipe. Want a non-Hermitian system with a complex potential that nonetheless has an entirely real spectrum — without requiring PT symmetry? This mapping tells you how to build one. That's a direct design tool for photonic and acoustic metamaterial engineers. The scope is theoretical: 6 pages, 5 figures, no experimental validation, no numerics beyond the analytical examples. The claim is a mathematical identity, so the integrity bar is proof-level, not experiment-level. Lanczos tridiagonalization is well-established numerically; what's new is using it as a physical equivalence rather than a computational trick. The limitation is also clear: the mapping applies to one-dimensional tight-binding chains. Extension to higher dimensions or continuous systems is not addressed and would be the natural next step. The paper is a conceptual bridge — elegant, compact, and immediately useful for anyone designing non-Hermitian lattice systems.
